Making Development Assistance for Health More Effective: Lessons Learned
More assistance is part of the answer to helping developing countries achieve more rapidly the improved health outcomes they seek and that are enshrined in the MDGs. To this end, we need to know how effective DAH has been and what can be done to make it more effective.
Despite valid criticisms of DAH, some health programs—inspired and supported by donors—have worked at scale and contributed to more than four decades of steady improvements in health, as measured by under-five mortality and overall life expectancy. The record of public health successes in developing countries is becoming increasingly clear, as noted in a recent review of four decades of experience (Levine and What Works Working Group 2004). The success stories cover a broad spectrum of circumstances. They are found in all regions and cover both communicable and noncommunicable diseases. They have been driven by new technologies, including vaccines, drugs, and diagnostics; community- and clinic-based care; and knowledge for behavior change.
Significant gains have occurred even in the poorest countries and in those with weak institutional environments. Consider, for example, the high levels of TB case detection in the Democratic Republic of Congo and in Myanmar (Stop TB Partnership 2003), the successes in polio eradication in African countries experiencing civil wars, and the growing availability of antiretroviral therapy in Haiti.
Some of this progress can be attributed to the general effects of economic growth and improvements in education, water, and sanitation. However, specific, compelling examples of the success of DAH-backed initiatives are available. For example, programs to immunize against measles, to control river blindness and guinea worm, and to fortify salt with iodine have had sustained and widespread effects (Levine and What Works Working Group 2004). These programs have been successful at scale, have generated sustainable health improvements at the population level, and have succeeded in a broad range of institutional environments.
Recent analyses and reviews of donor-supported successes in international health have noted a set of factors that tend to contribute to positive outcomes:
-
strong internal (governments) and external (donors) political leadership
-
collaboration across governments, donors, and nongovernmental organizations (NGOs) in program design and implementation
-
consistent, predictable funding support, even after success has been achieved
-
simple and flexible technologies that can be adapted to local conditions and do not require complex skills to operate and maintain
-
programmatic approaches that recognize and address the need to help build health system infrastructure, especially in human resources
-
household or community participation in the design, execution, and monitoring of program activities.
Policy Environment
Development assistance for health supports a vast array of activities and services, some focused on specific diseases (polio, TB, HIV/AIDS), some on strengthening health systems (disease surveillance, nurse and midwife training), and some on particular services (reproductive and child health services). But has DAH actually changed health outcomes? Recent work from the World Bank, the Commission on Macroeconomics and Health, and others suggests that it has (Rajkumar and Swaroop 2002). However, DAH does not work as effectively in countries where the policy environment is poor, even though some carefully targeted disease control activities can confer limited benefits. With good policies and institutions (strong property rights, reduced corruption, an efficient bureaucracy), an extra 1 percent of GDP in aid is estimated to reduce infant mortality by 0.9 percent. By contrast, where policies are average, the decline is estimated at only 0.4 percent, and where policies are poor, aid is estimated to have no significant effect on infant mortality (World Bank 2004b).
The issue is not black and white: there are gradations of good policy, and as policies improve, the productivity of aid increases. For example, Bangladesh has made large strides in reducing under-five mortality in recent years, relying on NGOs to deliver many services. If Bangladesh were able to raise the quality of its governance from below average to above average, even at its same public spending levels, it would realize more rapid gains. An additional dollar of government health spending would reduce under-five mortality by 14 percent, compared with 9 percent without such improvements (World Bank 2003c).
Tradeoffs may be necessary between targeting assistance toward the neediest countries and achieving the greatest effect from DAH, but even in needy countries with weak policies, some kinds of carefully targeted assistance for health (for example, immunizations delivered by reputable NGOs) can have a positive effect. In addition, in countries with weak policies, a focus by donors on policy dialogue and technical assistance to improve the environment for DAH can set the stage for a larger infusion of financial support down the road.
Conditionality
Conditionality is making the availability of funding dependent on a government completing an agreed task, such as enacting a new health law or spending a certain share of its budget on health activities. It can work, but only if the government is committed to making such a change. Tying aid to policy changes is a common practice, but recent studies have cast doubt on the ability of such conditionality to bring about reform (World Bank 1998, 2003a). If governments are not committed to reform, conditionality will not make them reform. Donors themselves often undermine the rigor and credibility of conditions because they usually face strong internal pressures to continue disbursement of the funds anyway, even when governments do not adhere to the agreed-on conditions. On the other hand, if governments are committed to reform, conditions can help by enabling governments to commit publicly to certain reforms and persuade private investors of their seriousness. For example, the government of Uganda's commitment to decentralizing the management of basic health services and to making local authorities accountable to communities was reinforced by conditions in the Uganda poverty reduction support credit, which several donors financed. Similarly, the Chinese government's commitment to reaching the poor with TB control services was reinforced by stipulations in donor-funded TB projects that they target the country's poorest provinces and reach out to deprived households (World Bank 2003c).
