Uncharted territories

Geographical choices of aid agencies

Dirk-Jan Koch | July 25, 2007

Non-governmental development organizations are expected to focus on countries with poor governance. New evidence shows that they do not. While they do tend to focus more than bilateral donors on poor countries, they make some curious geographical choices. It seems that they too have their ‘donor darlings’.

The Central African Republic is an extremely poor, landlocked country where around three million people live below the poverty line. Malawi also has no access to the sea, and has four million poor. One would expect these two countries to receive roughly the same volume of aid from organizations such as Care, Hivos and Plan. In reality, however, 50 of the world’s largest non-governmental development organizations (NGDOs, as distinct from humanitarian agencies) spend only about €1 million a year in the Central African Republic, compared with more than €70 million a year in Malawi.

In policy documents issued by the World Bank, the United Nations and bilateral donors such as the Netherlands and the UK, it is often stated that NGDOs are better suited to working in countries without good governance than are other development actors. Commentators such as Alan Fowler and William Easterly support this view. But are they indeed more active in these countries? Agencies like Action Aid have criticized bilateral donors for not focusing sufficiently on poor countries, implying that NGDOs have a better track record in this respect, but is that true?

Until recently, it was not possible to answer such questions because no one monitored where these organizations spent their money. However, A new database, compiled by the author, contains information provided by 50 of the world’s largest NGDOs, including Oxfam Novib, Save the Children and World Vision (see box NGDO country allocation database). It is now possible to see whether expectations regarding these organizations’ choices of countries to receive aid are reflected in reality. This article looks at the new database and tests the validity of these expectations.

Focus on bad governance

Many academics and policy makers assume that the governance environment is one of the main factors for NGDOs to decide where they go. In its influential report Assessing Aid: What Works, What Doesn’t, and Why (1998), the World Bank stated that ‘In highly distorted environments the government is failing to provide supportive policies and effective services. That is why government-to-government financial transfers produce poor results. Effective aid in such an environment often involves supporting civil society to pressure the government to change or to take service provision directly into its own hands (or to do both)’.

The United Nations shares this opinion. In its report Investing in Development: A Practical Plan to Achieve the MDGs (2005), it states that there are ‘countries that rank consistently low on civil liberties, political freedoms and human rights, while rating high on corruption, with little demonstrable will to achieve broad-based poverty reduction. In these cases, the international community can play a role in humanitarian assistance and deliver aid through NGOs’. This view is echoed in the policy documents of various bilateral donors. In Civil Society and Structural Poverty Reduction (2002), for example, the Netherlands Ministry of Foreign Affairs noted that ‘In cases of bad governance bilateral aid relationships are often underdeveloped; cooperation from civil society to civil society is the only way’. The UK’s Department for International Development states in similarly categorical terms: ‘… NGOs will need to play a more active role in providing public services in fragile states where governments are weak and direct support to governments is not yet possible’.

Fowler suggests that bilateral donors focus more on countries with good governance because numerical targets can be more easily achieved. NGDOs are then faced with the choice of whether to ‘follow the money’, or ‘to adopt strategic choices that, for example, include a focus on … less poor or less favoured countries’. In a report on Dutch government funding of co-financing agencies, Dietz  et al. argue that NGOs should opt for the latter, because that is where their added value lies – in countries that are not eligible for bilateral aid because of governance deficiencies. ‘These “bad governance cases” often demand politically more sensitive interventions, for which NGOs are thought to be better, and more appropriate agencies’.

In short, it is widely expected that private aid agencies will be involved in countries where good governance is lacking. But do they actually focus on such countries? The answer is a clear and simple ‘no’. With the best will in the world, it is impossible to find a link between the quality of governance in a country and the level of aid it receives from NGDOs. There is no distinction between the choices of countries by bilateral donors and NGDOs, in terms of the quality of governance. Thus the choices made by NGDOs contradict the expectations regarding their supposed added value in countries with bad governance. They focus on the same countries as bilateral donors.

The figures show that NGDOs spend relatively more money in countries with moderately good governance than in those with bad governance. They prefer Costa Rica to Cuba, Ghana to Togo, and Georgia to Azerbaijan. Why is this? Clearly, in a number of countries, like Turkmenistan or North Korea, the governments are so repressive that foreign private aid agencies cannot operate. But why do they not focus more on countries where they are expected to provide added value? There are, in general terms, two possible explanations for this. First, academics and policy makers are wrong – NGDOs are not at all specially suited to work in areas with bad governance – and the theory is therefore not valid. Perhaps civil society is particularly weak in such countries and there are therefore no partners available for Northern organizations. If that is true, the NGDOs are making good choices, but the policy theory is wrong. Second, the theory is correct and the situation in practice therefore needs to change. The NGDOs offer clear added value in these countries, but there is something going wrong in the translation of theory into practice, and that needs to be investigated. It is still too early to reach a conclusion on the first explanation. Some studies have shown that NGDOs have achieved fantastic results in countries with bad governance, but others reveal that they have set up unsustainable parallel systems in these countries. Because they take over tasks and staff from the government, this can sometimes lead to a weakening of the state.

