Issue 37 - Article 11

The UN Cluster Approach in the Pakistan earthquake response: an NGO perspective

April 15, 2007
Anne Street, ActionAid, and Gita Parihar, independent consultant

The South Asia earthquake on 8 October 2005 claimed at least 73,000 lives, and left many others without food, clothing and shelter. The many challenges faced by those implementing the emergency response included getting to grips with a new approach to humanitarian response being developed by the UN – the cluster approach. This article draws on a larger ActionAid report to present an analysis of the cluster experience in Pakistan from an NGO perspective. While it is difficult to extrapolate from this one experience, a number of lessons emerge that are relevant to other contexts where the cluster approach is being tried.

The emergence of the Cluster Approach

In 2005, the UN Under Secretary General for Humanitarian Affairs, Jan Egeland, commissioned the Humanitarian Response Review (HRR) to address failures in the international response to humanitarian crisis. The review, published in August 2005, found that the speed, quality and effectiveness of humanitarian responses was inadequate, and that no common basis existed for assessing and comparing levels of need. Levels and techniques of funding were also found to be inadequate. The HRR report aimed to address identified weaknesses in accountability, predictability and reliability by nominating organisational leaders for areas in which there was an identified gap in humanitarian response. Under this approach, these organisations would then be responsible for specific areas, or clusters.

Nine clusters were subsequently defined under three broad headings: Service Provision, which encompassed logistics and emergency telecommunications; Relief and Assistance to beneficiaries, which covered emergency shelter, health, nutrition and water hygiene and sanitation; and Cross-Cutting Concerns, with three clusters: early recovery, protection and camp coordination and management. Other sectors which the UN considered had clear leadership and accountability, such as agriculture (led by FAO), refugees (UNHRC-led), education (UNICEF) and food (WFP), were not included in the cluster approach.

The intention of the Humanitarian Response Review proposal was that clusters would be set up at both the global level – in order to strengthen preparedness and technical capacity across the humanitarian system – and at the country level, to strengthen coordination and effectiveness and encourage the participation of organisations across the humanitarian spectrum, including NGOs and other non-UN actors.

Main findings of ActionAid’s research

Following the implementation of the Cluster Approach in the aftermath of the Pakistan earthquake, ActionAid undertook research interviewing local and international NGOs, UN organisations and cluster leads, donors and affected communities. The interview questions were formulated after a desk-based review of selected meeting notes and reports on the earthquake response, progress on the HRR agenda and the on-going implementation of the Cluster Approach. ActionAid carried out interviews over a three-week period in February and March 2006 with ten local NGOs and members of affected communities, officials from six donor agencies, six UN organisations and cluster leads in several locations, both in the capital Islamabad and in the provinces, ten international NGOs (both in the capital and in regional hubs). We also consulted with local government officials. ActionAid researchers attended cluster and coordination meetings in Islamabad, Batagram and Mansehra, as well as camp management meetings in Muzaffarabad.

The Pakistan response marked the first implementation of the Cluster Approach. In broad terms, ActionAid’s research shows that the earthquake response was regarded as effective, particularly as the anticipated second wave of winter deaths was avoided. Key factors affecting the success of the response were the high level of cooperation from the Pakistan government and the relatively mild winter. It is unclear how much of a difference the Cluster Approach itself made.

Within the first 24 hours of the response, nine clusters modelled on the HRR recommendations were established in Islamabad (a tenth cluster covered education), and field cluster sites, dubbed ‘humanitarian hubs’, were set up. However, as the emergency progressed the number of clusters and sub-clusters grew exponentially, making it difficult for NGOs to keep track of them. There was also confusion among both UN and NGO staff as to what the Cluster Approach was about, primarily because it was still being developed at the global level when it was introduced in Pakistan. As a result, those implementing it 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, but the cluster approach was being implemented in Pakistan well before these were developed. Clusters were credited with providing an opportunity for information-sharing and coordination. International NGOs and donors reported that having a named agency responsible for coordinating efforts in a particular area was helpful.

Adequate attempts were not made to involve local NGOs and governmental structures. Local NGOs complained that cluster meetings – always held in English – did not pay sufficient attention to the ideas and issues they raised. Others regarded the meetings as talking shops and preferred to spend their time in the field. The UN made very limited efforts to involve local democratic structures, which were sidelined in the humanitarian response.

In assessing the successes and limitations of the Cluster Approach in Pakistan, we take as our starting point the issues which the HRR sought to address: improving the speed, quality and effectiveness of response, and enhancing accountability and predictability and leadership. Our research found that performance varied widely from cluster to cluster. One international NGO commented that clusters got ‘so bogged down with the mechanics they completely lost their focus’. Some clusters were carried forward by the charisma of their lead, while those oriented towards ground-level work, such as logistics, food and shelter, were more successful. Some clusters – those focused on future livelihood strategies and irrigation, for example – were regarded as not immediately relevant to the relief effort. Attendance at the Health, Education and Watsan clusters was low at the beginning of the response, and Livelihoods and Protection had a small membership throughout.

We also found that clusters were hampered by problems of communication between field staff and decision-makers based in Islamabad. High staff turnover inhibited the development of institutional memory and made it difficult to develop relationships with stakeholders. There was a lack of back-up support for cluster leads, who were essentially taking on two full-time roles, an agency role and a separate role as cluster lead.

