Lessons Learned from the Rwanda and Burundi Emergencies

EVAL/04/96/Rev.1

SCOPE OF THE REVIEW

This review forms part of an internal process of examining the lessons learned from the major refugee emergencies occuring in the Great Lakes Region of Africa during 1993 and 1994. The process began in early 1995 with preparations for a major workshop which would have brought together all the key players in these emergencies, from Headquarters and the field, to reflect upon and learn from the experience. Unfortunately, the continuing crisis in the Region led to the postponement, and later the cancellation of this workshop. The lessons learned exercise has continued, however, during the course of 1995 and 1996 through a variety of initiatives, and conference room papers summarising the progress made to date were submitted to inter-sessional meetings of the Executive Committee in 1995 and 1996.

In parallel to this internal process, the multi-donor Joint Evaluation of Emergency Aid to Rwanda, commissioned by OECD/DAC, which attempted to draw lessons from a much broader perspective, was launched in early 1995 and completed in early 1996. The recommendations of this major study included several of direct relevance to UNHCR, and the examination of these recommendations has, to some extent, been incorporated into the internal process of reflection.

The current paper brings together a number of related initiatives, the most recent of which was a small workshop held in Geneva in May 1996 whose participants included many of those involved in the Great Lakes emergencies two years earlier. The agenda of the workshop addressed not only some of the recurring themes emerging in-house over the past two years, but also some of the more pertinent recommendations from the Joint Evaluation. The aims of the paper are threefold: to record the main lessons learned at the time of the emergencies, both at Headquarters and in the field; to specify action taken since then to improve UNHCR's emergency response capacity further in the light of the experience gained; and to make recommendations on the many issues which still require action.

Stefan Sperl, a former UNHCR staff member who now teaches at the School of Oriental and African Studies in London, played the role of facilitator at the May 1996 workshop, and prepared this report, drawing on the extensive documentation gathered and written over the past two years and interviews with many of those involved. Alan Simmance, former Director of Assistance, worked with the Inspection and Evaluation Service during the first phase of the lessons learned process, and drafted an interim report for presentation to the inter-sessional meeting of the Executive Committee in June 1995. The work of both consultants was conducted in close collaboration with Christine Mougne of the Inspection and Evaluation Service who has coordinated the lessons learned process since early 1995.

I. OVERVIEW

1. The refugee emergencies occuring in the Great Lakes Region during 1993 and 1994 presented the international community with a challenge almost unprecedented in recent history. A series of increasingly complex emergencies, following rapidly one after the other, tested UNHCR's recently strengthened emergency response capacity to the limits. According to one of the senior managers involved in the emergency response, by the time the massive influx into Goma took place in July 1994, UNHCR had almost completely exhausted its resources. The fact that UNHCR's response to the latter emergency has since been acknowledged as "impressive" (OECD/DAC evaluation report) is a credit to the organisation and to the individuals involved, and is compelling evidence in support of the Emergency Response Team (ERT) formula.

2. The lessons learned process, initiated by the former Central Evaluation Service in January 1995 and culminating in the present paper, has, however, identified a number of issues arising from the Great Lakes emergencies which require prompt attention to ensure that UNHCR is in a position to maintain its leadership role and to respond appropriately to future large-scale and complex emergencies. Significant weaknesses have been identified in the protection response to the critical early stages of the emergency, which, in turn, impacted negatively on the organisation's ability to develop appropriate policy. The operational response, while generally viewed in extremely positive terms, still requires improvement in certain areas. Finally, significant management weaknesses identified in the response to the Great Lakes emergencies need to be carefully reviewed and addressed within the framework of the Delphi process.

Protection and Policy Issues

3. The internal lessons learned process has identified emergency protection capacity as being one of the major weaknesses in the response to the Rwanda emergency. The exodus from Rwanda in 1994 presented UNHCR with a host of very serious protection problems, yet few of them were addressed in the early stages of the emergency. The refugee leadership was implicated in genocide, a situation of anarchy reigned in many of the camps and there was wide-spread intimidation and abuse of the rights of women and children. It should be recalled that the mid-1994 outflow from Rwanda into eastern Zaire involved the virtual transplantation of a political, social and security structure, a structure which survived intact with the support of the international community. Indeed, the control by former authorities over the general refugee population was reinforced in the early stages of the emergency by their involvement in commodity distribution. Reports submitted by the ERTs document the lack of sufficient senior protection staff in the entire area during the early stages of the operation. Protection problems were, as a result, inadequately dealt with and protection considerations were not given sufficient priority in the formulation of an overall policy.

4. Another serious weakness, despite UNHCR's long involvement in the countries of origin, was a lack of understanding of the social, cultural and political background of the refugees. A stronger protection presence in the field, combined with a better understanding of the refugees' background, would have had a number of beneficial effects, notwithstanding the extreme complexity of the problems involved. The protection needs of women and children would have been identified and addressed more swiftly, the search for credible interlocutors among the refugees would have been facilitated and the crucial switch from a food distribution system involving a frequently corrupt leadership to a more family-based one could have been speeded up. In particular, UNHCR would have been in a position to develop more rapidly a coherent policy to deal with the various disruptive forces active in the camps; these included former militia members, army contingents and government authorities many of whom may have been subject to the exclusion clause due to their involvement in the genocide. Last, but not least, the viability of voluntary repatriation as a "quick solution" to the problem would have been more realistically assessed.

5. Considering the outstanding progress achieved in many technical areas of emergency response, and the fact that protection is UNHCR's principal mandate, the lack of comparable advances in the field of protection demands urgent attention, especially as it had already been the subject of criticism in previous emergencies. The need for remedial action was raised in the Conference Room Paper on Lessons Learned from the Rwanda Emergency which was submitted to the Sub Committee on Administrative and Financial Matters (SCAF) in 1995 (Annex C), but there appears to have been no follow-up to date. As noted in the recommendations of this paper, a strategy involving a variety of measures must be adopted urgently to ensure that the protection aspect of UNHCR's emergency response capacity is greatly strengthened.

6. It has to be stressed that a major factor which prevented UNHCR from carrying out its protection responsibilities effectively was the poor security situation in the camps, especially in Zaire. However, the lack of an early consensus on the appropriate measures to take to address the problem put the lives of aid workers and innocent refugees at risk. By late 1994, it was clearly impossible to separate the leaders from the rest of the population and senior protection staff on mission to the region, called for consideration of the possible withdrawal of UNHCR. Their call went unheeded. Serious security problems due to the weakness or absence of government law enforcement agents in the country of asylum have been a recurrent feature of recent emergencies. There is a clear need to develop, both in-house and on an inter-agency basis, a range of policy options and measures which may be adopted swiftly to increase camp security at an early stage in emergencies. This should be an integral part of preparedness and contingency planning in the protection domain.

7. UNHCR's position with regard to IDPs in the Great Lakes region has also been a focus of the lessons learned process. The organisation's reluctance to deal directly with IDPs in Rwanda before the genocide was, in retrospect, a lost opportunity, since subsequent events showed that there was no agency capable of taking a lead with regard to this population. A proposal made by the Division of International Protection in mid-1994 to integrate the response to both internal and external displacements into a comprehensive strategy was not pursued. The subsequent uncoordinated and peacemeal response to IDPs in post-genocide Rwanda was in sharp contrast to the well-coordinated assistance response to the refugees. The OECD/DAC evaluation referred to the lack of an overall UN strategy to bridge the gap between the country of asylum and the country of origin as "the hollow core". UNHCR should carefully reassess its potential role with regard to IDPs in major refugee crises in the future.

The Assistance Operation

8. From the assistance point of view, a number of lessons can be drawn from the Great Lakes experience. Firstly, it is clear that UNHCR's emergency response capacity, spearheaded by EPRS, PTSS and STS and supported by stand-by agreements with fellow agencies, represents a major operational asset for UNHCR which should not only be maintained but further strengthened to permit the organisation to retain its international lead in this respect. The experience in the Great Lakes demonstrated the soundness of the emergency response capacity which UNHCR had built-up after the Gulf War and consolidated further as a result of experience gained in former Yugoslavia. This is born out by the findings of the OECD/DAC report which comments on the "impressive performance of UNHCR's Emergency Response Teams", the successful technical coordinating role carried out by UNHCR and the quality of its staff which is described in several instances as "highly competent". In fact, the OECD/DAC findings show that no other UN agency had been capable of fielding a comparable level of leadership and well-coordinated technical expertise.

9. The emphasis on technical expertise in the reports of the OECD/DAC study underscores the importance of this element in effective emergency response. It is clear that without experienced technical support there can be no emergency response. The Great Lakes experience has shown that technical know-how is an indispensable element in UNHCR's lead agency function, which has proved vital for operational success, and in which PTSS staff have built-up a level of experience probably unrivalled in any other organisation.

10. Two related technical sectors in which serious weaknesses were identified during the Burundi-Rwanda emergencies are logistics and telecommunications, both of which are fundamental to an effective operational response. These weaknesses point to the need for a particular focus on these areas, starting with the contingency planning phase, which should involve logistics and telecommunications surveys of the potential theatre of operations so that gaps can be addressed as part of the preparedness activities. Another area where problems surfaced throughout the region concerns communication and coordination between UNHCR and WFP. Repeated blockages in the WFP food pipeline resulted in the lives and well-being of refugees being put at risk and became a source of much frustration for UNHCR staff who in several instances called for direct procurement of delayed food items by UNHCR. As has been pointed out on many previous occasions, there is a need for a wide range of measures to be taken from the outset of an emergency so as to establish all necessary parameters for effective joint action by the two organisations.

11. Another important lesson from the Great Lakes emergency operation is that emergency preparedness is a dynamic topic in a rapidly changing environment. It involves many sections within UNHCR and has increasingly important implications for inter-agency coordination. There is an urgent need to establish a framework for regular in-house consultation between the various players which will not only enhance preparedness but will also greatly improve effective cooperation between the sections concerned when an emergency occurs. The Great Lakes experience has shown how vitally important it is to get things right from the start of an emergency. UNHCR staff have the necessary knowledge and experience, and the ERT concept has proved to be the most effective way to get the right staff to the field at the right time.

12. With respect to administration, it was found that efforts made by UNHCR in recent years to strengthen the administrative aspect of its emergency response capacity have certainly borne fruit. The Guidelines for the Emergency Administrator were found to be of excellent use and the inclusion of Senior Emergency Administrators (SEAs) and Emergency Finance and Administrative Assistants (EFAAs) in the ERTs proved to be of vital importance. It was found, however, that in major emergency operations, a sufficient number of senior administrators need to be deployed in order to cater not only for the establishment of new offices in the emergency zones, but also for the strengthening of existing offices in the region and the training of new local staff.