Donors cannot force policies on governments, but they can help with policy design. Donors can alert governments to the reasons for reform and help nurture commitment, but at the end of the day, it is governments that have to sustain any reforms (box 13.1). Undertaking analytical work, providing training and technical assistance, disseminating ideas about policy reform and development, and stimulating debate in civil society can all be valuable activities for donors to support while a government's commitment to reform is growing.
[Box 13.1]
Vietnam in the late 1980s and early 1990s is a good example. At a landmark meeting in 1986, the ruling Communist Party decided to break with the past and introduce sweeping economic reforms. In the health sector, the reforms included introducing user fees at public facilities, legalizing private medicine, deregulating the pharmaceutical industry, and opening the pharmaceuticals and medical equipment subsectors to international trade. Initially, Vietnam saw no increase in aid, but such agencies as the United Nations Development Programme and the World Bank helped facilitate the reform process by organizing international workshops for the Vietnamese to exchange ideas on policy with their neighbors. This effort set the stage for a large inflow of donor financing, starting around 1995 and continuing to the present.
Fungibility of Development Assistance for Health
Much aid is earmarked, both across sectors and within them. One part of a development agency gives a grant to the ministry of health for a health sector reform, while another does the same for a primary education project. An agency makes a loan to the ministry of health for a TB project, while another makes a loan for a malaria control project. The donors' intent is that these activities remain tightly sealed: the funds for health sector reform are to be kept separate from the funds for the primary education project; the TB project funds are to be kept separate from the malaria control project funds. The idea is to ensure that the government makes a certain spending choice. It is based on the assumption that the choice would not be made if the government had been handed a blank check for the same amount.
The implied view of such aid is that what you see is what you get; that is, a government receives US$1 million for a water project and the net effect is US$1 million worth of extra spending on the water sector. This view has recently been challenged, with the alternative view being that aid is at least partly fungible (Burnside and Dollar 2000; World Bank 1998). Hence, as a result of the inflow of development assistance for a specific health activity, the government changes the way in which it spends the rest of its resources, both in the health sector and in terms of allocations between health and other sectors. As a result, for each dollar earmarked for a specific health project, spending on health rises by less than a dollar but by more than would have been the case had the government received an extra dollar in its overall budget. Similarly, spending for the specified purpose in the health sector rises by less than a dollar, but by more than would have been the case had the government allocated an extra dollar of its own resources to health generally.
Assessing whether aid is indeed fungible is not straightforward. The difficulty is knowing how the government would have responded had its own resources increased by the amount of the aid or had it received a blank check for the same amount. Recent research suggests that, despite considerable variation across countries, on average only 29 cents of each additional dollar of aid goes into government development programs, with the rest leaking out into nondevelopment programs such as military spending (World Bank 2004b).
One important implication of those findings is that donors should spend less time and effort trying to channel their external funding to specific programs for priority diseases and populations. Instead, a more useful exercise would be to engage in a dialogue with the government on basic changes in the overall patterns of public spending for health—that is, the total allocation and the amounts for, say, providing child health and communicable disease control services and for improving community and primary-level health delivery systems. If the government followed through on those basic changes, then donors would transfer their financial assistance to the health sector as a whole.
The finding that aid is indeed fungible has encouraged some donors to search for broader development assistance mechanisms that recognize the importance of the entire expenditure program and explicitly avoid earmarking. Such mechanisms include the Multi-Country AIDS Program in Africa, which supports national HIV/AIDS strategies; sectorwide approaches in health; and poverty reduction support credits (PRSCs) that back a broad public spending agenda.
An opposing viewpoint is that although, in general, good policies matter and the fungibility of aid tends to undermine donors' ability to earmark their funding effectively, many high-impact health services can be delivered to the population on a targeted basis even when national policies and institutions are weak—and this aid would not occur in the absence of DAH. This argument applies especially to services with simple technologies—for instance, basic childhood vaccinations that can be delivered on a single occasion through annual campaigns or disease treatment programs (such as short-course drug therapy for TB) that can be provided through tightly managed top-down efforts (Jamison 2004). The high coverage and treatment success using the directly observed short-course therapy approach for TB in the Democratic Republic of Congo in recent years, even during the civil war, can be cited as an example of how a well-protected enclave project with strong donor backing can be successful (Stop TB Partnership 2003).
