The Evolving UN Cluster Approach
in the Aftermath of the Pakistan Earthquake:
An NGO Perpective

A Report by ActionAid International
April 2006

Executive Summary and Action Points

INTRODUCTION

The South Asia earthquake on 8 October 2006 claimed at least 73,000 lives. Measuring 7.6 on the Richter scale, it left many others without food, clothes and shelter to face the harsh winter. So far, through ActionAid International's emergency response, thousands of tents, packs of food and basic supplies such as blankets, as well as medical supplies, have been distributed in Pakistan. Within one week of the earthquake, relief aid was being delivered by ActionAid staff in three different locations in an area which is extremely inaccessible, as well as being politically sensitive.

The many challenges faced by those who were implementing the emergency response included getting to grips with a new approach to humanitarian response being developed by the UN - the cluster approach. This report draws from the experiences of UN agencies, international, national, and local NGOs, and donors in responding to the earthquake to present an analysis of the cluster experience in Pakistan, along with recommendations for the future.

1. EXECUTIVE SUMMARY

The aim of this report is to highlight issues which need to be "factored in" to the development of the cluster approach, a key aspect of the UN humanitarian reform agenda. The intention is to provide an assessment of the practical value of the approach in Pakistan and the implications of this for the humanitarian reform agenda. A summary of the main findings can be found in "Action Points" in Section 2 of this report.

A point to be acknowledged from the outset is that the Pakistan earthquake was the first occasion in which the cluster approach was implemented in a disaster response situation, and therefore it is too early for the validity of the approach to entirely stand or fall by experiences in Pakistan. There are, however, valuable lessons to be learned from the implementation of the approach there.

In broad terms, the earthquake response was regarded as having been effective, particularly as the feared second wave of winter deaths was avoided. Pim Kramm, the Deputy Head of the Dutch Humanitarian department, commented that operations worked well in Pakistan and he felt a tangible difference between his visit to Pakistan and other disaster zones. Key factors identified as affecting the success of the response, after initial difficulties, were the high level of cooperation from the Pakistani government and the relatively mild winter. It is unclear how much of a difference the cluster approach made itself.

1.1 Outline of the Cluster Approach

The cluster approach was first set out in a Humanitarian Response Review (HRR) paper commissioned by the UN. The basic premise was that accountability, predictability and reliability could be improved by identifying organisational leaders for areas in which there was an identified gap in humanitarian response. These organisations would then be responsible for specific areas, or clusters.

Within the first 24 hours of the response, a set of nine clusters, Food and Nutrition, Water and Sanitation, Health, Emergency Shelter, Early Recovery and Reconstruction, IT Telecommunications, Logistics, Camp Management and Protection, modelled on the HRR recommendations, were established in Islamabad, plus a 10th cluster for Education. Field cluster sites were established in each of the main UN field presences and dubbed "humanitarian hubs". However, as the emergency progressed, the number of clusters and sub-clusters grew exponentially, making it difficult for NGOs to keep track of the number of clusters that existed.

There was a high level of confusion amongst both UN and NGO staff as to what the cluster approach was about, primarily because it was still in the process of being developed at the global level when it was introduced into Pakistan. As a result, those implementing the approach did not have Terms of Reference, appropriate support or training. Draft generic Terms of Reference for cluster leads at country level were developed in Geneva in January 2006, and these are attached as Appendix 2 to this report. However, it should be borne in mind that the cluster approach was being implemented in Pakistan well before these were developed. Clusters were credited with providing an opportunity for information sharing and for people in the relief effort to network on the fringes of meetings. They also provided the potential for coordination attempts, which had varying degrees of success, but which were regarded as worthwhile in any case. Both INGOs and donors commented that the fact that a named agency was responsible for coordinating efforts in a particular area was helpful.

Adequate attempts were not made to involve local NGOs and governmental structures. Local NGOs regarded cluster meetings as meetings of an elite group of foreigners, which, though helpful, did not pay sufficient attention to the ideas and issues raised by local NGOs. The vast majority of NGOs, both local and international, felt that cluster meetings, which were always held in English, should have had an Urdu interpreter present, to enable local NGOs to be involved. Even those who spoke English among the Pakistani NGOs said that they had difficulty in following all the UN acronyms and they often felt as if they were in UN internal meetings. Others regarded cluster meetings as talking shops and preferred to spend their time in the field.