13. As in previous emergencies, UNHCR had considerable difficulties in the Great Lakes in finding enough experienced personnel to fill all vacant posts, replace ERT members upon the termination of their deployment and allow sufficient overlap for briefing purposes. Too few senior staff members were willing to accept posts in remote and insecure duty stations: general service staff, able and willing to accept such posts, in some cases were prevented from doing so due to lack of seniority, and many newly-recruited staff members were insufficiently trained. This underscores the fact that while the ERT concept has been successful as such, the recruitment, training and deployment of staff for longer-term service in emergencies still requires further improvement.

Management Issues

14. A major deficiency identified during the Great Lakes emergency was the lack of a strong centralised regional approach to operations management. This stemmed from weaknesses both at Headquarters and in the field. At Headquarters, the operation was managed by a heavy country-based structure which did not provide for a post of sufficient seniority to ensure effective coordination of operations at a regional level. At field level, the Special Envoy appointed by the High Commissioner had neither the authority nor the staff to ensure a coordinated regional approach. These weaknesses were compounded by poor communications, weak information management and an excess of reporting and decision-making layers combined with a lack of clearly designated focal points for crucial areas such as logistics and staffing.

15. Another problem which had also surfaced on repeated occasions in previous emergencies concerned the relationship between the Regional Bureau and the Division of Programmes and Operational Support. The UNHCR emergency response capacity had been built-up as a result of initiatives taken by DPOS and the ERTs were principally staffed with and led by personnel from the Division, in particular EPRS, PTSS and STS. The integration of emergency staff into existing bureau and branch office structures has never been straightforward and led to very acrimonious problems at the height of the Rwanda emergency. While the situation approach which forms the central thrust of the Delphi process is likely to obviate some of these managerial difficulties, there is a need to give further thought to the issue of emergency management as a whole, both at Headquarters and in the field. This would seem to be all the more important in the context of the Change Management initiative.

16. In particular, the role of EPRS, PTSS and STS in future emergencies has to be clarified. In doing so, the proven competence and experience of staff in these sections must be recognised and they should, on that basis, share in the responsibility of handling the operational aspects of an emergency, beginning with the contingency planning stage. This cannot happen without the creation of an authority structure at Headquarters which will enable the sections concerned to work together effectively, and in close and harmonious cooperation with the regional bureau. A number of possible structures could be considered, including the establishment, as early as possible in a refugee emergency (and preferably before), of a Deputy Situation Manager (D1) at Headquarters, who would report directly to the Situation Manager in the field, and have line authority as director of operations in the event on an emergency. Alternatively, this role could be played by a Deputy Director in DOS. In either case an essential element would be the creation, at field level, of a clear chain of command, with unambiguous roles and responsibilities.

17. In view of the need for further discussion on these and related issues it is proposed that this might most appropriately take place in a workshop on Emergency Management as recently proposed by EPRS. It is strongly recommended that this workshop should take place at Headquarters within the next three months to permit the important lessons learned from the Great Lakes emergencies to be reflected appropriately in a strengthened protection and management response.

Summary of Recommendations

Recommendation 1

Ensure that the protection aspect of UNHCR's emergency response capacity is greatly strengthened by ensuring high level decision-making on protection policy at the outset of emergencies and strengthening of Protection Operations Support Section so as to produce an enhanced stand-by capacity of senior protection staff with emergency training.

Action: DIP, with support from EPRS

Recommendation 2

Examine, in the light of the experience gained in Cambodia, Rwanda and the former Yugoslavia, the options available to UNHCR in order to deal with a refugee leadership which may be excluded from refugee status under the Mandate of the Office on account of crimes against humanity and to obtain effective support from the international community in situations of this nature.

Action: DIP

Recommendation 3

Develop and adopt for the purpose of emergency preparedness and contingency planning, a range of policy options and measures which may be adopted swiftly to improve camp security in emergencies.

Action: DIP with support from FSSS and EPRS

Recommendation 4

To ensure that appropriate attention is paid to the specific protection and assistance needs of refugee women and children, all ERT members should receive prior training in designing assistance programmes that provide maximum protection of the rights of women and children as specified in the appropriate guidelines; such training and guidelines should also be provided to government counterparts and implementing partners.

Action: EPRS with support from DIP, PTSS and the coordinators on refugee women and refugee children

Recommendation 5

A range of options should be developed, for adoption at the earliest stages of an emergency, to strengthen community solidarity among the refugees and to integrate adolescents, in particular young males, into constructive activities.

Action: PTSS

Recommendation 6

A roster of anthropologists and other country/region specific experts should be developed, both in-house and elsewhere, with specialist knowledge of emergency prone areas, for provision of briefings on the cultural and political background of the target group, to participate in the production of Emergency Preparedness Profiles of emergency prone countries, and, if the need arises, to be deployed as part of an ERT.

Action: CDR

Recommendation 7

In major refugee emergencies UNHCR, in cooperation with DHA, should seek to assume lead agency status with respect to both external and internal population displacements resulting from the same sets of causes so as to be in a position to develop and implement a comprehensive strategy for assistance, protection and solutions.

Action: Executive Office

Recommendation 8

The early warning study recently initiated by CDR should seek to clarify UNHCR's position on early warning, make recommendations on the development of an in-house capacity, and define the parameters for cooperation with the integrated humanitarian early warning system recommended by the OECD/DAC report.

Action: CDR with support from EPRS

Recommendation 9

A number of steps are required to ensure proper handling of logistics in a large emergency, including; the regular review of stand-by agreements on logistics; the preparation of logistics profiles of emergency-prone countries as part of contingency planning; the deployment with the ERT of a sufficient number of logistics staff and the finalisation of the SIMS Commodity Tracking System for use in emergencies.

Action: STS with support from EPRS

Recommendation 10

A framework for regular dialogue between EPRS, Telecommunications Unit and the Desks should be established in order to review technical developments; to establish telecommunication guidelines for emergencies for inclusion in the Emergency Handbook; to assess possible needs in emergency prone areas and to establish a telecommunications strategy.

Action: Telecommunications Unit with support from EPRS

Recommendation 11

All possible measures must be taken to ensure effective collaboration between UNHCR and WFP in an emergency: this should include joint high level food assessment missions at the earliest moment leading to joint agreement on rations and needs; joint liaison with donor representatives in the field and at Headquarters; the inclusion in the ERT of a nutritionist to monitor the nutritional status of the refugees and advise on rations and the composition of the food basket, and the inclusion of a Food Aid Officer in the ERT to act as a permanent interlocutor with WFP and deal with all matters of joint concern for the two agencies.

Action: EPRS with support from PTSS and FSU/PCBS

Recommendation 12

UNHCR's creditable record of technical coordination in emergencies should be maintained and further enhanced. This will require a clear institutional commitment with focus on a range of measures including: the continued deployment of technical staff in key sectors as part of the ERT, remaining for at least 6-8 weeks to set up and hand over workable systems; the coordination of sectoral activities in major emergencies by UNHCR technical staff and consultants, and regional networking as well as ongoing coordination between PTSS and technical staff in the field.

Action: PTSS with support from EPRS

Recommendation 13

A formal framework should be established for regular consultations between PTSS, EPRS, STS and Telecommunications Unit to review contingency plans and emergency preparedness in technical sectors.

Action: EPRS with support from PTSS, STS and Telecommunications Unit

Recommendation 14

At the contingency planning stage, host governments must be strongly encouraged to take note of technical and environmental considerations in site selection. The involvement of technical experts in contingency planning will serve to reinforce this message.

Action:PTSS with support from EPRS and coordinator on environment

Recommendation 15

A comparative study of local salary practices in emergencies should be conducted, to establish a set of basic guidelines and to integrate salary policies into contingency planning.

Action: DHRM

Recommendation 16

There is a need for additional multifaceted efforts to enhance UNHCR's capacity to respond to emergency staffing needs; this should include, the development of crash training programmes for new staff, consultants and implementing partners; career opportunities and rewards for General Service staff who perform exceptionally well in emergencies, and additional incentives to induce senior staff to serve in emergencies.

Action: DHRM

Recommendation 17

The proposed expanded role of EPROs should be discussed at the Workshop on Emergency Management referred to in the final paragraph of the Overview.

Action: EPRS

Recommendation 18

In view of the management problems detailed in this report, a workshop should be convened to examine the functioning of emergency management within the new organisational structure which is being set up as a result of the Delphi process.

Action: DOS

II. BACKGROUND

Refugee movements in the Great Lakes region

18. The conflict between the Tutsi and Hutu communities in Rwanda and Burundi has brought about one of the most long-standing and intractable refugee problems in Africa. Despite repeated international mediation efforts, including nearly four decades of refugee assistance by UNHCR, the situation exploded into a series of dramatic emergencies in 1993/4 and continues to defy solution to this day. In both countries, the latent tension between a politically dominant minority, the Tutsis, and a relatively disenfranchised majority, the Hutus, was exacerbated by the colonial powers which adopted discriminatory policies that deepened the division between the two groups. With the coming of independence in 1962, a power struggle ensued whereby events in one country had serious and often detrimental effects upon the other.

19. In Rwanda, a Hutu social revolution begun in 1959 led to the flight of some 480,000 Tutsis to neighbouring countries, including Burundi, Zaire, Uganda and Tanzania. In the early sixties their repeated attempts to return home by force of arms provoked ethnic violence, massacres and further refugee flows. In Uganda, many of these refugees joined the armed forces and later provided the military core of the RPF (Rwanda Patriotic Front) which invaded Northern Rwanda in 1990 in a further and eventually successful attempt to enforce their return.

20. While post-independence Rwanda was thus ruled by Hutus and drove mostly Tutsis into exile, the opposite process took place in Burundi where a Tutsi oligarchy secured complete control over the country soon after independence. A Hutu uprising in 1972 was crushed with great loss of life and led to the flight of 120,000 Hutu refugees, mostly to Tanzania whose generous offer of citizenship to these refugees must not go unmentioned. While sporadic ethnic clashes continued, the 1980s also saw the beginnings of a democratic process which led to the creation in 1991 of a repatriation commission by the Government. By mid-1993, international and internal efforts to find a lasting solution to the crisis in the two countries seemed to be bearing fruit. In June, a Hutu President, Melchior Ndadaye, was democratically elected in Burundi and a programme of national reconciliation adopted. In August, a peace agreement signed at Arusha provided for the creation of a democratic process of government in Rwanda and the peaceful return of the refugees.

21. Extremist elements on both sides of the divide were, however, opposed to any compromise. The social tensions in both countries had been worsened by a severe economic downturn and increasing pressure on scarce land and resources. This was already creating problems for the repatriation of refugees to Burundi which had just begun under UNHCR auspices. The situation exploded with the assassination of the Burundian President Ndadaye in October 1993 by elements of the Tutsi army which was followed by massive reprisals and the flight of some 700,000 Hutu refugees, mostly to Tanzania and Rwanda. UNHCR's response to this became known as the "Burundi emergency" which is discussed in this paper.