One thing is clear – there is no logical alternative to aid through NGDOs in countries without good governance. To establish whether the policy theory is correct and practice must change, it is in any case necessary to look at another expectation regarding NGDOs choices: that they focus on the poorest countries.

Focus on poor countries

To what extent do geo-strategic interests influence the choice of countries to receive aid? The literature is rather critical of governments’ motives for giving development aid. Studies of the geographical distribution of the aid provided by bilateral donors show that, with a few exceptions, developing countries where they have political, economic and/or strategic interests receive more aid than countries where such interests are less evident, but which may need the support more urgently. The large volumes of aid provided by the United States to Egypt and Israel, for example, and the large-scale debt relief to Nigeria and Iraq have more to do with politics than with poverty. Such behaviour is not expected of NGDOs: they can focus on the poorest countries without having to take account of strategic interests.

The organizations themselves fuel these expectations. In a large advertisement in a Dutch national newspaper in 2006, for example, Plan Nederland announced: ‘The cameras have left, the journalists have left, the diplomats have left, the aid agencies have left, but we are still in Darfur’. In its 2005 annual report, Cordaid stated that ‘Cordaid does not only want to be active in regions and countries where infrastructure is in good shape, where there is good governance and where our investments in partners and programmes yield substantial returns. Cordaid wants explicitly to offer aid in areas where the population is on the run and terrorized, and where minimal protection is only offered by churches and NGOs’. The assumption that NGDOs focus on poor and marginal countries is so deeply rooted that researchers have only very recently started to investigate whether this is in fact the case.

A recent study by the International Monetary Fund (IMF) on this issue addressed what NGOs did with funds provided by the European Commission. It showed that they do take significant account of poverty levels, while the poverty focus of the Commission’s regular aid generally leaves much to be desired. If we look at a broader sample of organizations we see that there is a significant correlation between, for example, the life expectancy of people in developing countries and the quantity of aid per capita. Countries where people on average die younger (a poverty indicator) receive less aid from NGDOs. This does not apply to government-to-government (bilateral) aid. Recent research shows that there are large differences in the extent to which private aid agencies focus on the poorest countries. Norwegian and American agencies focus significantly more on the poorest countries than, for example, their German or Dutch counterparts. In general, however, NGDOs focus more on the poorest countries than bilateral donors do.

If this were the end of the story, the conclusion would be that private aid agencies in general succeed in reaching poor countries and that there is no cause for concern. Unfortunately, as can be seen from the map below, the reality is not that simple.

There are great inequalities in the levels of aid received by developing countries. The darker red a country on the map, the more aid per capita it received from the world’s largest NGDOs in 2005. There is a definite tendency towards clustering. Honduras, Nicaragua, the Palestinian Territories and Zambia, for example, received more than 20 times more aid per capita from NGDOs than, for example, the Central African Republic, Guinea and Yemen.

So NGDOs too have their ‘donor darlings’. Recent research provides more insight into the causes of this clustering. The factors that can lead private companies to cluster apply equally to NGDOs. They see no reason to start from scratch in a new country if they have staff in others that have already been trained, or if there are local organizations that are used to working with international agencies and can therefore fulfil their reporting obligations? If an NGDO has to choose, for example, between a country where the banks hardly function, and one where their funds can be safely deployed, the choice is easily made. In other words, why struggle in the Central African Republic if you can achieve more for the same money in Tanzania? This way of thinking results in the ‘first-mover’ problem. Everyone would be better off if aid were more equally distributed among countries, but many NGDOs are not prepared to make the first move.

There is therefore a wide gap between expectations of NGDOs and what they actually do in practice, especially with regard to their focus on countries with bad governance, but also their focus on poor countries. Here we are concerned not so much with the weaknesses in the theory as with trying to gain insight into this disconcerting discrepancy between theory and practice. Ongoing research into the geographical choices of Dutch NGDOs between 1989 and 2005 reveals two aspects that have so far received little attention: their relative rigidity or stability in terms of country selection, and the strategic nature of these choices.

