This note was written in February 2011 as a personal reflection on the earthquake that struck Haiti on January 12, 2010, and the lessons it could provide for Nepal.


The first main lesson of the Haitian experience in 2010 is that after widespread destruction of such magnitude, everyone and every organization is overwhelmed. The response is chaotic and it is almost impossible to coordinate or target. A major consequence of this is that organizations become introverted when faced with a disaster of a scale that they cannot fully comprehend. Much time and effort is spent on “restructuring” or working feverishly with only a partial understanding of the context in which they are operating and with little idea of how this fits in to the overall effort. What made the immediate emergency response in Haiti “successful” was that it was so massive that almost inevitably all the affected population received minimum life-sustaining aid. But the relative chaos of the relief response has meant that it is still going on, and the “reconstruction and development” phase has not kicked in. Difficult though it is in such circumstances, clearly the seeds of reconstruction need to be sown as the emergency effort is beginning and the follow-on should emerge “organically” from the relief effort.

My general conclusion of the first year after the quake is that a major problem is that humanitarian relief is basically top-down while reconstruction has to happen with a participatory, bottom-up approach, i.e., community-based. Somehow these two approaches need to be reconciled. As much as possible, this should be done pre-disaster, in calm. Level-headed policy making proved extremely difficult in post-quake Haiti. Short-term priorities tend to dominate “early response,” which often does not flow easily into (and may even hamper) “early recovery.”

Planning needs to assume that communications will be poor or non-existent for several days. This has a number of ramifications, notably the difficulty of coordinating relief efforts or of conveying arrangements to the affected population. It also makes it difficult for relief actors to communicate among themselves, restricting policy-making capacity. In the general frenzy, coordination with national and local authorities is often given insufficient priority.

This note is not a scientific or systematic review but is a series of observations by someone close to, but not part of, the relief effort. It is not meant as a criticism of the relief effort. The magnitude of the disaster was, as I say, overwhelming and it is unrealistic to expect the international community to be able to have a fully prepared adequate “stand-by” capacity to be mobilized immediately in response to a natural disaster of such dimensions. It is for this reason that organizations, even those with contingency plans and a knowledge of best practices—which only covers a percentage of relief organizations—have inevitably been obliged to improvise and adapt their normal practice as the response evolves.

One of the major policy decisions that should be determined is who will be entitled to relief aid and compensation, and on what basis. The behavior of internally displaced persons (IDPs) and the affected population is to a degree determined by the perceived assistance to which they hope they could be entitled. Often this behavior, therefore, is based on rumor and may hamper relief strategies. In Haiti, for example, many of those who moved to the more than a thousand IDP camps were not directly affected by the quake in the sense that they did not lose their homes. This is quite simply because they were slum dwellers who had no hard-wall accommodation prior to January 12, 2010. It is my estimate that this makes up at least half of the so-called “IDP population.” Thus, the primary challenge for relief workers is to develop a profile of the affected population. It is also urgent to quickly develop a common understanding within the humanitarian community of the priority target groups based on a systematic appraisal of “vulnerability.” This understanding of the nature of the affected population is important as the group will be heterogeneous whereas humanitarian response often assumes a homogenous population in need of assistance. These differences will determine the nature of the incentives required for various groups to participate in the “re-normalization” of the post-disaster situation.

One important policy puzzle is to decide how to deal with those who lose property—owned or rented—during the natural disaster and those who were “homeless” or living in “precarious settlements” prior to the quake. From a “needs or rights-based” approach, the vulnerable and homeless should be dealt with based on their current situation, not on how they got there, i.e. through the impact of the earthquake or as a result of longer-term poverty-related issues. In the case of Nepal these would inevitably be related to issues of caste and ethnicity as well as gender. Such issues need to be worked through prior to the disaster given the difficulty of developing policy in the frenetic post-disaster context.

Among the priorities for expectation management involves the views on timeframes. Recovery and reconstruction almost always take much longer than the affected population would imagine. An idea of the timescales can be provided by the experiences in Aceh and Sri Lanka and other areas post-tsunami. The slowness of reconstruction in Haiti is a harsh lesson of the difficulties of working in coordination with a weak state. It is also worth looking at historical examples, such as the Nicaraguan quake of 1972 where the center of Managua still remains in ruins. Any use of the post-tsunami experiences must take into account the peculiarities of the geographical impact of the tsunami (a coastal fringe, never far from an unaffected area). This is very different from the geographical impact of an earthquake. Crucially, also, the tsunami-affected areas enjoyed the support of unaffected capitals. The predicted quake in Nepal would probably share with Haiti the extremely negative factor of an affected capital which, in turn, means that many potential national first-responders are themselves victims.