The humanitarian intervention in Pakistan contrasted with that in complex emergencies by the presence of a strong, though not always constructive, state structure. The role of the Pakistani military in the relief effort was praised by the UN and NGOs alike, however there are concerns about its part in the relief effort, particularly due to the lack of parliamentary and civil oversight of reconstruction funds. There was very limited effort on the part of the UN to empower local democratic structures, which were already weakened by the policies of the military government prior to the earthquake. These were further damaged as a result of the earthquake and were then side-lined in the humanitarian response.

1.2 Relative Cluster Successes

There was a clear finding that the performance of clusters varied widely from cluster to cluster. One INGO commented that, whilst the cluster approach was a common- sense one, some clusters got "so bogged down with the mechanics they completely lost their focus". Some clusters were carried by the charisma of their lead, while those oriented towards ground-level work such as logistics, food and shelter were more successful. Those clusters that focussed on future livelihood strategies/irrigation were regarded as not being immediately relevant to the relief effort. Attendance in Health, Education and Watsan was low in the beginning and Livelihoods and Protection still has a small membership; Protection has not been a successful role. Clusters that operated well need to assist not so successful colleagues and there should be feedback to Geneva on what worked and what did not in terms of running clusters.

Clusters were hampered by a lack of full attendance at meetings and problems with, for example, operational/field staff located at hub clusters while decision-makers, such as heads of agencies in Islamabad, a problem compounded by communication problems between hubs and clusters. In Pakistan, the cluster approach appears to have been expanded to cover the entire humanitarian response, rather than simply gap sectors identified in the Humanitarian Response Review. This may have had an impact on its effectiveness. Meetings were too long and too frequent, there was a general feeling among NGOs that clusters were overly compartmentalised and there was no need for so many. There was also a spread of sub-clusters, and complaints of duplication and overlap. NGOs described a non-participatory attitude on the part of the UN, where they were treated simply as implementing partners, or "policed", rather than having an input into conceptual thinking.

There was also a criticism that there was not enough analysis, synthesis and thinking ahead within meetings. Clusters provided some back up support on technical matters, such as shelter design and heating arrangements, but not enough. Too much time was spent agreeing on guidance on what kind of support to provide. This led to people by-passing clusters in deciding what kind of response was necessary. There was a lack of monitoring and evaluation, and synthesis between clusters and across hubs and clusters. This lead to duplication and omissions. For example, in the week immediately following the earthquake in Muzzaffarabad, children were emptying mineral water bottles into a large barrel for washing and bathing purposes, presumably as a result of a lack of water for washing and a surplus of drinking water.

Representatives of donor agencies attended clusters sporadically but did not have decision-making authority in terms of funding proposals. In general, donors supported the cluster approach as something to be developed. They were cautious of coming to any firm views on its success in the light of what happened in Pakistan as the cluster approach had not been finalised at global level. Most donors stated that the approach had not yet affected funding attitudes to the UN and they would continue to channel funding where this was felt to be most effective, including to a strengthened humanitarian coordinator function.

There were structural problems with the clusters, such as very little back up support for cluster leads who were essentially taking on two full time roles, an agency role and a separate role as head of cluster. There were also concerns about limited information flow between cluster and hub level, and lack of staff at hub level. In addition there was a high staff turnover, which inhibited the development of institutional memory and the ability to develop relationships with stakeholders. The lack of staff and high turnover may need to be addressed as part of wider issues within the UN. A number of donors stressed the need to support OCHA.

1.3 From the Outside

Geneva's role in the implementation of the cluster approach in Pakistan is unclear, but the Inter Agency Standing Committee (IASC) Earthquake Task Force missed an opportunity to help overcome problems concerning gaps, duplication and analysis. If adequate support is not available at global level, cluster leads may not wish to "carry the can" for all cluster failings. There does not appear to be any mechanism in existence for stakeholders in Pakistan to input into the humanitarian reform agenda in the light of their experiences.

