Tim Harford had an interesting article in the FT in August arguing that we are better off in most walks of life if there is experimentation and a multiplicity of approaches.

But how do we value diversity in the aid business, when the prevailing consensus, embodied in the Paris Declaration, is that proliferation of aid agencies is a growing problem which is making aid less effective?

The aid system could in principle benefit from the emergence of new kinds of donors (specialised multilaterals such as GAVI, new donors such as China and Brazil, philanthropic foundations such as Gates, private non-profits such as Marie Stopes) working alongside conventional bilateral and multilateral aid.  Different kinds of organisations could bring particular strengths which complement each other’s work.

However, in practice these different types of organisation do not seem to be playing to their strengths. Like kids playing football, everybody follows the ball instead of holding their position on the pitch.

Proliferation is a significant problem

We will come to the benefits of diversity among donors. But first let’s acknowledge that proliferation is causing real problems on the ground. Developing countries are having to deal with a large and growing number of partners, each with separate agendas, priorities, and requirements. Meetings, reports, milestones and systems multiply. Skilled staff are hired away to serve in local agency offices or NGOs. Funding is fragmented and unpredictable, which means that developing countries are often unable to bring together the scale of long-term, predictable finance needed to undertake significant institutional reform and service delivery. Donors lose influence, because they undermine each other; and yet developing countries are not able to keep track of, let alone exercise sufficient ownership and control over, an increasingly fragmented system of aid delivery. Public accountability is impossible, since nobody has a clear view of what resources are being used, by whom, or for what purpose. Donors face rising administrative costs when agencies proliferate, and the costs of coordination and harmonization rise exponentially with the number of aid agencies.

Here are three real life examples of the problems that are caused by the proliferation of aid agencies:

• In Vietnam, it took 18 months and the involvement of 150 government workers to purchase five vehicles for  a donor-funded project, because of differences in procurement policies among aid agencies. (source: Knack/World Bank)
• In 2007 alone the EU countries launched 22,000 new aid projects inn developing countries, with an average budget of €0.7-1 million. The total costs of preparing new projects by EU donors (not the money needed to fund them, just the administrative cost of putting them in place) is estimated at between €2-3 billion per year. (source: EU)
• In the aftermath of the tsunami disaster a local doctor in Banda Aceh, one of the most affected areas, wrote: “In February, in Riga (close to Calang) we had a case of measles, a little girl. Immediately, all epidemiologists of Banda Aceh came in, because they were afraid of a propagation of measles among displaced people, but the little girl recovered very fast. Then, we realized that this was not a normal case of measles and we discovered that this girl has received the same vaccine three times, from three different organizations. The measles symptoms were a result of the three vaccines she received.” (source: Djankov et al)

(For more examples of proliferation badness, take a look at ‘The Governance of the aid system and the role of the EU’ by Owen Barder, Simon Maxwell, Mikaela Gavas and Deborah Johnson.)

Different types of agency could make different contributions

These problems are caused by a growing number of aid agencies doing broadly the same thing.  That proliferation imposes substantial costs on donors and on recipient countries and this makes aid much less effective.  The question is whether there are also benefits to having this large number of agencies, compared to delivering the same amount of money through fewer channels.

In principle a greater variety of different types of donor, if they focused on their specialisms, could strengthen the aid system, because they can make different kinds of contribution which could complement both existing donors and each other.

Here are some ways in which different types of donor can make different contributions:

