disaster management

What can the development industry learn from disaster responses?

Graeme MacRae is not an expert on disasters or development – just an ordinary anthropologist who first visited Indonesia in 1977 and has conducted research there since 1993. He spent several weeks in Jogjakarta shortly after the earthquake, observing the international response as well as visiting affected villages and listening to local people. Back in New Zealand he spends most of his time teaching at Massey University in Auckland. The story he tells here can be read in more detail in Development in Practice, Volume 18, Issue 2, 2008 , pages 190 – 200.

Disasters happen – seemingly with increasing regularity (in Burma as I write, then China as I revise) and ever-larger consequences. When they happen, the international humanitarian relief system (IHRS) moves in to help. The IHRS is not the development industry, but they have a lot in common: they work on similar ideological principles, use similar methods and many agencies work across both fields. It is useful to consider the IHRS as a sub-section of the larger industry, in which we can see its practices and principles at work in a particularly concentrated form – stripped down to the basics without a lot of their usual packaging. Studying the IHRS in action can provide useful insights for thinking about the development industry as a whole.

The Indian Ocean tsunami in December 2004 led to the biggest ever response from donors and thus to the biggest and best funded relief and reconstruction operation ever. It also led to some very bad practice and bad results as well as gross wastage of money – reminiscent of the most spectacular failures of the development industry. The earthquake in Jogjakarta a year and a half later (May 2006) was less spectacular, less costly in terms of human life, and attracted much less media and public attention as well as much less funding. In terms of housing destroyed however, it was actually larger – over 300,000 houses destroyed and over a million people left homeless.

The nature and scale of the problem were fairly clear within days of the earthquake and transport and communications infrastructure remained relatively intact – a large emergency but not a complex one. The priorities were also clear, to local people and to anyone else who bothered to look and ask – get shelter over people’s heads as fast as possible. The initial, emergency phase did happen fairly quickly – 200,000 tents and tarpaulins were distributed in the first few weeks. But everybody knew these would not be adequate for the rainy season and that more reliable shelter was needed - and fast. However, by the time the rains came in November, very little had been built and hundreds of thousands of people still had very inadequate shelter. While this was not a complete failure of the system it was a partial failure – too little, too late – and it could have been avoided.

The reasons for this slowness were multiple and complex. Firstly, because donor funding was much less than for the tsunami, the resources were simply insufficient for the magnitude of the task. Agencies were also frustrated by donor conditions and timelines which were not always compatible with what they felt needed to happen on the ground. Secondly, both the Indonesian government and international agencies who had their fingers burnt in the tsunami operation, had lost confidence and moved very cautiously, even defensively and thus slowly. Furthermore the conclusions that the government and the IHRS drew from the failures of the tsunami response were rather different, so their approaches to the shelter problem moved, at least in the early stages, in opposite directions. When they finally did manage to coordinate their efforts, the demand for materials and the logistics of such a huge rebuilding programme were obstacles to speed. IHRS workers I spoke too were unanimous in their assessment that the response had been relatively successful, compared with other disasters, especially the tsunami. However they also acknowledged that there were problems of coordination between over 100 agencies with their own agendas, policies and methods and this was the main focus of subsequent reports on the response.

All these factors undoubtedly contributed to the delays and some were difficult if not impossible to avoid. However, my observations of the response, in the critical stage when it was moving the emergency phase into reconstruction, suggest that the “too little, too-late”-ness of the response was also a direct result of the way the system itself is organized and operates.

The IHRS consists of resources and expertise embodied in thousands of people in hundreds of organizations dispersed all over the world. They (collectively and to a large extent individually) see disasters as essentially similar sets of problems (shelter, health, water and sanitation, etc.) which can be analysed and addressed in essentially similar ways. In other words it is a global system based on a set of standard assumptions and universalised knowledge. All that needs to happen is for it to be mobilised, get the scene fast, and move into action.

In fact, when the system arrives at the scene of a disaster the reality it encounters is very local and very specific – in terms of geography, environment, history, politics, culture, social organization, religion and language – not to mention people. One of the most extraordinary things about the system, is that this is not even seen as relevant, let alone a problem. Advertising for field staff for IHRS organizations rarely even mentions local experience, knowledge or language skills. Of the hundreds of international aid workers who arrived in Jogjakarta, very few had any of these. Consequently when the system arrives, with all its skill, experience and resources, the one thing it does not know is how to communicate with local people, communities, organisations or government, with the partial exception of those who happen to speak some English. But this is not seen as particularly important, because the real priorities seem to be setting up offices, data collection and analysis systems, logframes, maps and matrices – then meetings and systems of communication – between agencies – all in the city and all in English. If local organisations are involved, they usually have to fit into this system as best they can.

But what about the ruined villages just down the road – and the thousands of people huddled under their blue plastic tarpaulins? Many aid workers rarely if ever get to meet them. The interface between the system and the people it is supposed to serve is mediated by what are known as “partnerships” with local NGOs, who know the local scene and how to talk to people. This is in theory not a bad idea, adapted from current practice in the wider development industry. But even here it is not without its problems, because of the widely varying priorities, styles and capacities of local NGOs, and real cultural, social and often political gaps between them and on one hand the international organizations and on the other, local communities. Partnerships take time and hard work to build and maintain at the best of times. In the worst of times - the overheated, chaotic, time-scarce conditions of a disaster response, the chances of success, let alone instant success, are much less.