22. There is little doubt that the assassination of the Hutu President of Burundi strengthened the extremist Hutu elements in Rwanda in their determination to undermine the Arusha Peace Agreement and lent added credence to their violent anti-Tutsi propaganda. To prevent the implementation of the Agreement, President Habyarimana was killed in an attack on his plane on 6 April 1994 and a meticulous plan to exterminate the Tutsi population of Rwanda as well as moderate Hutus was carried out in the ensuing three months. The loss of at least 500,000 lives in the genocide did not prevent a rapid conquest of the country by the RPF. As a result, some 1.2 million Hutu refugees fled to Tanzania (Ngara) and Zaire (Goma, Bukavu and Uvira) in front of the advancing RPF forces.

23. The Burundi and Rwanda emergencies are examined here together since they took place in quick succession in the same theatre of operations and the lessons learned from both operations are related. However, there are also crucial differences. The Burundi emergency was a classic refugee situation comprising mostly civilians who fled spontaneously and their plight attracted little international attention. The Rwanda emergency, on the other hand, involved a strategic retreat by the former Government of Rwanda and its armed forces following a genocide. The tragic events in Rwanda, the sudden and massive nature of the exodus and the cholera epidemic which afflicted the refugees in Goma attracted world-wide publicity, and a huge international aid effort was set in motion.

The UNHCR Response

24. Following the problems experienced during the Gulf War in 1990-1991, UNHCR took a range of measures to strengthen its emergency response capacity. The resulting package, summarised in the Catalogue of Emergency Response Resources, is based upon the following key elements:

• UNHCR internal emergency staffing, consisting of Emergency Preparedness and Response Officers (EPROs) backed up by Senior Emergency Administrators (SEAs) and Emergency Finance and Administrative Assistants (EFAAs) and an emergency roster of UNHCR staff world-wide; these form the core of the Emergency Response Team (ERT) and have the task of opening UNHCR Offices and initiating the operations;

• participation in the ERT of technical experts from PTSS, STS and Telecommunications to assume sectoral responsibilities and establish a logistics and communications network;

• stand-by agreements with other agencies, including Nordic NGOs, UNVs and the State Committee of the Russian Federation for Civil Defense, Emergencies and the Elimination of the Consequences of Natural Disasters (EMERCOM) to provide emergency staffing and/or additional technical and logistical support to the ERT; and

• stand-by arrangements for the rapid provision of administrative and operational resources (vehicles, relief goods etc.) to support the ERT.

25. As will be shown in the following pages, this response system proved to be vital in dealing with the Rwanda and Burundi emergencies. A large number of ERTs were deployed to the region and their mission reports, in as far as it has been possible to trace them, proved to be a particularly valuable source for lessons learned purposes. The table which appears below indicates the locations to which these ERTs were deployed and gives information on the duration of their assignment, the number of refugees involved and the reports consulted.

The lessons learned process

26. Shaken by the scale of the Rwanda emergency, the Executive Committee in 1994 requested UNHCR to conduct a comprehensive lessons learned exercise involving other UN agencies, governments and implementing partners. Almost simultaneously, a similar system-wide study was proposed by members of the OECD/DAC. After further consideration, the SCAF decided in December 1994 that UNHCR should limit itself to an internal lessons learned exercise to be conducted in parallel to the larger OECD/DAC Study.

27. When OECD/DAC launched its Joint Evaluation in February 1995 the Deputy High Commissioner issued a statement explaining the importance of the exercise and calling on staff concerned to give it their full cooperation (IOM/15/FOM/14/95 of 9/2/95). However, there appears to have been communication problems early on with the team conducting the crucial Study Three on Humanitarian Aid and Effects. Having failed to interview any of the key EPRS staff who had been deployed to the region between April and July 1994, the team's findings contained errors and omissions which were detrimental to UNHCR. As a result, a great deal of time and effort had to be expended later in the year to obtain a more balanced description of events.

28. The document as finally published is generally complimentary about the work of UNHCR and comments repeatedly on the "impressive performance" of the ERTs. The coordinating role played by UNHCR's technical experts is singled out for particular praise (see Study Three, pp.16, 17, 31, 38 etc.). Most of the findings of the OECD/DAC Study concern the need for a more coherent inter-agency approach to humanitarian emergencies and the related recommendations are currently the subject of inter-agency consultations. However the study also contains a number of critical observations including recommendations of direct relevance to the work of UNHCR. As noted in the following pages, while many of the points they raise have since been addressed, others still require attention.

29. The UNHCR internal lessons learned process began with the preparation by the former Central Evaluation Section of background documentation for a major lessons learned workshop to be held in April 1995. A detailed Annotated Agenda was produced on the basis of interviews conducted with numerous staff-members. A background paper on Recurring Lessons from Previous Emergencies was prepared by J. Telford (Annex A) while Y.Stevens contributed a discussion paper on Lesson Learned from the Rwanda Emergency on Management and Coordination of Technical Sectors (Annex B). Additional material included reviews of the histories of Rwanda and Burundi and a detailed comparative chronology of events during the emergency.

30. Considering this large preparatory effort it is unfortunate that the workshop had to be postponed and later cancelled at short notice as the continuing crisis in the region made it impossible for all key players to be withdrawn from the field at the same time. The cancellation meant that the UNHCR lessons learned process became rather protracted and that the Office was not as well equipped as it should have been to contribute to the findings of the OECD/DAC evaluation. All that could be organised in 1995 was a short meeting with a selected number of field staff which took place in Nairobi in May. The material prepared for the aborted evaluation workshop and the discussions at the Nairobi meeting provided the basis for the Conference Room Paper on lessons learned from the Rwanda emergency which was presented to SCAF in June 1995 (see Annex C).

31. Following the publication in spring 1996 of the OECD/DAC Report, a two-day workshop was convened in Geneva in May to discuss the implications of its findings for UNHCR's activities and prepare a follow-up paper on the subject which was submitted to the Standing Committee in June 1996 (Annex D).

32. The present document represents the conclusion of the UNHCR internal process. It aims to: (a) identify the main lessons learned from the deployment of emergency response teams to the Great Lakes region between October 1993 and December 1995; (b) specify action taken since 1994 to improve UNHCR's emergency response capacity in the light of the experience gained in the Great Lakes; and (c) identify areas in which further action is required and make concrete recommendations to that effect. The document is based upon the following sources: the reports of the ERTs deployed to the region; the material prepared for the aborted 1995 lessons learned workshop; Y.Stevens' paper on technical sectors and the overview of J. Telford's paper on recurring lessons; the OECD/DAC Joint Evaluation of the Rwanda Emergency; the findings of the May 1996 workshop on the OECD/DAC recommendations of relevance to the work of UNHCR; and interviews with staff members who were involved in the operations. A considerable number of other relevant documents were consulted and a full list of sources is provided in Annex G.

III. PROTECTION AND POLICY ISSUES

Insufficient protection coverage

33. J. Telford's paper on Recurrent Lessons in Previous Emergencies notes that "protection in emergencies has frequently been the subject of debate and criticism" due to lack of strategies, policy guidance or response capacity (Annex A, p.2). This apparent weakness in the protection domain came very much to the fore in the Rwanda operation. As documented in several ERT reports, there was a lack of sufficient senior protection staff in virtually all areas which hampered the ability of the ERTs to address a range of serious protection issues in a proper manner. The ERT leader in Ngara reported that "we made a mistake in not sitting down and discussing the role of protection at an early stage. One Protection Officer is insufficient in an emergency of this size" (M.Connelly). In Goma, the lack of a Senior Protection Officer "was the most serious delay in an otherwise efficient staff deployment and it had direct consequences on the definition and implementation of a protection strategy" (F.Grandi's report). Similar problems were also encountered in Uganda: "there was a series of protection issues which would have been more efficiently handled if that team had counted with the presence of a Protection Officer throughout" (E.Demant).

34. The weakness in this area was referred to in the Conference Room paper on Lessons learned from the Rwanda Emergency presented to SCAF in June 1995 which concluded that "efforts should continue to ensure the integration of protection activities within the overall UNHCR response to emergencies" (Annex C para. 14). The only follow-up action undertaken so far is the appointment within EPRS of one permanent staff-member with protection background. This is clearly insufficient, not least because EPROs are not intended to assume sectoral functions but are expected to act as team leaders and managers. Additional measures are required; these should include the strengthening of the Protection Operations Support Section (POSS) in the Division of International Protection in terms of staff and financial resources so as to enable it to respond more effectively to emergency situations. It could thus serve as a protection bridgehead for EPRS by providing support in defining and coordinating protection input and ensuring subsequent monitoring.

35. The lack of sufficient protection staff in the field was, however, only one aspect of a more fundamental problem. What was especially lacking in the Rwandan situation was an early assessment of the nature of a refugee caseload which was far from straightforward from the point of view of UNHCR's mandate as discussed in more detail below. The lack of such an assessment meant that the conditions for UNHCR's involvement in the assistance operation were not clearly defined from the outset and protection considerations were not given sufficient weight in the formulation of an overall policy. While UNHCR has made outstanding progress in many technical areas of emergency preparedness and has specially trained administrative staff on stand-by for emergency service, it is a matter of serious concern that no comparable effort has been made to bring about an effective strategic response to protection concerns. Technical expertise in any other field can, in theory, be provided by outside agencies, but only UNHCR is competent to provide international protection.

Recommendation 1

Ensure that the protection aspect of UNHCR's emergency response capacity is greatly strengthened; this must involve:

• high-level decision-making on protection policy at the outset of emergencies to establish an appropriate strategy with effective monitoring and evaluation arrangements;

• strengthening of POSS so as to produce an enhanced stand-by capacity of senior protection staff with emergency training and ability to work in different languages; and

• preparation of a protection crash course package to train implementing partners and new staff members deployed in emergencies

A leadership implicated in genocide

36. From the point of view of UNHCR's mandate, the refugees from Burundi and Rwanda were subject to group status recognition under Article 1, paragraph 2 of the OAU Convention which grants refugee status to persons having fled their country as a result of events seriously affecting public order. While this presented no problems with the Burundians, the situation with the Rwandese was different as many of them, in particular their leaders, had unquestionably instigated, or directly participated in, the genocide and were thus subject to the exclusion clause. The dilemma this presented for UNHCR was perceived early on by the UNHCR Representative in Tanzania who raised the issue of the presence of suspected perpetrators of atrocities in Rwandan refugee camps with Headquarters already in May 1994.

37. In their Position Paper on Protection Issues produced in October 1994, S.Lombardo and K.Paul described the circumstances in Zaire as follows:

"With the massive influx of Rwandans into Zaire we have witnessed not only a movement of people but also a transplantation of a very well organised political, social and security structure into another country. This structure continues to function and is thereby controlling the daily life in the camps. Each day, with the consolidation of the assistance programme this Rwandan political structure is growing stronger."