With the withdrawal in April 2010 of the majority of international military forces that had responded to the quake in Haiti, the approach became purely humanitarian with the IDPs being seen as victims requiring assistance but having no actual voice to determine the nature of that assistance. In the latter months of the year, as the emphasis moved toward reconstruction, the IDPs became seen more as communities, as members of a neighborhood. They are now seen as a social movement or social problem, according to the perspective, particularly those who are living on the many public squares and other sites such as sports grounds. In the case of Nepal, it would be desirable to, from the start, take a multi-dimensional, multi-disciplinary approach rather than a sequence of one-dimensional approaches.

Two families share a tent in Tundikhel, Kathmandu after the earthquake that struck Nepal on April 25, 2015. Pramila K.C.

Determining where people should go

One of the main lessons from Haiti was that, in the days following the quake, some 600,000 people fled the capital for their hometowns or other areas they considered safe. These figures (unlike many others) are reliable as they are based on the data from the main cell-phone provider who could track the movement of the phones. There should be a plan, therefore, with the main cell phone providers in Nepal to provide a similar service. UN operations in Nepal could be in touch with the university students from the Karolinska Institute and Columbia University who worked on the raw cell phone data to track movements. According to them, some 80 percent have now returned to Port-au-Prince. This represents a missed opportunity: the capital is vastly overpopulated and simply does not have the infrastructure to support the current population. Even before the quake, the infrastructure was insufficient, with some two-thirds of the population living in slums and depending on the informal economic sector. Donors and humanitarians should have targeted host families and communities to give them the support to allow those who left the capital to stay in their new location. Instead nearly all aid went to areas physically affected by the quake. Support could range from basic emergency food aid to more durable solutions such as the provision of health and education and even job creation. This would have been a major step toward decentralization, of which there is much talk but little action.

It is predictable that following a major earthquake in Kathmandu, a significant percentage of the population would leave the capital, mainly returning to their areas of origin but potentially also simply going to the nearest unaffected towns. As in Haiti, it is likely that all international relief efforts will target exclusively the physically affected areas. Pre-disaster sensitization should be carried out with relief and humanitarian organizations to actively ensure support to host families and communities to delay the return of IDPs to the capital, which will have limited capacity and infrastructure to receive them. This could range from basic emergency food aid to more durable solutions such as the provision of health and education, and even job creation.

The geographical relocation of the displaced is, to a significant degree, determined by the policies of humanitarian relief providers. For many relief providers, distribution is simplest if the displaced population is gathered in large camps, reducing the number of distribution points. On the other hand, the creation of camps tends to lead to dependency on aid, and there are many other social problems associated particularly with the creation of large camps, which have a tendency to become semi-permanent. If large-scale camps are seen as undesirable, what measures or incentives can be implemented to either keep people outside the capital, i.e. in their place of first relocation? It should be noted that in a situation of large numbers of IDPs, large-scale secondary and tertiary relocations are sure to take place. For example, if there are a number of IDP camps set up, a family may disperse its members to receive assistance in more than one location or a family that has fled to the provinces may send back one or more breadwinners to the capital to rejoin the informal economy if there is insufficient incentive to remain in their new location.

So while locations are hard to predict, and depend on the precise impact of the quake, it is possible to develop policies prior to the disaster on the model that the donors and humanitarians would like to promote, such as measures to keep communities together with a view to an early return to their place of origin or environs. Given the high-rise, high-density housing in most of Kathmandu, it should be borne in mind that the erection of temporary shelters on collapsed property is contingent on the removal of rubble. Also, for the interim period, these sites will only be able to provide shelter of one story, so not all the residents can be housed on the site of their previous multi-family dwellings.

Registering IDPs

The registration of a dispersed population of around one million people in over one thousand camps proved to be practically impossible to achieve in Haiti in a reliable and timely fashion. Such registration would have provided data for planning and policy-making. In a situation of this scale, it would perhaps be more useful to devise some method of sampling that would provide a predetermined data set without a futile attempt to do comprehensive registration. If the registration process takes much longer than a month, the data becomes redundant due to secondary and tertiary relocations. At the same time, it would be useful to define the key information required to obtain a profile of the make up of the IDPs as part of the design of incentives and remedies.