• Philanthropic foundations, such as Gates, Ford, Hewlett and Rockefeller, are still tiny in comparison to government aid agencies, but they are increasingly important in particular sectors, notably health.   In their recent book, Philanthrocapitalism, Matt Bishop and Mike Green argue that the growth of philanthropic giving should be welcomed, because these foundations bring a “businesslike approach to solving society’s problems”.  According to this view, philanthropic donors bring new attitudes and ways of working. Foundations are frequently founded by successful entrepreneurs, so they may be more inclined to operate along business principles, such as making decisions based on evidence, tightly controlling overheads, adopting new technologies, and focusing more sharply on results. They may be willing to take more risks and accept more failures in return for bigger success than risk averse governments. Foundations may be more able and inclined to work closely with the private sector, which plays a key role in development, which official agencies have not found easy to do.  Because foundations do not depend on public support for future funding, they may be willing to support unpopular causes, or investments which do not easily capture the public imagination (e.g. supporting statistical systems in developing countries).
• New government donors such as China and Brazil are playing an increasingly important role (though the Economist was wrong to suggest that Brazil’s aid budget is comparable to that of Canada and Sweden).  This has caused concern among traditional donors, who worry that their implicit cartel is undermined by donors that are less concerned about governance and human rights, and that are prepared to be more open about its desire for access to raw materials and minerals. These new donors do not feel constrained to follow the DAC development model, and in many ways developing countries prefer the approach which tends to respect the sovereignty and ownership of developing countries. These donors rarely poach skilled staff; and they do not overstretch developing country governments with meetings, reports and workshops.   They are also willing to invest in sectors that the DAC donors have moved away from, such as infrastructure, irrigation and university scholarships.
• The number of private charities is also growing, funded both by institutional donors and by private giving. Here in Ethiopia there are about 3,500 NGOs, spending about $1.5 billion a year (compared to the Ethiopian government budget which is about $4 billion a year). Private aid through charities tends to focus on supporting communities and individuals rather than governments. It tends to be more opportunistic and closer to the ground. These organisations can bring about results more directly although it is harder to bring about systemic change this way.
• Specialised multilateral global organizations – such as the Global Fund against AIDS, TB and Malaria (GFATM) – continue to grow in number. In principle, they can bring apply specialist skills and expertise, they can learn more systematically and spread knowledge more quickly, they can bring together a number of different donors, the public and the private sector to work in a more joined-up way on a particular issue, and they can raise money from the public because they can be more specific about what they do.

This changing landscape could benefit the aid system …

In an ideal world, if these different development actors played to their strengths, and stuck to their specialities, this growing diversity could strengthen the international aid system as a whole. Foundations could act like venture capitalists: taking bigger risks, and backing it up with rigorous evaluation and evidence, but leaving long-term financing of scaled up successes to official aid donors. Official aid agencies could focus on long term funding and resource transfer, and they could provide sustained support for institutional change and capacity. Private aid could focus on achieving community and individual level results. Specialised global organizations could provide particular expertise not available through generalist support. The growing number of official donors could build up expertise in particular countries or topics, and specialise in these, and they could respond to evidence generated by foundations and NGOs about what works, by taking those activities to scale.

If these actors could all focus on their strengths, and if the aid system enabled them to work together well, these changes in the development landscape might substantially improve the effectiveness of development assistance.

… but in practice it does not work like that

That’s all very well in theory, but most people working in the aid business will tell you that back on planet earth, it doesn’t work like that.

Rather than differentiate, development organisations have strong incentives to converge.  So instead of specialisation we get duplication.  The philanthropic foundations say that they have a more entrepreneurial, risk-taking approach; anecdotal experience suggests that in many cases they prefer the implicit validation of being part of a multi-donor group.  (This may be a form of political correctness: agencies seem to think that the Paris Declaration on Aid Effectiveness requires that they be part of a shared funding arrangement rather than doing anything alone.)

For example, consider the bandwagon on restoring funding of health systems.  Increasing the funding of health systems is something of which all right-thinking people should approve.  The arrival of the big global health initiatives, particularly GFATM and GAVI, coincided with a collapse in funding for health systems which led to many unnecessary deaths in developing countries. Donors are now seeing that the shift away from health systems to vertical funds was an error (one which was predictable and predicted), and the pendulum is swinging back to funding health systems.  The institution with the mandate and greatest capacity for supporting developing countries to strengthen their health systems is the World Bank. So why are the Global Fund and GAVI being allowed on the health systems bandwagon?  The logic of establishing these specialised multilateral agencies was that they would bring particular depth and expertise to specific activities which would be available from more generalised aid agencies. If we offer competition to World Bank concessional loans in the form of grant finance through GAVI and and the Global Fund, most developing countries will look to these institutions instead.  As a result of the proliferation of health funds offering grant finance for health systems, the core role and capacity of the World Bank is eroded, and we put at risk the benefits of specialisation by GFATM and GAVI.   Similarly, the International Finance Facility for Immunisation (IFFIm) was set up to enable donors to secure the benefits of front loading spending on vaccination, for which there is a clear economic rationale.  Now it is proposed that it should also finance health systems: if there is an economic rationale for using IFFIm on health systems, I’d like to hear about it.