Path dependency

Although NGDOs are expected to be flexible, recent studies show that there is a certain rigidity in the geographical distribution of their activities. Anthony Bebbington of the University of Manchester, UK, in an analysis of the geographical choices of NGOs in Latin America, has observed what he refers to as ‘path dependency’. By this he means that the choices made by NGDOs in the past leave a clear mark on their current choices. He has observed that, once an organization has chosen to work in a certain region, its staff become involved in the local situation and encourage the agency to continue its activities, even if circumstances change and the reasons for becoming involved in the first place no longer apply. In addition to these established interests, Bebbington has also observed that formal and informal networks evolve, and then perpetuate themselves. In this way, Northern organizations become acquainted with new partners in the regions in which they are already active, and ignore equally good potential partners in other countries. NGDOs often say that they work with organizations they have come into contact ‘by coincidence’. Bebbington calls this into doubt, and says that these ‘coincidental’ choices often arise from self-reinforcing social networks.

To return to the example of Tanzania and the Central African Republic, if you are the manager of a Northern NGDO attending an international conference, there is a greater chance that you will find yourself talking to the manager of an NGDO from Tanzania than from the Central African Republic. An analysis of the choices made by Dutch NGDOs between 1989 and 2005 has shown that they remained very stable. Countries that received large amounts of NGDO aid in the late 1980s (e.g. Peru, the Philippines and South Africa) were still among the major recipients more than 15 years later, despite the fact that the situation in these countries had improved considerably. It is of course possible to think of reasons for this, such as the predictability of aid and the desire to avoid capital destruction, but it also leads to a discrepancy between policy theory and the stability observed in practice.

Donor-led behaviour

According to critics of NGDOs, all the explanations given above for the choices they make (poverty levels, clustering, path dependency) are not the true determinants of these choices. They see NGDOs as organizations concerned with ensuring their own survival and will therefore follow loyally in their own donors’ footsteps.

One way of determining whether this is true is to imagine what you would do if you were one of these organizations. You receive 80% of your funding from one donor, who says that you are free to choose in which countries you wish to operate. If you act strategically, you will make sure that your organization is as visible as possible to the donor. You will therefore choose to operate in countries where the donor is active, and avoid those in which your donor is not interested. Between 1989 and 2005, Dutch NGDOs were in this situation: they received on average 80% of their income from the Dutch government and were free in their choice of countries. However, a comparison of the countries they chose and those receiving aid from the Dutch government shows that there is a clear link between the two. The more aid countries receive from the Dutch government, the more they receive from Dutch NGDOs (Cordaid, Hivos, ICCO, Oxfam Novib and Plan in this sample). This relationship is shown in the graph below by the ascending line.

Thus we can conclude that NGDOs imitate the choices of bilateral donors. The more an organization depends financially on its official donor, the greater is the correlation between their choices of countries. The focus of Norwegian agencies on the poorest countries reflects the choices of the Norwegian government, and the fragmented nature of German private aid that of German bilateral assistance. Countries that receive a lot of funding from the Dutch government also receive a lot of money from Dutch NGDOs. The influence is probably mutual: the choices of NGDOs also affect those made by bilateral agencies. If the Dutch government tries to stop providing aid to a country where many Dutch organizations are active, the latter may start an intensive lobbying campaign to persuade the government to continue its aid presence in the country concerned. But irrespective of who influences whom, the result is the same: NGDOs are active in the same types of country as bilateral donors. Even though there is probably some positive synergy between bilateral and NGDO aid in the same recipient countries, it is not in line with policy theory, which suggests that they should focus on different countries.

Why are the geographical choices of NGDOs described as curious in the introduction? There are many expectations for NGDOs’ country choices (which are reinforced by the agencies themselves), but these have been little researched, let alone understood. Now that, for the first time, there is a database on the geographical choices of the major NGDOs, it is clear that practice only partially concurs with theory. It is striking that the assumed added value of these organizations in countries where good governance is lacking is not reflected in how they allocate their aid. They do focus more than bilateral donors on the poorer countries, but their aid is unequally distributed. A small number of countries receive more than 20 times more NGDO aid per capita than other comparable countries. Initiatives by NGDOs themselves to make high-risk investments with the aim of redressing this inequality – such as the renewed involvement of Cordaid in the Central African Republic – should therefore be welcomed wholeheartedly.


The author wishes to thank Boriana Yontcheva (IMF), Thea Hilhorst (Wageningen University and Research Centre), Ruerd Ruben (Centre for International Development Issues, Radboud University Nijmegen, CIDIN) and Jan Willem Nibbering (Netherlands Ministry of Foreign Affairs) for their critical comments on this article.

Photo credit main picture: HH / Roel Burgler


  • 1.