Linked to this is the development of a common set of assessment priorities with a shared and agreed methodology (this could be based on existing methodologies such as UN Disaster Assessment and Coordination, etc.). At the same time, it is crucial to carry out a rapid, commonly accepted, survey of vulnerability to allow prioritization of target groups, particularly in case of secondary disasters (monsoon, disease, etc.).

Supporting host communities and families

The physically affected areas of an earthquake exercise an almost hypnotic effect on relief organizations. The needs are so glaringly obvious and overwhelming that it seems something of a diversion to look at non-affected areas. But in the case of a disaster hitting the capital or major city, there are repercussions for the whole country. Central government administrative and other service delivery is inevitably weakened.

More importantly, as stated previously, many families will flee the site of an earthquake. Those who do become almost entirely dependent on host families and communities that often have few spare resources and whose coping mechanisms are rapidly exhausted. It is in the overall interest of the recovery effort to keep these people in their new locations at least until infrastructure problems at the quake site have been resolved. In the case of Haiti, this will probably be between five and ten years.

Support to these groups requires rapid identification of movement (best done through mobile phone records). In the case of Nepal, chief district offices and local development offices should be in a position to provide reasonably reliable data, though self-interest will often lead to the provision of inflated figures in an attempt to maximize resource mobilization. Support to host communities and to smaller camps at the quake site requires a humanitarian response with a mobile capacity to allow as many distribution points as possible.

A family’s belongings sit outside their damaged home in Chautara. Gyanu Adhikari/The Record

Avoiding organized mass relocation 

Despite the logistical attraction of large camps (a corollary of which is the temptation to set up new sites to receive large numbers of vulnerable groups), the Haiti experience suggests that smaller camps near to the place of origin of the displaced are often a preferable solution. In the short term, if communities are broadly kept together, there is a reduced risk of sexual and gender-based violence and other forms of violence and anti-social behavior. In the medium term, these smaller more cohesive units are more likely to lead to recovery on a neighborhood-based approach. It should also be easier to reestablish basic services, including education, if a community or neighborhood approach is adopted. These services are important for immediate survival but also speed a return to “normality,” notably, for example, the return to regular schooling.

Determining shelter provisions

The last year in Haiti has seen a lively, on-going debate on shelter options. While a uniform policy on the provision of shelter would be optimal, it is notable that even the various national sections of the International Federation of the Red Cross present in Haiti all offer different shelter options. The general view is that t-shelters (temporary or transitional) are the best option, though some point out that “permanent” solutions cost little more than temporary shelters. It should also be kept in mind that funding will be reduced year on year as donor fatigue sets in. This is already the case in Haiti where the November 2010 UN appeal for funds to combat the cholera outbreak has had great difficulty in fund raising despite high-level UN support for the appeal. [May 1, 2015: Although precise figures are not available, in 2010 there were estimated to be 1.5 million people who had lost their homes in Haiti. Roughly 300,000 new units would have been required to meet this deficit. The latest available figures suggest, however, that only 10,000 have been built. Consideration should be given to whether or not it is feasible to support self-build community-based programs.]

Explaining what “packages” people can expect

In the post-disaster period, the decision-making of the affected population is in part determined by their anticipation of aid packages or incentives to relocate, either to their places of origin or to new sites. The arrival of numerous NGOs offering aid creates a great deal of expectation, some, but not all, reasonable. These expectations are exacerbated by the communications vacuum that is almost inevitable in the days following a quake. This creates a context in which rumors are generated about what assistance or “compensation” will become available. In the Haitian case, with no basis, many in IDP camps still expect to benefit from land titles. The anticipation of significant financial packages or the distribution of land titles has been a pull factor to keep people in the IDP camps, even many whose houses were only marginally damaged. One sector that requires special attention is renters. Often landlords take the opportunity to raise rents after a crisis making them possibly unaffordable to previous tenants.

Preparing for land register issues

The Haiti experience demonstrates the negative effect on reconstruction planning of a poor and incomplete land register and the widespread existence of property disputes. This makes emergency reconstruction planning slow and complicated. One of the main impediments to reconstruction in Haiti has been the unfavorable legal context: the land register covers only a small percentage of land ownership and much of it is contested. The government was slow to carry out compulsory purchase (eminent domain) on land that could be used for new settlements.