The authors of the paper go on to note that UNHCR was in danger of facing "a moral and ethical dilemma" by extending its protection function to elements of the population that may be responsible for genocide. They called for the urgent deployment of an international force to separate suspected perpetrators from the rest of the population. Otherwise "UNHCR should seriously consider and be prepared to pull out". In such an eventuality it was recommended that UNHCR should promote the establishment of a Border Relief Operation modelled on UNBRO with no protection role but only relief and assistance functions.

38. UNHCR Headquarters for its part did not decide on such a drastic course of action. Its policy on the matter was summarised in a paper presented to the High Commissioner by the Director of International Protection on 18 November 1994 and was based on the following considerations: (a) that UNHCR was "deeply concerned" at the presence of suspected perpetrators of "war crimes or grave violations of international humanitarian law" among the refugees; however, neither UNHCR nor the countries of asylum were able to exclude them from refugee status on account of "serious security constraints"; and (b) that UNHCR was ready to cooperate with the activities of the human rights machinery of the United Nations, including the International Tribunal and the human rights monitors and investigators dispatched to the field, however, the primary responsibility to comply with requests issued by the Tribunal lay with States. The paper concluded with the operative statement that "the mandatory function entrusted to UNHCR by the international community in assisting host countries to cope with refugee influxes further justifies our continued presence even under the most trying circumstances".

39. Considering the fact that this was the policy of the house, the work of the ERTs in the field can in no way be faulted. What they were attempting to do, as documented in their reports, was to implement a successful emergency response "in the most trying circumstances", notwithstanding the fact that many segments of the refugee leadership were undoubtedly implicated in crimes against humanity. One of the consequences of this was, however, that the food and commodity distribution system initially set up in the camps relied upon that very leadership and strengthened its control over the refugees.

40. The September 1994 ERT report from Goma is the only one by an ERT leader which makes an issue of this fact. It states that "one of the most excruciating moral dilemmas faced by UNHCR in Goma has been the contradiction between need to assist and assistance to 'bad' leaders" (F. Grandi's report p.12). The report goes on to present a number of practical arguments in favour of providing assistance which concur with the overall policy adopted by the Office. In particular, the report points out the impossibility for UNHCR to identify those leaders who may fall under the exclusion clause and the dangers involved in any such attempt considering the volatile security situation in the camps. In this context it should mentioned that the Tanzanian authorities' move to arrest a suspect former bourgemestre in Ngara earlier in the year had to be abandoned after it sparked a riot among the refugees.

41. Two years later, the situation on the ground gives rise to disquieting thoughts. Despite repeated requests, no international force has materialised to separate suspect elements from the refugee camps and less than a handful of persons have been apprehended for trial by the International Tribunal. The "Rwandan political structure", on the other hand, has been greatly strengthened by a continuing assistance programme and remains firmly in charge, actively discouraging repatriation and, most likely, involved in supporting cross-border raids into both Rwanda and Burundi. Indeed, recent reports appear to indicate that Interahamwe elements are currently infiltrating into Rwanda in an attempt to eliminate remaining witnesses of the genocide.

42. Despite the resounding success of UNHCR's efforts to provide emergency relief to the Rwandan refugee population, the long term political consequences of the approach taken by the international community are profoundly disturbing. The entrenchment of an extremist leadership has been strengthened, their constituency - the refugees - are maintained by international aid and a peaceful solution to the problem appears more remote than ever. There must be a lesson to be learned from this. Indeed, a number of fundamental questions which are beyond the scope of this report, remain to be answered. Did UNHCR in fact have a choice? Should the Rwandan refugees have been granted group status recognition despite the enormity of the genocide which had been set in motion by their leaders? Could UNHCR have promoted a firmer approach by the international community? Should UNHCR have opted for a policy of complete or partial withdrawal after the provision of emergency assistance in the initial stages?

43. One of the regrettable consequences of the cancellation on the lack of the internal UNHCR lessons learned workshop in Spring 1995 is that these issues have still not been collectively examined by the staff members who were originally involved in the Rwandan operation. There remains a need for further thought on how UNHCR and the international community as a whole should deal with a refugee leadership which may be subject to the exclusion clause. In this respect, the Rwandan case is not unique: other cases of genocide have occurred in recent history and may yet occur again.

Recommendation 2

Examine, in the light of the experience gained in Cambodia, Rwanda and the former Yugoslavia, the options available to UNHCR in order to:

• deal with a refugee leadership which may be excluded from refugee status under the Mandate of the Office on account of crimes against humanity; and

• obtain effective support from the international community in situations of this nature.

Need for enhanced security in emergencies

44. In the camps of Eastern Zaire UNHCR had "no possibility to carry out its protection activities due to the lack of authority and power of the local Zairian authorities" (ERT report N.K. Mbaidjol). The local security forces were poorly led and unpaid, which conspired to create an atmosphere of corruption, extortion and anarchy. The problem had already been noted by the first ERT in Goma when various arrangements were tried out unsuccessfully "until a system of direct payment of daily allowances to the military was implemented" (ERT report F. Groot). Similar difficulties also surfaced in Tanzania during the Burundi emergency where "security provided in the camps by the local authorities was non-existent" as staff were not being paid their allowances (ERT report N. Coussidis).

45. Serious security problems caused by the weakness or absence of government law enforcement agents have been a feature in several recent emergencies. In such situations the debate has often arisen whether and to what extent UNHCR should engage in paying the local security forces for their work and whether international forces should become involved in camp security arrangements. In Eastern Zaire, the lack of early agreement on such measures led to a long period of anarchy which put the lives of both aid workers and innocent refugees at risk.

46. In the light of these problems, the OECD/DAC study recommends in finding C-1 that UN peace missions should be given the authority and means to ensure the protection of camp populations and relief staff. Experience has shown, however, that such peace missions are costly and slow to mobilise. A pragmatic solution was found with the Zairian Camp Security Contingent (ZCSC), an elite force from the presidential guard financed by UNHCR and trained by a small international police force. The Inspection and Evaluation Service is planning an evaluation of the work of this contingent.

47. In order to be better prepared to deal with security problems in future situations, the issue should be studied as a matter of principle so as to reach agreement on a range of measures which may be adopted without a long and time consuming debate surfacing anew in every case and resulting in potentially life-threatening delays. Ensuring adequate security arrangements should be an integral part of contingency planning.

Recommendation 3

On the basis of the findings of the IES study develop and adopt, for the purpose of emergency preparedness and contingency planning, a range of policy options and measures which may be adopted swiftly to improve camp security in emergencies. These may include:

• guidelines on reinforcing local capacity including payment of local security forces (bearing in mind previous experience);

• stand-by arrangements with outside agencies such as the international police force which was running the ZCSC;

• deployment of Security Officers as part of the ERT; and

• coordination with DPA/DPKO to ensure system-wide approach.

Protecting women and children, and catering for the needs of adolescents

48. The speed of the exodus, the large numbers involved and the overcrowded and chaotic conditions in many of the camps had dire consequences, especially for women and children which were aggravated by the lack of security arrangements. With regard to women, the problems described in the ERT report by N.K.Mbaidjol (for Zaire) and a memo by D.Goodman (for Tanzania) include marginalisation during food distribution, fire wood and water collection, physical abuse and sexual exploitation as well as discrimination in the day-to-day running of the camps.

49. The Rwanda experience shows that in a large-scale emergency, the structure of the assistance programme should from the outset be designed in such a way as to promote the equal access for male and female refugees to protection and assistance and minimise the opportunities for the marginalisation of women. D.Goodman's memo recommends a number of concrete measures to that effect based upon the experience in Ngara. These include the establishment of separate distribution points for men and women, the recruitment of equal numbers of female and male guardians and guardians chiefs and the appointment of specially trained counsellors for refugees who have been victims of sexual violence. Vital preconditions for measures of this kind are close cooperation between protection and community services staff , field workers and the refugees themselves and the timely organisation of protection training courses for implementing partners and government counterparts.

50. The extent of the problems involving children is illustrated by the number of unaccompanied minors identified during the Rwanda-Burundi emergency. With 117,000, it is the highest on record during recent refugee situations. The lessons learned from dealing with their problems have been carefully studied and have resulted in a number of specific actions including a Memorandum of Understanding with UNICEF and the development of a emergency registration kit. A summary is provided in a paper by the former Senior Coordinator for Refugee Children Unaccompanied Refugee Children - Lessons learned and some Strategies for the Future (A. Skadvedt April 1996).

51. What measures for women and children have in common is that a) all necessary information and guidelines to establish them in accordance with best practice exist and b) in order to be effective, the measures have to be set up from the beginning of the emergency. As stated by A.Skatvedt, "if we can establish good systems for assistance and protection of unaccompanied children during the very first days of a new emergency, suffering - and costs - can be reduced" - and, indeed, lives might be saved.

52. Measures for women and children do have implications for all assistance sectors, starting with camp design, and the responsibility for putting them in place should therefore not be left to Protection and Community Services Officers alone. All members of an ERT should be equally conversant in this field, and adequate training for implementing partners and government counterparts should be organised as a matter of priority. This has also been one of the findings of the recent review of the Women Victims of Violence Project in Kenya.

Recommendation 4

In accordance with the recommendations of the IES' Review of the Women Victims of Violence Project in Kenya, it should be ensured that:

• all ERT members receive prior training in designing assistance programmes that provide maximum protection of the rights of women and children as specified in the appropriate guidelines; and

• staffing resources and training materials are available to provide appropriate training of government counterparts and implementing partners.

53. While the protection of women and children requires careful attention, their plight should not be looked upon in isolation. Abuse by men may take place because family and community structures have been destroyed or because young males are condemned to a life of inactivity and frustration in camps. In this respect, the Rädda Barnen report on Community Participation in Rwandan camps in Tanzania notes among the lessons learned that the "situation of adolescents was not given enough attention in the early emergency" (p.28) and stresses the need for targeted self-help programmes for this group. The report also notes that good camp organisation at the start of the emergency can be vital to maintaining community solidarity and co-operation. This finding was also reflected in the IES study of the Women Victims of Violence Project in Kenya.

Recommendation 5

Develop a range of options which can be adopted from the early stages of the emergency to strengthen community solidarity among the refugees and integrate adolescents, in particular young males, into constructive activities.

54. Further useful recommendations on measures to address the needs of refugee women and children in emergencies may be found in the Joint Evaluation of the Agreement between Rädda Barnen and UNHCR, prepared in 1995.

Protection and Cultural Awareness

55. For purposes of both protection and assistance, a good understanding of the nature of refugee society is of vital importance. This is hard to develop during the early stages of an emergency when staff are under pressure and numerous practical decisions have to be taken. Without such an understanding, however, these very decisions may be ill-informed and have undesirable consequences. Open communication with refugees may be seriously impaired.