The cluster approach carries a considerable amount of potential, but there is clearly some way to go before this can be fully realised. Clusters cannot be an answer without the availability of trained staff, adequate support and appropriate engagement with outside actors to ensure that measurable improvements in humanitarian response are made.

2. ACTION POINTS

2.1 Engagement with Local Democratic Structures

• It is within the gift of those involved in the earthquake response to empower such civil structures as exist after the earthquake, at the most local level. For the reconstruction phase in Pakistan, clusters should focus on independent assessment of the government's reconstruction priorities and on capacity building with secular NGOs and elected local bodies.

• While forming clusters in a country there needs to be a clear exit strategy. Key actors in local authorities should be identified early through existing UN agencies operating in a country and other sources. All possible support must be provided to them so that all the information, networking and capacity building done in the early stages is not lost, but is relevant to and built on for the recovery and reconstruction stages.

2.2 Encouraging Local Organisations to attend Cluster Meetings

• OCHA, the Humanitarian Information Centre (HIC), and the cluster leads need to proactively identify key local actors and make sure that they know how to get access to minutes of cluster meetings even if their staff are not attending them.

• A clear strategy for local involvement is required. This will require the IASC country team to strengthen its knowledge of and engagement with local/national NGOs prior to an emergency occurring.

• The monitoring of a cluster's performance, particularly at hub level, should include an assessment of the attempts made to involve local communities and civil structures.

2.3 Language

• Having an Urdu interpreter present at cluster meetings would have enabled local community groups and citizens to engage with the process, and assisted with capacity building. It would also have helped ensure that the response was implemented in a culturally appropriate way. In the same way, cluster meeting notes, agendas etc., should have been translated into Urdu as far as possible.

2.4 Engagement with the Pakistani Government

• There needs to be analysis of the differences between the implementation of the cluster approach in fragile states, as compared to that in a strong but relatively autocratic state, including any lessons learned.

• For the reconstruction phase in Pakistan, there needs to be an independent assessment of the government's reconstruction priorities.

2.5 Encouraging the Involvement of non-UN Organisations

• Terms of reference need to be developed for cluster leads which take into account the experience in Pakistan.

• UN agencies need to make significant further outreach to NGOs, particularly local NGOs, to set out the role of clusters and how NGOs can interact with them. Feedback from NGOs should be assimilated into future work plans.

• In particular, UN agencies need to adopt a more participatory approach with regards to NGO involvement in clusters.

• NGOs should be involved in the conceptual stages of planning and treated as genuine partners rather than implementing agencies. The UN should avoid calling or changing meetings at short notice without consultation. This is an issue of "attitude" which needs to be resolved if NGOs are truly to engage with clusters.

• As the cluster approach appears to have been adopted for the entire UN earthquake response in Pakistan, the rationale for having different clusters to cover "gap" areas no longer applied and it may have been better to merge linked clusters, as suggested by one respondent. This is an issue to consider for future sudden onset emergencies.

• Better communication connections between hubs and clusters, and across clusters, could counter problems of not having key decision makers in the same place at the same time.

• More senior and experienced technical staff need to be deployed at field level. This would assist with support and capacity building and improve analysis.

2.6 Linkages/Analysis

• Structural issues such as administrative support and information management need to be resolved so that adequate analysis of information can take place and there is free flow of information across hubs and clusters. This will address the issue of "connectivity", ensuring that analysis benefits from input from all relevant clusters and hubs.

• An appropriate mechanism needs to be developed for feedback from cluster leads to the global level concerning what worked during the Pakistan response and what did not, so that lessons can be learned and good practice developed.

• Linkages between earthquake response and ongoing development activities need to be identified and bolstered for the recovery phase.

• Mechanisms need to be developed for forward planning within clusters from the outset of an emergency.

• The lack of strategic analysis, overview and in-depth discussion was universally felt to be missing from the cluster approach and incorporating this into the work of clusters is likely to have a dramatic effect on NGO attendance.