It would be useful in the Kathmandu valley to identify available land for potential emergency settlement and to verify its status. This would be a register of potentially useable land and would include unused land and land currently used for agriculture. A further recommended action would be the prepositioning to the degree possible of essential infrastructure, for example, sanitation, water, electricity, access roads, etc., or planning toward how these would be organized post-disaster.

Efforts are underway to begin removing the rubble from Kathmandu Durbar Square. Gyanu Adhikari/The Record

Dealing with rubble

Although it is obvious that earthquakes produce rubble, there are some lessons to be learned from Haiti. For example, the sheer scale of the quantity of rubble produced, though predictable, is still overwhelming. Oxfam and the UN Office for the Coordination of Humanitarian Affairs recently reported that in the first 12 months, only some 5 percent has been removed. This is difficult to quantify with any precision, however, as even the overall estimate of 20 million cubic meters of rubble is far from scientific. Recently some have been suggesting that in fact the percentage removed is actually greater, up to 20 percent. The main point is that rubble prevents other activities and it is an urgent priority to identify which rubble is the most obstructive and to remove that first. Some rubble just sits there, not affecting anything, so that can wait.

Planning should consider what legal provision is needed for rubble clearance teams to work on destroyed buildings, and how permission for demolition or rubble removal will be obtained. Even a flattened building is owned by someone and nearly all rubble has an “owner.”

There have been great difficulties in identifying sites for rubble dumping. Much time has been lost identifying the best technology for rubble crushing. No preparatory work was done to look at the challenges of recycling rubble—for example, whether or not it is toxic, and what purposes it could be used for. There are many useful ways of using rubble, for example, river control, flood avoidance, road building, etc.

Any decisions on rubble recycling, however, depend on a determination of whether or not it is toxic. The most appropriate technology for the rubble disposal needs to be decided as quickly as possible. The removal plans for rubble have to take into account that the “production” of rubble in an on-going process. Only a small proportion of buildings damaged beyond repair on January 12, 2010 were “destroyed,” and the demolition (largely by hand) of damaged buildings is far from complete one year after the quake. It is predictable that ongoing demolition will be creating more rubble for years to come.

Cash for work policy

How will cash for work be used given the lessons learned from Haiti? Large-scale funds have been dedicated to cash and/or food for work. Certainly in the initial period after the quake the objectives and criteria for these programs was hard to perceive. This is clearly a useful instrument, but it would be worth carrying out a rigorous evaluation of the various initiatives in Haiti. It is obviously a good way of getting some cash into the hard hit economy via some of the poorest in the community. However, the selection of the participants and the tasks they are set are often unclear. A year on, it is common to see groups of several dozen “cash for workers” clearing roadside drains or similar. They are easy to spot in their branded t-shirts but usually it appears that fewer than 10 percent are actually working at any given time. Some who run these programs privately admit that in the initial phase things were disorganized but now improvements have been made. In the case of Nepal, it would be important to decide in advance who would qualify to take part, what is an appropriate remuneration, how to manage the work teams (they require some management and coordination), and how to avoid the oft-reported phenomenon in Haiti of gangs taking control of cash for work programs and creaming off significant sums. In Nepal this would probably be the case of political party appropriation rather than criminal organization. It would be worth reviewing previous experiences of cash transfers and cash/food for work in Nepal. I imagine the key example would be the Koshi flooding of a few years back.

It might be worth researching traditional forms of community cooperation and working patterns to see if cash for work programs could be adapted to local custom and practice. It is certainly necessary to ensure that these programs are “purposeful” rather than “make work” programs. One obvious way to do this is to take a neighborhood approach with groups of residents paid to carry out initial recovery and reconstruction work even if it is as basic as rubble removal or latrine building.

Potential transport issues

Contingency plans should be made in case Tribhuvan International Airport (TIA) is damaged. It may be necessary to examine the feasibility of upgrading Simra or Pokhara airports to receive larger planes if TIA is damaged. It should be taken into account that the transport network connecting Kathmandu to the rest of the country and its neighbors is very limited and therefore represents a potentially serious impediment to aid delivery. In practical terms, I would suggest that any work done on TIA, particularly with international support, should take disaster risk reduction into account, which of course will add to the costs.

Potential political impact of the disaster and response

Most large-scale disasters have political impacts that are difficult to predict. In Aceh the tsunami clearly contributed to the peace-making process, whereas it is arguable that the same disaster contributed to a return to conflict in Sri Lanka as both sides sought to use the inflow of massive amounts of international aid to their political advantage. It is argued by many that the poor government response to the Nicaraguan quake of 1972 contributed to the demise of the long-standing Somoza dictatorship seven years later, and there is little doubt that the low-profile response of the Haitian political leadership contributed to the rejection of President Préval’s candidate in the election of November 2010. If an earthquake occurred in the polarized political context of today’s Nepal, any aid effort would have to understand and take into account the political dynamics and should not use the issue of humanitarian neutrality to justify an ignorance of the political and cultural context.