56. Despite UNHCR's long presence in Rwanda before the emergency, this was certainly an issue. The OECD/DAC study found that "relief agencies had only a very limited understanding of Rwandese society" and "very little account had been taken of the views of beneficiaries" (Study 3, p.143). This may have delayed the vital switch from commune to cellule level for food distribution and rendered the search for credible interlocutors among the refugees more difficult. The above-mentioned Rädda Barnen report also raises the issue and states that emergency team members "need assistance with as much background information as possible ... both before and during deployment" (p.27).

57. When this matter was discussed at the UNHCR workshop in May 1996, there was general agreement that in complex situations "we need help to understand our clients" and expertise may be required at an early stage. Where this expertise does not exist in-house, anthropologists and other regional experts with specialist knowledge should be consulted for purposes of orientation and briefing. This, however, requires advance preparation and should become an integral part of contingency planning for major emergencies.

Recommendation 6

Develop a roster of anthropologists and other country/region specific experts, both in-house and elsewhere, with specialist knowledge of emergency prone areas who can be:

• consulted for purposes of cultural background briefing;

• requested to participate in the production of Emergency Preparedness Profiles of emergency prone countries; and

• be deployed as part of an ERT if the need arises.

Voluntary Repatriation

58. During 1994, at least three different types of voluntary repatriation were taking place in the Great Lakes region: the largely spontaneous return of the Burundians who had fled during the 1993 emergency; the return of the "old caseload" of Rwandese Tutsi which took place in the wake of the RPF victory; and the sporadic return of some Hutu refugees who had fled Rwanda during the 1994 emergency.

59. Information of use for this paper could only be obtained on the last of these three groups. This concerns in particular the repatriation of the Hutu refugees in Zaire which was, and still is, a very complex problem. UNHCR's policy towards the voluntary return of this group went through two distinct stages during 1994. From the beginning of the exodus in July until the end of September, the Office was convinced that a massive and rapid return to Rwanda was a distinct possibility. This seemed to be justified by the rapid return of large numbers of refugees from Goma in the early days of the emergency. As stated in a Conference Room Paper issued by UNHCR for the UN/DHA Pledging Conference on 2 August 1994, repatriation to Rwanda "could in itself become an essential part of the process of national reconciliation and peace rather than the result of it".

60. With the benefit of hindsight, it must be said that this initial optimism was misplaced for at least two related reasons: (a) there was insufficient understanding of the specific circumstances created by the genocide in Rwanda, including the largely justified fear of most refugees that they would be treated as perpetrators upon their return, and the desire for retribution on the part of the survivors and the RPF forces. In fact, the chasm between Hutus and Tutsis had never been deeper and a sudden "national reconciliation" was therefore highly unlikely, especially as the Hutu leadership which was implicated in the genocide exerted a high degree of control over the refugees; and (b) UNHCR had not sufficiently examined to what extent the security situation in Rwanda was in fact compatible with a massive return of the refugees before deciding upon its policy towards repatriation; the latter was determined more by the "untenable situation" in the countries of asylum rather than by conditions in the country of origin. When a survey of conditions in Rwanda was conducted in August 1994, there was found to be a "pattern of abuse by the new Rwandan army" and all repatriation arrangements had to be suspended as of 28 September 1994 (telex HCR/BDI/1481 of that date).

61. UNHCR's early misjudgement of the situation was probably brought about by factors already noted above: insufficient attention to protection considerations combined with a lack of real understanding of the socio-cultural circumstances of the refugees and the situation in Rwanda. Similar observations are also made in the OECD/DAC report which concludes that "the view in some quarters of the international community that major repatriation and reintegration could occur quickly was clearly unrealistic and mistaken" (Synthesis Report, p.64). This criticism, however, is not just aimed at UNHCR. Other actors, in particular UNAMIR, were far less cautious in their early advocacy of repatriation as the quick solution.

The "Protection Gap": the Internally Displaced

62. While this paper is concerned with UNHCR's emergency response in the countries of asylum, it is not complete without a mention of the associated problem of the internally displaced in Rwanda. UNHCR's approach to this issue went through two distinct phases. In early 1993, following a request from President Habyarimana for assistance to persons displaced by the armed conflict in Rwanda, UNHCR concluded that any direct intervention on its part would have to be linked to a wider international effort to resolve the root causes of the problem. It was, therefore, decided to make a limited financial contribution to the Government with a view to preventing refugee flows into neighbouring countries.

63. Following the genocide of April-June 1994, UNHCR was obliged to address the problem of IDPs in Rwanda in dramatically changed circumstances. The Office then took the position that it should focus on the development of a broad operational framework encompassing all aspects of violence-induced displacement resulting from the same set of causes rather than subscribing to a potentially discriminatory delineation of competences and activities based on specific categories of persons. While UNHCR was thus disposed to become more closely involved in IDP assistance in Rwanda after June 1994, however, it was to be only one player among several other international agencies involved and was not in any way to assume a leading or coordinating function.

64. As strikingly analysed in the OECD/DAC report, a major problem with UN operations in Rwanda was that no agency proved capable of assuming such a responsibility. In fact, the Great Lakes operation as a whole was characterised, on the one hand, by a well-funded and well-coordinated assistance operation spearheaded by UNHCR on behalf of refugees in the countries of asylum and, on the other, by an underfunded and ill coordinated response to the problems of IDPs and other persons in need in the country of origin. The lack in the UN operation of a central leadership function, with ability to develop and implement an overall strategy that bridged the gap between countries of asylum and country of origin, caused the OECD/DAC evaluators to use the term "hollow core" to describe an organisational centre which was "weak, poorly resourced and lacking in clarity" (Study 3, p.133). This realisation led to a series of OECD/DAC recommendations which aim at strengthening DHA so as to enable it to assume overall control of large humanitarian relief operations and thus fill the "hollow core".

65. One element at the heart of this issue is the internally displaced and the obvious lack of an international agency with a mandate to address the protection and assistance needs of this group. In the Rwandan case, this had at least two major negative consequences for the work of UNHCR. Firstly, the lack of such an agency meant that no international body had responsibility for monitoring the movements of the internally displaced in the crucial days before the Goma exodus; the lack of advance warning of the massive influx that occurred is at least partially attributable to this; had UNHCR assumed a larger share of responsibility for IDPs in Rwanda already in 1993, it might have thus taken more effective precautionary measures. Secondly, in the case of the Kibeho incident, succinctly analysed in a recent study with the apt title "The Protection Gap", the lack of a coherent international response to IDPs had tragic consequences when numerous inhabitants of an internally displaced persons camp were gunned down by members of the Rwandan armed forces (for reference see Annex G). The impact upon the region was dramatic as UNHCR-sponsored repatriations practically ceased for a long time thereafter.

66. These events illustrate that in major international refugee crises, it would seem to be very much in the interests of UNHCR to assume the status of lead agency not only for refugees but also for IDPs; the Office certainly has the experience and credibility required for such a task. This line of thinking was in fact advocated at the time of the Rwanda crisis when DIP proposed that "UNHCR should consider both external and internal population displacements resulting from the same sets of causes and integrate both sides of the problem into a comprehensive strategy for assistance, protection and solutions" (P. Bertrand, memo to the Deputy High Commissioner, 12 July 1994). The best illustration that such an approach can be very effective indeed is the Tajikistan operation where "IDPs were the first to benefit from UNHCR's assistance, before any refugee had come back" and a climate of confidence was created which was vital to the success of the entire programme (Tajikistan, Lessons Learned from a Country of Origin Operation, p.iii). The implementation of an integrated strategy for both population groups, however, is only feasible in practice when UNHCR is officially assigned the authority and the leadership role to coordinate operations in the countries of asylum and the country of origin. Such was not the case in the Great Lakes emergency.

Recommendation 7

In major refugee emergencies, UNHCR in cooperation with DHA should seek to assume lead agency status with respect to both external and internal population displacements resulting from the same sets of causes so as to be in a position to develop and implement a comprehensive strategy for assistance, protection and solutions.

IV. OPERATIONAL ISSUES

Early Warning

67. The massive influx of refugees into Goma took UNHCR and most other international relief agencies by surprise. Noting the lack of a mechanism to provide advance warning of major population movements in crisis areas, the OECD/DAC study recommended the development of an integrated humanitarian early warning system (Finding C-2). When the topic was raised at the UNHCR workshop in May 1996, participants felt that while the Office should support any relevant inter-agency initiatives, the experience to date was not encouraging; UNHCR should therefore endeavour to build up an in-house early warning function which at present does not exist. It had originally been planned that EPRS should assume this responsibility but this had been overtaken by other priorities.

68. UNHCR's current approach to early warning is far from clear. CDR has developed an impressive network to obtain and collate information of relevance to early warning and an extensive database linked to the country of origin information project is already in existence. There, is, however, no systematic approach to information management and analysis for the purpose of early warning and the CDR database is not yet accessible on-line to users in other sections of Headquarters. The technology exists but human resources and the methodology are lacking.

Recommendation 8

The early warning study recently initiated by CDR should seek to: clarify UNHCR's position on early warning, make recommendations on the development of an in-house capacity and define the parameters of cooperation with the integrated humanitarian early warning system recommended by the OECD/DAC report.

Emergency Preparedness and Contingency Planning

69. It is important here to underscore that the ERT concept developed by EPRS withstood its most severe trial to date well enough in the Great Lakes to prove that it is indeed the right approach. The recommendations in this paper aim at further strengthening of UNHCR's capacity to field such teams in the future.

70. Since 1994, considerable advances have been made in the emergency preparedness field and several weaknesses identified by the OECD/DAC Study have been largely remedied. In particular, the technical specifications and inter-agency mobilisation procedures of Government Service Packages have been clarified (see Annex E) and Contingency Planning Guidelines have been issued. For additional details and caveats, see para. 5-9 of the 1996 Conference Room Paper (Annex D).

Logistics and Procurement

71. The procurement and dispatch of relief goods from Headquarters to the Great Lakes was generally speedy and efficient. Some bottlenecks did develop, however, because STS was overburdened at the peak of the emergencies and the staffing of the Unit was not strengthened until October 1994. STS also experienced difficulties because of the lack of a single focal point within the Regional Bureau to coordinate the various needs of the field and engage in advance planning for longer term needs.

72. During the Burundi emergency in Tanzania, in-country logistics was "the most serious limiting operational factor" (ERT report N.Coussidis, p.8). Ironically, the knowledge of the country's logistics profile gained from that operation became, together with the quantity of relief goods dispatched to the area, a major element in the success of the Rwandese operation which took place some months later. The success at Ngara should therefore not obscure the fact that UNHCR was initially ill-prepared to handle the logistics of a major emergency in Tanzania. Insufficient knowledge of the logistics infrastructure in the Great Lakes region may also have contributed to the fact that the Goma airlift went on for several weeks longer than necessary as pointed out by the OECD/DAC study (3, p.101-2). The study also notes that "nothing was done in advance to increase the capacity of Goma airport ... and consequently the first weeks of the airlift were chaotic and inefficient" (3, p.99).