What’s missing?

The growing number and diversity of development organisations could be a source of strength in the aid system, if different organisations could stick to their specialities and if they worked in an aid environment which enabled them to work together effectively.

In competitive markets, firms tend to focus on their strengths, because this is how they make the biggest profits. Firms that diversify into another line of business either need to make a success of that new work, or they will start to make losses and eventually decide to withdraw or they will go bust. So appropriate specialisation is the consequence of individual decisions by profit-maximising firms, and not a result of a collective compromise.

Unfortunately, the political economy of aid encourages the opposite behaviour.    The “operating system” which supports the work of aid agencies creates pressures against specialisation.  For example:

• Organisations which work collaboratively and holistically across a wide range of activities are likely to attract more donor funding than organisations which are effective in a particular niche.  One reason for this is that many donors either don’t have, or don’t systematically use , information about impact and cost effectiveness when they make resource allocation decisions – so there are rewards for aid organisations getting involved in as many activities as possible, and no penalty if this mission creep makes them less effective.
• Lack of transparency and access to information about who is doing what means that organisations cannot make sensible individual decisions about how they can increase their own impact with finite resources and avoid duplication.
• There are no mechanisms by which innovative ideas can be pioneered by foundations or NGOs and, if they are successful, taken up and taken to scale by official donors and multilateral funders. There too little venture capital to support innovation; too little rigorous analysis of what actually works; and the mechanisms for taking successful programmes to scale are too unpredictable and capricious.
• Donors, NGOs and foundations are all under pressure from well-meaning activists to be engaged in everything everywhere.   For example, last year the Lancet criticized the Gates Foundation saying that it should “do more to invest in health systems and research capacity in low-income countries, leaving a sustainable footprint”.  DFID is criticised for a perceived lack of investment in agricultural research.  In a sane world it would be perfectly sensible for the Gates Foundation, which has very little in-country presence, to fund technological research in health and agriculture, but not to invest in health systems in developing countries; and for DFID, which has an extremely professional presence on the ground in developing countries, to invest in developing country systems but not to spend money on research, in which it has no discernible comparative advantage.  We could have the same total spending on both research and systems, managed by organisations specializing in those activities and reducing coordination and transaction costs.  But development activists and politics apparently make such a division of labour impossible for both organisations.
• The Paris Declaration on Aid Effectiveness and Accra Agenda for Action are being implemented in ways which create strong peer pressure on donors to collaborate and harmonise, to engage in pooled funding and joint activities, rather than to diversify and specialise.  Where there are efforts towards a better division of labour (e.g. this EU initiative), the approach is based simply on getting down the numbers by committee, rather than creating incentives which push development agencies towards focusing on the areas in which they have a comparative advantage.

What should we do?

The proliferation of development organisations, which could be a great strength, is instead becoming a growing handicap for the aid system, because the system is not well adapted to taking advantage of that diversity and encouraging appropriate specialisation.

Some possible measures that might address this are:

• a step change increase in transparency about aid.  The International Aid Transparency Initiative offers the promise of this, as it will provide up-to-date, detailed information about aid projects in an accessible form.
• agreement to an international standardized system for describing and measuring outputs and unit costs, to facilitate cost-effectiveness comparisons across development organisations;
• explicit use of unit costs and cost-effectiveness in aid allocation decisions, in a way that penalises organisations which are engaged in activities in which they are relatively ineffective
• the development of a mechanism for “venture capital” funding with an associated process for scaling up success;
• self-restraint by development activists who do more harm than good by trying to push every development organisation to be involved in everything.

As ever, I’d welcome further suggestions in the comments section.