• The role of global clusters with regard to the above points needs to be considered.

2.7 Cross Cutting Issues

• Specialist UN agencies should have an independent role in monitoring thematic issues in a cluster response, for example UNIFEM for gender, the Human Rights Council for human rights and UNEP for environmental issues. This may be through representatives of these organisations attending cluster meetings/acting as advisory support for all clusters on relevant issues.

• ActionAid International notes the existence of the joint UNEP-OCHA environment unit in Pakistan and takes the view that crosscutting agencies should take the lead in evaluating the performance of clusters in their own right against clear thematic standards.

2.8 Funding

• The role of donor agencies representatives within the cluster process should be clarified. IFIs in particular need to be encouraged to attend meetings.

• The relationship between clusters, the CAP and the Humanitarian Coordinator (HC) needs to be considered and delineated.

2.9 Mass Media

• Mass media could be better utilised to spread awareness about clusters and to carry out coordination activities. The UN needs to have a bold, centrally devised strategy which can then be implemented and adapted by IASC Country Teams according to the needs of the emergency and the particular country in which it occurs.

2.10 Structural Issues

• There needs to be improved information flow between cluster and hub level, and across clusters. A central coordination system at each cluster and hub will resolve overlaps between clusters and meetings.

• Full training should be provided for clusters on their roles and responsibilities. Guidance should be provided to cluster leads on how to avoid unnecessary fragmentation into sub-clusters and control "cluster spread", whilst avoiding cluster meetings that are so large as to be unwieldy.

• The number of cluster meetings should be rationalised so that they become manageable.

• There is an "over-clusterisation" at national level, when it is actually the hubs that need the most focus, as they are most directly involved in the relief effort.

• Administrative support for cluster leads is a necessity, and the co-location of cluster heads should be considered.

• The frequent turnover of cluster leads must be resolved. This may require consideration of the length of the employment contracts of UN staff and the manner in which they are rotated in and out of emergencies.

• It is necessary to distinguish between the duties of heads of agency and cluster leads, whilst ensuring that the lead organisation attends meetings in its own right.

• National clusters should concentrate on information consolidation from hubs and policy support to hubs.

• The role of OCHA within the cluster system needs to be clarified and supported.

2.11 Monitoring and Evaluation

• The cluster approach relies strongly on information provided by stakeholders about their activities, and it is important to ensure that the information provided is accurate. Mechanisms should be set up to re-survey and monitor areas of activity in which stakeholders operate.

• Data formats and standards need to be agreed for making the information more usable, as the variation in these caused a lot of confusion.

• Databases need to be kept up to date.

• As recommended by the HIC and Real Time Evaluations (RTE), information management requires a dedicated cluster with additional investment in upgrading skills, raising awareness and communications technologies, e.g., internet cafes. This cluster could also play a role in engaging the media.

2.12 Engagement with the Cluster Approach at Global Level

• The RTE identifies the confusion amongst NGOs between the cluster and sector approaches. ActionAid International's research identifies a further lack of clarity concerning the role of cluster leads at global cluster level.

• Interviews with local, national and international NGOs on the ground have disclosed an almost complete lack of knowledge of the role of cluster leads at global level.

• A clear demarcation between the different roles of clusters at global, as well as national and hub level would enable stakeholders to decide whether it was appropriate to provide input into the global process, including through an NGO coalition such as International Council of Voluntary Agencies (ICVA).

• The IASC task force on earthquake response in Geneva appears to have added little value. The task force could have assisted in analysis of the information coming in or strategic overview, flagging gaps, etc. However, the meetings mainly dealt with numbers of items distributed etc. This was a missed opportunity.

• As the cluster approach is implemented in different countries, mechanisms need to be created for affected populations and stakeholders, such as those in Pakistan, to input into the global reform process, should they wish. This would enable the global process to benefit from a "bottom-up" participatory approach.

• Global clusters/the IASC need to consider how best to support the cluster approach as implemented in a particular country, how to be a resource in respect of lessons learned in the past and how to draw out best practice for the future. The mechanisms by which they intend to do this should be explained to stakeholders so that they can input into the process if they wish.