73. Another difficulty that surfaced in the entire region was poor commodity tracking. Cargoes arrived with insufficient prior notice and incomplete manifests, and the Commodity Tracking System set up by UNHCR caused nothing but "endless problems" (ERT report M.Connelly). It appears that a system suitable for use in emergencies is currently under study as part of the Supply Information Management System (SIMS).

Recommendation 9

To ensure proper handling of logistics in a large emergency, the following steps should be taken:

• stand-by agreements on logistics should be regularly reviewed;

• a study of the logistics infrastructure of emergency-prone countries must be part of contingency planning;

• an adequate number of logistics staff must be deployed with the ERT;

• an STS Procurement Officer should be deployed with the ERT to coordinate local and international procurement needs;

• STS must be strengthened rapidly if the need arises

• arrangements in the logistics field must be regularly updated and revised; and

• development of the SIMS Commodity Tracking System for use in emergencies should be completed.

Telecommunications

74. Virtually all ERT reports mention the serious problems in the installation of a regional telecommunications system in the crucial early stages. As explained by the Telecommunications Officer, there was indeed a six week delay in the delivery of six V-Sat stations for the region which was caused by component parts having been temporarily lost during dispatch by the US Air Force. Clearly, commodity tracking is not only a problem for UNHCR. While this may be seen as a one-off incident, the Telecommunications Officer also raised other more fundamental issues. In particular, he felt that there was a need for more regular consultations between Telecommunications and EPRS to review developments in the rapidly changing communications field and obtain advance notice of possible needs. Generally, he felt that there was a tendency not to involve Telecommunications staff "until things go badly" and that other sections were not sufficiently aware of the problems involved in installing and maintaining increasingly sophisticated systems.

Recommendation 10

Establish a framework for a regular dialogue between EPRS, Telecommunications Unit and the Desks in order to

• review technical developments;

• establish telecommunication guidelines for emergencies for inclusion in the Emergency Handbook;

• assess possible needs in emergency-prone areas and establish a telecommunications strategy; and

• identify equipment needs and staff training requirements.

Food Aid

75. Among the problems that have arisen in this sector, two must be highlighted in particular: difficulties in establishing an equitable food aid distribution system and problems with WFP concerning ration size and delays in the food pipeline.

76. In Goma and Ngara, food was initially distributed through commune leaders, which, as noted above, was "potentially beneficial to the militia and those who had been involved in the genocide" (OECD/DAC Study 3, p.96). Following the recent development of UNHCR Food Distribution Guidelines, the Office is now in a stronger position to establish appropriate distribution arrangements from the outset of an operation. Further comments on this and related matters are found in paras. 13-17 of the 1996 Conference Room Paper (Annex D).

77. Blockages in the WFP food pipeline were particularly serious during the Burundi emergency and resulted in malnutrition and even death among refugees in both Rwanda and Tanzania (see ERT reports by N.Coussidis and F.Groot). Sporadic delays were also experienced during the Rwanda emergency in Tanzania (see OECD/DAC Study 3, p.92), Uganda (see ERT report by E.Demant, p.7) and Zaire (see below). Witnessing the misery caused to many refugees by these delays was a source of great frustration for UNHCR staff and resulted in frequent calls for direct procurement by UNHCR of essential food items.

78. The WFP/UNHCR Memorandum of Understanding contains an Emergency Response Contingency Clause which does indeed provide for UNHCR procurement of food items in certain circumstances but it is not easy to invoke and requires consultations initiated by WFP. A delay in the shipment of CSB for Goma was only averted because senior WFP and UNHCR staff who happened to be on mission at the time were able to sign an ad hoc agreement in the field which allowed UNHCR to purchase the commodity "on an exceptional basis".

79. As noted by the OECD/DAC study, WFP staff have also had their fair share of grievances. They have often doubted the effectiveness of UNHCR's distribution arrangements and the reliability of information on refugee numbers. Only very close cooperation based upon openness and mutual trust can help to mitigate these recurrent tensions between the two organisations.

Recommendation 11

All possible measures must be taken to ensure effective collaboration between UNHCR and WFP in an emergency; these must include:

• joint high-level food assessment missions at the earliest moment resulting in joint agreement on rations and needs;

• joint liaison with donor representatives in the field and at Headquarters;

• regular consultations and information-sharing through in-country and regional meetings;

• the inclusion in the ERT of a nutritionist to monitor the nutritional status of the refugees and advise on rations and the composition of the food basket;

• the inclusion of a Food Aid Officer in the ERT who should act as a permanent interlocutor with WFP and deal with all matters of joint concern for the two agencies including assessment missions, distribution arrangements and approaches to donors; and

• a mechanism for the rapid resolution of supply problems through recourse to designated senior officials in the respective Headquarters.

Technical Support

80. As repeatedly stated in the ERT reports and confirmed by the OECD/DAC study, one of the greatest successes of the UNHCR response to the Rwanda emergency lay in the quality and competence of its technical input. Time and again, strong coordination of technical sectors by UNHCR proved essential to establishing the necessary framework for effective collaboration with implementing partners. These achievements are primarily due to the input of highly experienced PTSS staff and consultants from fellow agencies (Rädda Barnen, ODA etc.).

81. Vital in this respect was the experience gained from the Burundi emergency which had confirmed that technical experts in the fields of health, site-planning, water, sanitation and social services should be deployed as an integral part of the ERT. In addition, it had been found that "deployment should not only be for one-time needs assessment but rather a continuous involvement throughout the emergency phase including early phase intervention" (E.Morris, Lessons learned from the Burundi Emergency, p.2). This was acted upon during the Rwanda operation where the active and sustained involvement of technical experts in programme design, coordination and implementation led to such commendable results.

82. While the achievements were particularly notable in Ngara, the key role of UNHCR technical experts was also recognised in Goma. As stated in the ERT report, "sectors where UNHCR had technical experts early on during the operation (eg. health) got their programmes under way faster than sectors where coordination had to be delegated initially to NGOs or non-specialist UNHCR staff (eg. water, a most crucial sector in the beginning of the crisis)" (ERT report F.Grandi, p.10). Water was indeed crucial, not least because of the disastrous cholera epidemic which affected the refugees in Goma within days of their arrival. The problems encountered in this sector were exacerbated by the delay or inappropriate nature of technical equipment procured for the purpose of water treatment and distribution. In retrospect one must conclude that the water sector in Goma would probably have been better handled if UNHCR technical staff had been involved already at the contingency planning stage as they might have been able to identify the potential problems in advance and planned for a more appropriate response. The need for the involvement of technical experts already at the contingency planning stage is certainly one of the lessons to be learned from the Great Lakes operation.

83. A positive consequence of the serious problems encountered in Goma has been the much enhanced dialogue between technical agencies involved in emergency response which has taken place since that time. This was stimulated further by the in-depth discussions which have been held with donors in order to clarify the Government Service Package concept. The Interagency Technical Cooperation Meeting on Water and Sanitation Issues hosted by UNHCR in May 1996 is a direct result of these developments. The meeting was part of an ongoing process which should ensure a much better coordinated technical response to future emergencies.

84. Considering the fundamental importance of sound technical expertise for UNHCR's emergency response, it is preoccupying that there still appears to be a lack of institutional recognition and understanding of the role of technical support in UNHCR programming; indeed, the posting of technical staff is still sometimes viewed as optional rather than essential. What the Great Lakes emergencies have proved, however, is that technical know-how is an indispensable aspect of UNHCR's leadership function and its coordination and monitoring of the work of implementing partners and government counterparts. This lends added emphasis to the recommendation below which also reflects the very similar conclusions reached by Y.Stevens in her paper on the lessons learned from the Rwanda emergency with respect to management and coordination of technical sectors (see Annex B).

Recommendation 12

UNHCR's creditable record of technical coordination in emergencies should be maintained and further enhanced. This requires a clear institutional commitment with focus on the following measures:

• technical staff in key sectors should continue to be deployed as part of the ERT and remain for at least 6-8 weeks to set up and hand over workable systems;

• in major emergencies the coordination of sectoral activities should be undertaken by UNHCR technical staff and consultants;

• UNHCR technical staff should also be involved in contingency planning both in the field and at Headquarters;

• further stand-by agreements should be made with technical agencies to ensure a wider range of coverage and skills (there was a lack of French-speaking experts during the emergency);

• qualified technical staff should be recruited to support field operations as and when recommended by PTSS;

• to ensure the provision of adequate community services in emergencies the recommendations of the Joint Evaluation of the Agreement between UNHCR and Rädda Barnen (1995) should be fully implemented;

• regional networking as well as ongoing coordination between PTSS and technical staff in the field should become routine;

• PTSS should assist in finalising the revised version of the Emergency Handbook and develop further guidelines for certain technical sectors (the availability of UNHCR sectoral guidelines proved very beneficial during the emergency; for details see Annex B);

Recommendation 13

A formal framework should be established for regular consultations between PTSS, EPRS, STS and Telecommunications Unit to review contingency plans and emergency preparedness in technical sectors.

Environment and local population

85. Lack of early attention to environmental issues has been noted by several ERT members, including A. Moussa and M.Connelly. The tendency of host governments to see site selection only as a political issue is partly to blame. In both Tanzania and Zaire, for instance, the authorities identified sites with insufficient water resources. The resulting problems are still not resolved and have caused enormous recurrent costs.

Recommendation 14

At the contingency planning stage, host governments must be strongly encouraged to take note of technical and environmental considerations in site selection. The involvement of technical experts in contingency planning will serve to reinforce this message.

86. The recently published UNHCR Environmental Guidelines represent a significant advance in this field and contains a chapter on the emergency phase. It recommends that where there is a risk of serious environmental impacts "an environmental specialist should be included in the emergency team"; in all cases, one team member should be designated as environmental focal point (pp.16, 17).

87. The OECD/DAC Study noted that the local population is just as severely affected by a massive refugee influx as the environment, and recommends the establishment of a quick-disbursing fund to compensate host communities. This was raised at the UNHCR workshop of May 1996 and the participants' responses are summarised in paras. 18 and 19 of the 1996 Conference Room Paper (see Annex D). The consensus was that depending on the situation in the area concerned, such a proposal might be very costly and difficult to implement. While UNHCR's limited financial resources should be disbursed exclusively for the benefit of refugees, the organisation should endeavour to act as a catalyst to initiate programmes by other agencies that have the capacity to assist local populations in need.