There was a general feeling among NGOs that clusters were overly compartmentalised, and that there were too many meetings. Although the HRR had originally proposed the cluster approach as a way to fill identified gaps, in Pakistan it appears to have been expanded to encompass the humanitarian response as a whole. Some NGO interviewees complained of ‘cluster creep’, and reported duplication and overlap. NGOs felt that UN agencies treated them simply as implementing partners, or actors to be ‘policed’, and did not allow them adequate input into conceptual thinking.

Respondents also noted insufficient analysis, synthesis and thinking ahead within meetings. Clusters provided some support on technical matters, such as shelter design and heating arrangements, but this was inadequate. Too much time was spent on agreeing guidance as to what kind of support to provide, leading people to bypass clusters and decide for themselves what response was necessary. There was also a lack of monitoring and evaluation.

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, but reported that it had not affected their attitudes to funding the UN. A number of donors stressed the need for continued support to the Office for the Coordination of Humanitarian Affairs (OCHA).

Issues to address

ActionAid’s research identified the following issues that need to be addressed.

  • Engagement with local democratic structures. Clusters must work to empower such civil structures as exist at the most local level after a major emergency. From the very beginning, clusters should have a clear exit strategy, through the identification of key actors in local authorities. Cluster leads must provide as much support as possible 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 stages.
  • Encouraging local organisations to attend cluster meetings. Inter-agency Standing Committee (IASC) Country Teams need a clear strategy for local involvement. OCHA, the Humanitarian Information Centre and cluster leads need to identify key local actors, and the monitoring of a cluster’s performance should include assessment of attempts to involve local communities and civil structures.
  • Language. Having interpreters at cluster meetings is an important means of enabling local community groups and citizens to engage with the process, and assists in capacity-building.
  • Encouraging the involvement of non-UN organisations. UN agencies need to make clear to NGOs, particularly local NGOs, what the role of clusters is, and how NGOs can interact with them. A more participatory approach is needed, particularly in the conceptual stages of planning. NGOs must be treated as genuine partners rather than simply as implementing agencies.
  • Donors/funding. The role of donor agency representatives within the cluster process should be clarified, and international financial institutions such as the World Bank or the regional development banks should be encouraged to attend cluster meetings.
  • The media. The media needs to be better utilised to spread awareness of the clusters and to carry out coordination activities. The UN needs a bold, centrally devised strategy which can be implemented and adapted by IASC Country Teams according to the needs of the emergency and the particular country in which it occurs.
  • Structural issues. A number of structural problems need to be addressed, such as the distinction between the duties of heads of agency and cluster leads, the provision of training for clusters on their roles and responsibilities, and how to avoid fragmentation into sub-clusters.

Both OCHA and the IASC have worked to clarify the anomalies in the cluster system and provide better guidance to staff and other actors. A Preliminary Guidance Note on cluster leadership is due to be approved, after consultation, by the IASC meeting in December. The minutes of the meeting say that the IASC ‘Agreed to provide continued support for the implementation of the Cluster Guidance Note, taking into account the need to monitor progress and report back to the IASC on cluster implementation at both global and field level’. Generic Terms of Reference for cluster leads have been developed. OCHA has provided its staff with ‘specific preliminary guidance on carrying out their role in support of the cluster approach at the field level’, and at the global level individual cluster leads have set up working groups to address priority issues. In addition, self-assessments of the Cluster Approach have been carried out in the roll-out countries (the Democratic Republic of Congo, Uganda, Liberia, Colombia and Somalia. A Desk Review of more than 50 documents from UN and non-UN sources was carried out by the Humanitarian Reform Support Unit (HRSU).


The UN has come a long way in developing the Cluster Approach since 2005. There are many positive aspects to it, including increased opportunities for collaboration and dialogue, improved coordination and greater predictability. However, several of the problems we identified in the Pakistan response remain. In particular, we are concerned about inter-cluster coordination. Our research in Pakistan found that, in some cases, participants found the increased demands due to the need for inter-cluster coordination to be a burden rather than a help in ensuring adequate and timely responses, in managing information and avoiding both gaps and duplication. Local and national NGOs still find it difficult to understand and engage with the clusters.

On a more general level, these issues relate to the role of the IASC Country Team and its interaction with Humanitarian Coordinators, Emergency Relief Coordinators and cluster leads. According to the IASC guidance note: ‘The establishment of the IASC Country Teams in all countries with Humanitarian Coordinators should help to strengthen partnerships and complementarity amongst humanitarian actors, in line with the overall aims of the cluster leadership approach’. Yet an in-country investigation in Uganda in May 2006 found that humanitarian coordination was woefully inadequate. Although the IASC Country Team has been operational since 2004, it has not been able to develop its role as a forum for discussing strategic issues and humanitarian policy. The same report also noted that each cluster was carrying out its own needs assessments, capacity mapping and strategic planning.

As the international humanitarian community, led by the UN, moves forward with improving and refining the cluster approach, the findings of agencies such as ActionAid and the experience of our partners on the ground must be taken into account in building a timely and reliable response system which truly responds to the needs of the people it is attempting to serve.

Anne Street is a Policy Analyst in the International Emergencies and Conflict Team at ActionAid. Her email address is Gita Parihar is a UK-based consultant and author of the ActionAid report on which this article is based. Fawad Khan, a Pakistan-based consultant, conducted some of the interviews for the report.


Comments are available for logged in members only.