As a result, in Jogjakarta, weeks, even months after the earthquake, most IHRS staff knew little about the people they were there to help, and the people in the villages knew just as little about the IHRS. Some partnerships worked smoothly and efficiently – but many did not. This fundamental gap of different worlds, world-views and communication between them was a major obstacle to accurate understanding, and seriously delayed decision-making and hindered effective action. It also leads us directly to another weakness of the system - scale.

Concepts such as “participation” and “community-based development” are well established within the development industry, for the good reason that evidence tells us (like commonsense) that things work better when they happen at the level of real communities and people have a sense of knowledge and involvement in them. However, like “partnerships”, these do not happen automatically or quickly – they take time and local knowledge to build. The former is a scarce resource in post-disaster conditions and the latter is even scarcer in the IHRS. So, in practice, in Jogjakarta, resources for construction of both temporary and permanent shelter, were delivered mostly via impersonal, bureaucratic chains from donors to international NGOs, to local NGOs and/or local government to local leaders, to communities and eventually to households. Like the communication gaps, this too led to complication, high costs and ample opportunities for inefficiency, corruption and delays.

Much in this story resonates with the critical literature on development more generally, but it begs the usual defensive question – its easy to criticise, but is there any alternative? In Jogjakarta the answer is yes. There was at least one alternative: faster, cheaper, technologically innovative, culturally appropriate, aesthetically pleasing – and also small-scale, initiated, controlled and implemented at the level of the local community.

Ngibikan is a village of 65 houses right in the heart of the effected area. After the earthquake only one house was left standing. Three months later, all were rebuilt and occupied. The new houses were more earthquake-resistant than the old ones, better ventilated and arguably more beautiful. At this stage the IHRS was still distributing tarpaulins, collecting data and planning for the mass construction of very minimal temporary shelters. How did they do it?

Immediately after the earthquake a national newspaper launched an appeal to raise funds, then looked for ways to spend them. They contacted Eko Prawoto, a local architect with a reputation for beautiful, innovative, cost effective designs based on traditional local styles, materials and technologies. One of his building teams came from Ngibikan, so he contacted them and within days they had worked out a standard design, built a prototype of the main structure and cleared most of their sites of wreckage. The community decided to build all the houses together, organising their collective labour by a traditional system known as gotong-royong. They divided themselves into teams led by their most experienced builders, each specialising in one stage of the process, and they worked from house to house. Eko and Pak Maryono, the headman of Ngibikan, used their contacts to get good deals on good materials. Less than two weeks after the earthquake, they started rebuilding. When I first saw them a month later the frames were nearly all prefabricated, and most of them had been erected and the first roofs were going on. Once the roofs were on, the walls were filled in with bricks and joinery salvaged from the old houses supplemented by new materials. By the end of August, they were all complete, sufficient for their owners to complete them individually and move in when they were ready. At this stage few other villages had any but the most rudimentary temporary shelters. Does this mean Ngibikan did well, or did the IHRS do badly?

Certainly Ngibikan began with some advantages: adequate (but not extravagant) funding up front, with no bureaucratic strings attached or complicated procedures to access it and some good building skills within the community. But there was money in the IHRS too and many communities had building skills. According to Eko and Maryono the most important factors were the ordinary resources already existing in the village: collective, democratic decision-making, the egalitarian distribution of resources and sharing of labour as well as the absence of unrealistic expectations in terms of style and size of houses. Together these seem to have led to a sense of common ownership of and commitment to the project. The fact that the whole project took place at the “natural” scale of the existing community, in which everyone knew each other undoubtedly facilitated the factors mentioned above. What Eko and Maryono did not mention were their own roles in the process, as key people with complementary roles, one inside the community, the other outside, but both known and trusted by each other and the community, and working closely together. They are also both men of exceptional ability and commitment. Taken together, these factors enabled them to lead the community to quickly and accurately assess the situation, make decisions based on real understanding and implement them efficiently without complicated procedures. So they did very well, but this was no excuse for the (partial) failure IHRS.

What has this story got to do with development? On one hand it reads like a perfect illustration of all the ideological principles of the development industry: “community-based”, “participation”, mobilising “social capital”, etc. On the other hand, the list of factors which enabled the Ngibikan project reads like almost the opposite of what the IHRS system was doing in practice and what the development system often seems to do. So, is there perhaps a gap between between the ideology and practice of the international relief/development system? Is the centralised, globalised, bureaucratic, top-down structure of the system, the way it actually works, out of step with its ideologies of community scaled participation? If there is, maybe its time to think about changing it. There are unlikely to be magic bullets, but if there is anything to be learnt from this story is that small can be beautiful, local knowledge is critical and that working relationships between real people who know and trust each other are what makes things happen. If it is possible in a disaster situation, surely it is more possible in development ones. I like to imagine a global development system that looks a bit more like lots of Ngibikans than what we have now.

© Kim Williamson 2008