The nature and scale of a future quake in Kathmandu and the political context in which it occurs could require the contemplation of some provisional governance arrangements if the government of the time becomes dysfunctional. This may require a government of national unity.

The polity of Nepal is much more articulated and structured than in Haiti where political parties have played little or no role. Discussions in advance with political parties would be desirable.

Magnitude of the disaster and impact on relief planning

One key lesson from the immediate aftermath of the Haitian quake is that its magnitude made planning and policy-making extremely difficult after the event. This made emergency relief targeting extremely difficult. Planning for a disaster of this scale needs to include this problem in its assumptions, and targeting needs to be done prior to the event. This is a challenge given the endless variables in a disaster of this scale, but everyone needs to know more or less what they are going to do in the days following the quake and adapt those standing operational procedures (SOPs) themselves, individually or as organizations, in the light to the actual event. These SOPs need to be ingrained so they become automatic reflexes—people need such guidance as a handrail given the traumatic effect of a quake.

Coordinating the NGO sector to maximize impact

The Haiti quake provides a useful insight into the problems of coordination of relief. The sheer quantity of NGOs and other organizations that pitched up, many with no contextual knowledge of Haiti, much less experience of work there, made coordination extremely challenging and almost certainly reduced their impact. Many NGOs simply do not want to be coordinated. OCHA never really got on top of the situation to the extent of being able to give a useful overview, so was not an attractive partner for many groups.

Anecdotal evidence suggests that NGOs proved to be very reluctant to act with transparency with national and local authorities, which led to repeated complaints from the prime minister who is also the Minister of Planning. In the case of Nepal, it would be worth investigating whether there is any way of predicting which NGOs will be present after a quake. Related to this is the issue of potentially pre-assigning either thematic tasks or geographical responsibilities in advance. There should also be a pre-agreed policy on NGO “registration” and what that might consist of. (It is worth consulting the recent report of the US-based Disaster Accountability Project, which was highly critical of the lack of NGO transparency.)

In Haiti, the cluster model seems to have worked effectively in the regions where those working in any particular sector were a manageable number (10 to 15) whereas the sector clusters in the capital turned into assemblies with 50 to 70 participants not being uncommon. While some level of information exchange took place, the clusters were inadequate as coordination or decision-making forums.

It is important to develop incentives for relief organizations to participate in coordination. This would ideally come about by leadership on information management. However, it has to be realized that many of the organizations have their own institutional interests and will target groups and areas which feed back into their fundraising base.

In the case of Haiti, it has been a clearly more attractive option for many organizations to work with obviously vulnerable groups in live-saving work rather than, say, working on essential but “boring” issues such as rubble removal.

Moreover, the functioning of the cluster system clearly depends on subjective factors, in this case the leadership qualities of agencies assigned to be focal points for each sector. It may be worth considering ways of preparing UN agencies for this role.

Managing donor institutional imperatives

Donor and NGO expectations need to be managed. Their institutional imperatives push them to spend the funds raised for a specific disaster in a timely fashion. This may be a contributory factor in the transition from emergency response to reconstruction in that relief efforts offer ways of disbursing funds at a time when the lack of clear government or international community policies leads to a dearth of opportunities for spending on medium to long-term solutions. In Haiti, one major agency has been trying to disburse large sums through direct and unconditional cash transfers to avoid accusations of failure to spend funds raised, a move that was widely opposed by the international community and government who preferred such measures to be linked to return to safe housing or registration of children in schools. Some suggested that the agency involved wanted to avoid a repetition of accusations of failure to spend after a previous high-profile disaster.


Apart from the available UN documentation of the past year, I would recommend the Oxfam Briefing Paper of January 6, 2011, “From Relief to Recovery.” ActionAid have also produced a report, “Building for the future – homes and security in Haiti 1 year on,” but I am more confident about the data in the Oxfam report.

Cover photo: People survey the earthquake damage in Kathmandu Durbar Square. Gyanu Adhikari/The Record


Updated: Saturday, May 2, 2015
This piece was updated to inlcude a section on cash for work policy from the author’s 2011 note.