    Dollar, D. and Pritchett, L. (1998) Assessing Aid: What Works, What Doesn’t, and Why. World Bank Policy Research Report. New York: Oxford University Press, p.104.

  • 2.

    UN Millennium Project (2005) Investing in Development: A Practical Plan to Achieve the MDGs . New York: UNDP/Earthscan, p.113.

  • 3.

    Netherlands Ministry of Foreign Affairs (2002) Civil society en structurele armoedebestrijding (Civil Society and Structural Poverty Reduction).

  • 4.

    Department for International Development (2006) Eliminating World Poverty: Making Governance Work for the Poor . UK government White Paper on International Development, p.53.

  • 5.

    Fowler, A. (2005) Aid Architecture: Reflections on NGDO Futures and the Emergence of Counter-terrorism , p.6.

  • 6.

    Dietz, T. et al. (2006) Evaluation of the Theme-based Co-financing Programme (TMF) of the Dutch Ministry of Foreign Affairs: Synthesis Report , p.15.

  • 7.

    Koch, D.J. (2006) Blind Spots on the Map of Aid Allocations: The Geographies of Civil Society Aid . Paper presented at the WIDER conference on Aid, Principles, Policies and Performance, June 2006.

  • 8.

    This of course says as much about the choices of bilateral donors as of NGOs. Donors spend large quantities of money in corrupt countries. This is illustrated by the billions provided in debt relief and export credits, for example, for Nigeria and Iraq. But Iraq has to be left out of most calculations, because the levels of international ‘development aid’ it receives are so high that this has a distorting effect on the statistics. The vast amounts of money thrown at these countries completely negate the growing bilateral aid to countries with relatively good governance, like Ghana and Tanzania.

  • 9.

    Neumayer, E. (2003) The Pattern of Aid Giving: The Impact of Good Governance on Development Assistance. London: Routledge. See also Levin, V. and Dollar, D. (2004) Increasing the Selectivity of Foreign Aid, 1984-2002. World Bank Policy Research Working Paper 3299.

  • 10.

    Judith Tendler described this pattern of expectations back in 1982. Tendler, J. (1982) Turning Private Voluntary Organizations into Development Agencies: Questions for Evaluation. AID Program Evaluation Discussion Paper 12.

  • 11.

    De Volkskrant, 16 December 2006.

  • 12.

    Cordaid (2005) Annual Report 2004 (Jaarverslag, in Dutch).

  • 13.

    Nancy, G. and Yontcheva, B. (2006) Does NGO Aid Go to the Poor? Empirical Evidence from Europe . IMF Institute Working Paper WP/06/39.

  • 14.

    Koch, D.J., Westeneng, J. and Ruben, R. (2007) Does marketisation of aid reduce country poverty targeting of private aid agencies? European Journal of Development Research (forthcoming).

  • 15.

    Koch, D.J. and van der Laan, J. (2007) The New Geographical Economics of Non-profit Organizations: Explaining the Agglomeration of Development NGOs . Paper presented at the AEGIS European Conference on African Studies, 11 July 2007.

  • 16.

    Robert Chambers described these biases in Rural Development: Putting the Last First (1983). Harlow, UK: Longman (now Pearson Scientific).

  • 17.

    Bebbington, A. (2004) NGOs and uneven development: Geographies of development intervention, Progress in Human Development, 28(6): 725-45.

  • 18.

    Cooley, A. and James, R. (2002) The NGO scramble: Organizational insecurity and the political economy of transnational action, International Security, 27(1): 5-39; Fruttero, A. and Gauri, V. (2005) The strategic choices of NGOs: Location decisions in rural Bangladesh, Journal of Development Studies, 41(5): 759 –87.

  • 19.

    Michael Edwards and David Hulme described the relationship between bilateral donors and NGOs as ‘too close for comfort’. Edwards, M. and Hulme, D. (1998) Too close for comfort? The impact of official aid on non-governmental organizations, Current Issues in Comparative Education (1)

  • 20.

    R-squared 0.38 and correlation 0.62.

  • 21.

    Koch, D.J., Westeneng, J. and Ruben, R. (2007) Does marketisation of aid reduce country poverty targeting of private aid agencies? European Journal of Development Research (forthcoming).

Global Development Strategies


Global Development Strategies

Research article


Non-governmental development organizations are expected to focus on countries with poor governance. New evidence shows that they do not. While they do tend to focus more than bilateral donors on poor countries, they make some curious geographical choices. It seems that they too have their ‘donor darlings’.

About the author

Related articles

Special report: Collective self-interest
Aid is a knowledge industry
Aiming global
The good, the bad and the governance