Administration

88. The Rwanda-Burundi emergency confirmed once more the validity of deploying Senior Emergency Administrators and Emergency Finance and Administrative Assistants as part of the ERT. The EPRS Guidelines for the Emergency Administrator were found to be excellent. In this context, the following points of note were stressed in some of the ERT reports:

• the vital need for an experienced senior administrator to set up proper administrative structures and procedures from the very start of the operation; omissions and mistakes are hard to rectify later

• the possible need for two (or more) administrative staff on the ERT: one to strengthen an existing Branch Office, the other to set up a Sub-Office in the theatre of operations; and

• the possible need, in major emergencies, for an international secretary to the ERT Leader if suitable staff cannot be recruited locally; a deputy with external relations responsibilities may also be required to receive the large numbers of visitors to be expected.

89. The establishment of fair and consistent local salary scales was found to be a recurrent problem noted also by the OECD/DAC Study which concludes that "a salary policy is needed before a relief programme begins" (3, p.146). Differences between local salaries (or their absence!), local UN salaries and salaries paid to refugees did indeed become a source of friction and confusion. When the matter was raised at the UNHCR workshop in May 1996, there was a consensus that the issue merits closer attention.

Recommendation 15

Initiate a comparative study of local salary practices in emergencies, establish a set of basic guidelines and integrate salary policies into contingency planning.

Staffing

90. As so often in previous emergencies , UNHCR experienced a range of problems in trying to find experienced personnel to fill posts in remote, insecure non-family duty stations in the Great Lakes. The following specific constraints arose. It proved difficult to attract a sufficient number of experienced senior staff to replace ERT members upon the termination of their assignment which should, as a rule, not be extended beyond two months. General Service staff were frequently deployed during the emergency and many performed very well in posts subsequently graded in the professional category. However, because of their limited career opportunities in many cases they were not eligible to occupy such posts on a longer term basis. Meanwhile, new staff deployed as part of the emergency response were given insufficient training prior to their assignment. Finally, the lack of standardised organigrams, job descriptions and the absence of a focal point for staffing at Headquarters compounded the situation and led to further delays.

Recommendation 16

There is a need for additional multifaceted efforts to enhance UNHCR's capacity to respond to emergency staffing needs; this should include:

• the development of crash training programmes for new staff, consultants and implementing partners;

• career opportunities and rewards for General Service staff who perform exceptionally well in emergencies; and

• additional incentives to induce senior staff to serve in emergencies

91. The Rwanda-Burundi emergency exposed staff to high levels of stress and physical danger. The lessons to be learned from working in such an environment are currently being examined in a global study conducted by the Inspection and Evaluation Service and are not discussed here. In Ngara appalling living conditions were an additional cause of stress and ill-health. With mobile accommodation now part of the emergency stockpile, this type of problem at least should be a matter of the past.

The Role of EPRS

92. The professional staff of the EPRS are selected on the basis of their proven operational competence and experience. The Great Lakes experience has highlighted a number of important and related areas in which their skills could greatly benefit the work of the office, particularly in the period between major emergencies. Some of these activities would further enhance the response in the case of an eventual emergency.

• participation in early warning activities, an important area in which the policy of the House has still to be defined with greater clarity;

• participation in contingency planning exercises both at Headquarters and in the field; this should include fact finding missions to prepare information packages and standard logistics profiles on locations identified as likely places of mass influx;

• using EPRS staff, in particular emergency administrators, to help in opening new UNHCR offices also in non-emergency settings; their training and know-how will ensure that proper structures and procedures are established right from the beginning;

• the updating of existing training materials for emergency related topics, the production of new materials and the development of an integrated training strategy, issues touched upon in several recommendations of this paper;

• liaison with other agencies on issues of joint concern in the field of emergency preparedness and response; this might also include loaning EPRS staff to assist fellow UN Agencies in dealing with non-refugee emergencies (eg. natural disasters or the establishment of human rights monitoring networks); and

• participation in practical training activities in the field, including simulated deployments modelled on comparable exercises undertaken by professional bodies such as civil defence forces or the police.

Recommendation 17

The proposed expanded role of EPROs should be discussed at the Workshop on Emergency Management referred to in the final paragraph of the Overview.

V. MANAGEMENT ISSUES

93. The emergency brought to the fore a number of problems related to the managerial structure of UNHCR. The most salient of these are discussed below under three headings: Headquarters level, regional and field office level. The concluding recommendations take account of the Change Management proposals currently under discussion as a result of the Delphi process.

Emergency Management at Headquarters

a) Support Services and Regional Bureau

94. The consolidation of UNHCR's emergency response capacity since 1991 was undertaken by support service sections, notably EPRS, PTSS and STS. Responsibility for operational management was to rest with the Regional Bureau. However, the deployment of an elite of highly operational ERTs by DPOS and their insertion into existing Bureau structures engaged in demanding ongoing programmes did not always run smoothly. In some cases the Bureau did not change priorities fast enough and failed to provide rapid support for the ERT in the field. These structural problems which had been evident for some time came to a head during the Goma emergency as a result of an unfortunate constellation of personalities, staff absences and staff rotation at this crucial time. At one point the High Commissioner intervened to assign overall responsibility for the operation to the Head of DPOS until, several weeks later, the Regional Bureau was once more put in charge following the arrival of the new Director. It speaks for the professionalism of UNHCR staff that this period of uncertainty at the top did not more seriously hinder the operation in the field.

b) Management by the Bureau

95. The Africa Bureau's response to the emergency went through a number of phases during 1993-1994 which resulted in the creation of the Special Unit for Rwanda and Burundi (SURB). The experience at the time showed once more that immediate strengthening of Desks upon the onset of a major emergency is as important as a rapid presence in the field. Nevertheless, there appear to have been moments during 1994 when some Desks were totally overburdened, particularly during the summer period when some staff were on leave and no replacements were designated. Serious problems due to understaffing were also experienced by STS as noted earlier. Splitting responsibilities between ongoing and emergency operations also created problems particularly with respect to Zaire.

96. It is questionable whether SURB in its current form was the most effective way to manage the operation. The unit consists essentially of country desks linked by an ill-defined decision-making structure which includes a Head of Operations of similar seniority to the Heads of Desk and a Coordinator whose function focuses primarily on external relations. Altogether, the Bureau structure involves too many reporting and decision-making layers (Desk Officer, Head of Desk, SURB Head of Operations, SURB Coordinator, Deputy Head of Bureau, Head of Bureau). In several instances, it appears to have been unclear who was ultimately responsible and decisions taken during task force meetings were later overturned by more senior colleagues. The work of support services such as STS and DHRM was on occasion hindered by the lack of clearly designated focal points to decide on priorities and regional approaches. It should be noted, however, that these criticisms deal with aspects of managerial structure only and do not pertain to the work of individual staff-members of SURB who have clearly done their utmost to support the operation in the field.

c) Information Management at Headquarters

97. The lack of a good regional UNHCR telecommunications network early in the emergency combined with the abundance of often conflicting reports from NGOs, donors and the media made it difficult for Headquarters to gain a full and accurate picture of what was happening in the field. Dissemination of information by Headquarters was also problematic with key documents not copied to the field and messages from the field not copied to all Headquarters sections concerned. The resulting confusion was noticed by donor missions liaising with Headquarters during the height of the emergency.

98. The Goma II ERT report concludes that "Headquarters must seriously reexamine systems through which information is received, circulated , assigned for action and responded to in Geneva" (F. Grandi's mission report, p.18). This has not been followed up and there remains a need to establish a clear set of procedures for information management in emergencies. It is a highly technical full-time occupation which is rendered increasingly complex through the information glut produced by instant communications. The lessons learned from the successful Sarajevo Airlift Information Centre may be a helpful starting point for addressing this problem (see report by M.Alford of 17/7/92).

d) Management of multiple emergencies

99. Both the Burundi and the Rwanda operations were emergencies taking place simultaneously or in quick succession in different countries. Given the atomised structure at Headquarters and in the field, it proved to be difficult to maintain a balanced and consistent regional approach. In the Burundi emergency "there were disproportionately more resources allocated to Rwanda than to Tanzania and Zaire" (J.Lim, Lessons Learned from the Burundi Emergency). When the Goma tragedy unfolded, attention at Headquarters quickly shifted away from all other areas. As a result, morale in Ngara suffered and even illnesses among staff increased. ERT leaders returning to Geneva from other localities in the region were not even given the possibility of a Headquarters debriefing. For staff working in emergency conditions, the link with Headquarters is a crucial psychological lifeline. Headquarters must be aware of this and ensure that operations which are not in the limelight "still continue to receive the attention they deserve" (F.Groot, Goma I report).

Operations management in the region

100. Following the Burundi emergency, there was a call for a standard model for establishing the level of staffing and support needed to run an operation so as to ensure a more equitable distribution of resources in multiple emergencies. The idea was not followed up in view of the extent of variation of conditions from one situation to another and it was agreed that the solution must lie instead in a well coordinated regional approach which permits staff and resources to be moved from one locality to another depending on needs.

101. The management set-up during the subsequent Rwanda emergency, however, made such a regional approach difficult to realise. Both in the field and in Headquarters, management continued to be primarily country-based. The post of Special Envoy for Rwanda and Burundi created on 20 January 1994 did little to mitigate this. The first envoy had no staff and no executive authority. Field offices were only required to "keep him regularly appraised of activities" (IOM/46/FOM/48/94) and he was not consulted even when crucial decisions had to be taken. His role was limited to "supporting and enhancing UNHCR's efforts", "assessing the capacity of UNHCR's activities" and "providing advice and guidance where appropriate".

102. His successor who took over on 12 December 1994 also found himself working very much in limbo. The lack of a structural link between the Special Envoy, the field representatives and SURB made friction almost inevitable. Staff with regional responsibilities were now placed under the Envoy's supervision but they were not concentrated in one location and instead remained in their original duty stations. This was done to avoid creating another heavy structure in the field in parallel to SURB at Headquarters. However, in practice, it turned out that the physical dispersal of these staff members "prevented effective coordination of their activities" (C.Faubert).

103. The result of all this was the effective absence, at Headquarters and in the field, of a strong, centralised regional approach to operations management both during the emergency and thereafter.

Field Office Management

a) The ERT and the local network

104. The establishment of good working relationships between newly arriving ERT staff, the existing UNHCR Branch Office and aid agency staff already on the ground must be managed with care. The ERT may be perceived by some as an unwelcome imposition making demands on scarce resources. Access to office vehicles has, for example, been difficult for ERT logistics staff on mission, a fact which "has on many occasions prevented them from carrying out their duties expeditiously and created conflict within the office" (H.R.Leefe, Burundi Emergency mission report).

105. The first ERT arriving in Goma in April 1994 was not well received by aid agency staff already there who viewed them as potential rivals (see ERT Report by F.Groot). Relations with ICRC never improved and the latter's failure to provide UNHCR with information on population movements in the country of origin contributed to the lack of advance warning of the massive influx that took place in July 1994.

106. A number of lessons can be learned from this experience. Firstly, the role of the ERT within the existing Branch Office/Field Office structure must be clearly defined by Headquarters. In situations where the area of operations is far removed from the Branch Office, a direct chain of command between the ERT and Headquarters may have to be established as was done in Goma. As a rule, however, this should be avoided, since by-passing the Branch Office inevitably leads to difficulties of in-country coordination. Secondly, while the Branch Office should give priority to the ERT's needs, Headquarters must ensure the ERT has the capacity required to strengthen the resources of the Branch Office both in terms of staffing and equipment. Thirdly, before the deployment of an ERT, it is advisable for Headquarters to inform agencies with offices in the locality concerned (in particular ICRC) of its arrival and terms of reference to encourage their full cooperation.

107. The Ngara operation provides a good example of an ERT functioning successfully within its local network. A number of different elements had to come together to make this possible. A crucial element was the organisation of basic contingency planning and inter-agency discussion at the local, national and international levels which provided for an immediate coordinated response "without confusion over who is doing what as a first step" (ERT leader's report, Ngara). There was also exemplary cooperation with the Tanzanian Government which requested UNHCR to coordinate the operation and allowed "the ERT considerable clout in allocating responsibilities to the NGOs that were present" (OECD/DAC Study 3, p.31-2). The ERT was given excellent support by the Branch Office and by Headquarters which together responded rapidly to field requests and ensured the speedy arrival of relief goods. Co-operation with other UN Agencies, including UNICEF and WFP was positive and there was a high level of donor support, in particular from ECHO. At camp level the refugee leadership " (for whatever reason of their own) worked with UNHCR and the agencies" (ERT leader's report, Ngara). Last but not least, the UNHCR team displayed an exceptional degree of managerial and technical competence.

108. The Ngara experience, and the Great Lakes operation in general, provide a number of lessons on implementing partner selection in emergencies and its relationship to overall coordination.

b) Management within the field office

109. As noted above, the management skills of the ERT leaders during the emergency were outstanding and their efforts provided the cornerstone for the operational success noted by the OECD/DAC report. Nevertheless some field offices experienced difficulties in the division of responsibilities between technical coordinators, programme officers and field officers. In some cases the technical and programming functions reported separately to the head of the office which made programming as an integrated activity more difficult. A related topic is taken up in the lessons learned paper by Y.Stevens who recommended that "sectoral coordinating functions should also imply responsibility for programming in the given sector" and calls for technical staff to be given relevant training (see Annex B, paras. 14, 15). F. Groot noted the need for a training package to provide field officers inter alia "with the criteria and methods to carry out basic technical monitoring" (Rwanda ERT report, p.19). In some cases a micromanagement mentality precluded direct communications between technical staff in the field and their respective support services at Headquarters.

110. The different possible interfaces between technical, programming and field officer functions should be examined critically in order to establish a set of standard organigrams and reporting lines with recommendations on best practice which would be most useful for training purposes and the establishment of new offices.

c) Management Systems

111. Several reports have commented negatively on the suitability of the FMIS programme in emergencies. It has been described as "very user unfriendly, complex, slow and frustrating". It is important to ensure that the FMIS replacement currently under discussion will be more adaptable for use in an emergency.

Lessons learned and the Change Management Process

112. The Situation Manager concept established by the Delphi process would probably have obviated many of the managerial difficulties encountered during the Rwanda-Burundi emergencies. It provides for a strong, field-based approach to regional management which is precisely what was lacking most. It is unlikely, however, that the proposals as summarised in the Report of the Working Group on Operations Management would automatically remove all the difficulties encountered. The report rightly recommends that the Emergency Preparedness and Response function should remain based at Headquarters. The deployment of ERTs from Headquarters into an existing office network dealing with ongoing operations is still going to be a potential source of friction unless priorities are clearly established and resources allocated accordingly. Information management will also be no less problematic.

113. The Report of the Working Group does not appear to make explicit mention of a centralised technical support capacity. In this context it must be stressed that, as shown in the Rwanda-Burundi operation, UNHCR's emergency response capacity depends to an equal measure on the input of EPRS, PTSS and STS. Without experienced technical support there is no emergency response.

114. In the light of the wide range of managerial problems experienced during the operation, the recent EPRS proposal to convene a Workshop for Managing Emergencies at Headquarters should be strongly supported. It would be most opportune to hold it in the near future in order to ensure that the new organisational structure takes full account of the specific needs of effective emergency response management. The issues that require priority attention are summarised in the following recommendation.

Recommendation 18

In view of the management problems detailed above, a workshop should be convened to examine the functioning of emergency management within the new organisational structure which is being set up as a result of the Delphi process; in doing so, particular attention should be given to the following:

• a strong, regional managerial structure is required with few reporting layers and clearly designated focal points for individual sectors;

• a mechanism with a clear decision-making structure and clearly designated focal points must be established to coordinate Headquarters' support for operations in the field;

• clear reporting lines and divisions of responsibility must be established when emergencies have to be managed within the context of an ongoing programme;

• a plan of action should be devised to set up effective information management systems;

• the suitability of the newly-devised financial, administrative and programming procedures for the purposes of emergency management should be examined;

• standard organigrams should be devised to ensure effective coordination of technical, programming and field office monitoring functions; and

• an updated staff and management training package for emergencies should be devised to take account of the managerial changes and ensure that lessons-learned become part of operational practice.

VI. CONCLUSIONS

115. Underlying the various recommendations of this paper are three main areas of concern which are briefly discussed hereunder.

The composition of the ERT

116. The most important single decision concerns the composition of the ERT. Taking into account the recommendations made in this paper as well as in the ERT reports and the workshop discussions the team required to cope with a large emergency may comprise any of the following:

• Team leader

• Senior Protection Officer

• Senior Programme Officer

Technical Experts (Health, Water, Sanitation, Nutrition, Physical Planning, Environment, Community Services, Anthropology, see Recommendations 6, 12)

• Logistics Officer

• Procurement Officer

• Telecommunications Officer

• Food Aid Officer

• Security Officer

• External Relations Officer

• Senior Emergency Administrator

• Emergency Finance and Admin. Assistant

• International Secretary

117. Such a large team will rarely if ever be fielded in practice. The team's composition, however, will determine the course of the operation. Good contingency planning and expert professional advice by EPRS may be the best guarantee to ensure that the right decisions are made. This must be combined with extended stand-by arrangements to ensure that a sufficient number of properly trained staff in all areas of expertise noted above are readily available as required. Special attention must be given to Protection.

The need for an in-house forum for regular consultations on Emergency Preparedness

118. During the interviews conducted for the preparation of this paper it transpired that different sections at Headquarters whose work is of relevance to UNHCR's emergency response are not sufficiently in contact with each other, are not fully aware of each others' activities and sometimes even feel excluded or by-passed. Considering the fact that emergency preparedness is a dynamic topic in a rapidly-changing technical environment which involves many different sections of the house and has increasingly important implications for inter-agency coordination, it may be advisable to establish a framework for regular in-house consultations as noted in Recommendations 10 and 13) above. The purpose of such a forum which might meet on a quarterly or six-monthly basis when no emergencies are taking place should be to:

• examine contingency plans;

• exchange information on relevant technical developments;

• review and expand stand-by arrangements;

• discuss guidelines and training issues;

• take note of ERT reports and their recommendations;

• review lessons learned from past operations;

• examine developments on the inter-agency front;

• agree on objectives and priorities for further action;

The establishment of such a forum might be adopted as an additional agenda item for the Workshop on Emergency Management at Headquarters proposed by EPRS and endorsed in Recommendation 18 of this paper.

An integrated approach to guidelines and training

119. UNHCR has now become so awash with guidelines, handbooks and training materials that it has become difficult for staff to absorb all the information they contain. The need for more material of this kind noted in nine of the 18 recommendations in this paper may therefore be met with some caution. However, it is a fact that the production and dissemination of such documents has contributed immeasurably to the professionalism of the organisation and increased its credibility with governments and implementing partners. What is required for the purposes of emergency preparedness and response is an integrated approach which links the information available in guidelines and handbooks with a training strategy that caters for a variety of clients ranging from experienced staff to newcomers and provides materials for crash courses to be held in an emergency setting.

120. Central to this endeavour must be the finalisation of the revision of the Emergency Handbook to replace the outdated 1982 version. While it is useful to have separate documents dealing with different aspects of emergency response, there is no substitute for a revised one/two volume publication. Much preparatory work on this has already been accomplished earlier in the decade before work was interrupted by other priorities; it should be resumed and completed without delay.

121. Action required to establish the training strategy recommended above may be summarised as follows:

• complete the revision of the Emergency Handbook (see also Recommendation 12);

• initiate the drafting of further guidelines on issues of concern (see Recommendations 5, 12, 15);

• produce simple concise summaries of guidelines on all sectors which may be used as hand-outs for purposes of briefing or training;

• review and continually update existing training materials with a view to producing packages suitable for trainees with different levels of experience; and

• continue to liaise with emergency training activities taking place in other agencies and organisations.

Such a strategy would be able to build on the considerable advances already achieved by UNHCR in the field of emergency training and may provide the best guarantee to ensure that lessons learned from previous operations are passed on to other staff members, new and experienced alike.

Acronyms

CSB: Corn Soya Blend

CDR: Centre for Documentation and Research

DCMS: Division of the Controller and Management Services (UNHCR)

DHA: Department of Humanitarian Affairs (UN)

DPA: Department of Political Affairs (UN)

DPKO: Department of Peace-Keeping Operations (UN)

DPOS: Division of Programmes and Operational Support (UNHCR)

ECHO: European Community Humanitarian Office

EFAA: Emergency Finance and Administrative Assistant

EMERCOM: State Committee of the Russian Federation for Civil Defense, Emergencies and the Elimination of the Consequences of Natural Disasters

EPRO: Emergency Preparedness and Response Officer

EPRS: Emergency Preparedness and Response Section

ERT: Emergency Response Team

FMIS: Finance Management Information System

ICRC: International Committee of the Red Cross

IDP: Internally Displaced Person

IES: Inspection and Evaluation Service (UNHCR)

NGO: Non governmental organisation

OAU: Organization for African Unity

ODA: Overseas Development Administration (UK)

OECD: Organization for Economic Cooperation and Development

POS: Protection Operations Support

PTSS: Programme and Technical Support Section

RPF: Rwandese Patriotic Front

SCAF: Sub Committee on Administrative and Financial matters

SEA: Senior Emergency Administrator

SIMS: Supply Information Management System

SMC: Senior Management Committee (UNHCR)

STS: Supply and Transport Section (UNHCR)

SURB: Special Unit for Rwanda and Burundi (UNHCR)

UNAMIR: United Nations Assistance Mission to Rwanda

UNBRO: United Nations Border Relief Operation

UNICEF: United Nations Children's Fund

UNV: United Nations Volunteers

WFP: World Food Programme (UN)

ZCSC: Zairian Camp Security Contingent