Another argument against broad budget support of health sector funding or in favor of the more traditional earmarking of DAH is that it helps maintain governments' and donors' focus on implementing and monitoring specific health service interventions and on the necessary technical and managerial improvements to ensure the achievement of targets in those areas. On the basis of experience in countries such as Bangladesh, Ghana, Mozambique, and Zambia, donors are increasingly of the view that broad support to a national health sector program leads to superficial oversight of bureaucratic processes and a corresponding loss of technical focus and depth (Foster, Brown, and Conway 2000). Although a sector-wide approach is theoretically fully compatible with careful monitoring of key outputs and health outcomes and with in-depth technical improvements, in practice, achieving this mix of objectives may prove difficult.
In summary, the debate on earmarked versus broad DAH support for national health programs continues. More analysis is needed to produce clearer conclusions on the advantages and disadvantages of the two approaches.
Transaction Costs of Aid
In a single low-income country, more than 20 donors—including bilaterals, multilaterals, global programs, foundations, and large NGOs—may be involved in the health sector. The demands placed on recipient countries can be huge, and donors are starting to acknowledge this burden. They are recognizing that their individual procedures for reporting, accounting, and managing funds—which often encompass different budget structures; different ways of measuring progress toward objectives; different regulations for the procurement of goods, services, and works; and different approaches toward and cycles for disbursing funds—place heavy and unreasonable demands on recipient countries. Demands are particularly heavy in poor countries that are forced to allocate limited human resources away from service delivery to manage donor funding.
The donor community is working to harmonize and simplify its procedures to reduce these transaction costs. In the health sector, experiments are taking place in several developing countries, including Bangladesh, the Kyrgyz Republic, Mozambique, and Zambia, to determine how best to lower the costs (OECD 2004b). Some of the principles of improved donor action include the following:
-
ensuring that countries, not donors, drive the coordination
-
fostering strategic coherence through a poverty reduction strategy and the health, nutrition, and population analyses that feed into it
-
promoting financial coherence through a medium-term expenditure framework and an agreement that all donor funding will respect the government's overall spending plans and limits
-
pooling donor funds in a single account and untying aid so that the government can procure goods and services from the lowest-cost source and not just from the donor countries
-
limiting the number of country coordination bodies that can bring together national and international actors involved in health.
At the same time, some of the experiments in country-level coordination of DAH reveal the difficulties of implementing the principles of better donor harmonization. In some instances, persuading donors to pool their funds and sever the links between funding and procurement has proven difficult. Some donors face pressure from their legislatures to maintain separate accounting for the use of their DAH allocations so they can claim credit for achieving progress and thus "plant flags" on individual health projects. In addition, monitoring and evaluation systems have frequently not been strong enough to yield timely and meaningful data on progress, a critical failure if disbursements are linked to performance rather than to spending on specific inputs. Multiple national coordination bodies for government, donors, and NGOs for different diseases (AIDS, TB) and services (immunization, polio eradication) also persist in many settings. In short, a number of political and technical changes are needed to ensure the successful implementation of a harmonized donor agenda at the country level in the poorest developing nations.
Unpredictability of Development Assistance for Health
Some donors have taken steps to put in place instruments for DAH that extend the length of their financial commitments. The development of multiphase funding "slices" of 3 to 5 years embedded in a 10- or 15-year program of support is one way to lengthen commitments. Nevertheless, donor financing for health is not yet as reliable or sustained as is often claimed or hoped for, even under these new, long-term arrangements. In some developing countries, cuts in DAH have been sharp. Donors' budgets are subject to the usual business and political cycles and may go up or down during their annual budgetary processes. For such countries as the Comoros and Eritrea, where the year-to-year changes in external funding can amount to as much as a fifth of all public spending for health, the fluctuations are so great that they make planning and implementing coherent national health programs nearly impossible (figure 13.1).
[Figure
13.1]
Further work is needed to design mechanisms for DAH that provide greater assurance of sustained financial support. The challenges are to overcome the factors that result in interruptions to long-term DAH. Those factors include changes in political leadership and aid agency management that lead to modifications of earlier agreements and end up reducing external funding levels or reallocating those funds to other activities.
One example of an innovative approach is the recently proposed establishment of an international finance facility. This facility would use financial commitments from governments of developed countries to tap funds in capital markets and would use those funds to frontload development assistance so as to accelerate progress toward the MDGs. The International Finance Facility for Immunization, which the United Kingdom formally announced at the World Economic Forum in January 2005 and is aiming to launch later in the year, will expand external financing for childhood vaccinations. The pilot project could be used to pledge funding against multiyear advance-purchase contracts for new vaccines (such as rotavirus vaccine) that may not reach the market for several years. Such an effort could help create a more assured market and reduce the risks for vaccine companies, thereby speeding up the introduction of these new health commodities to low-income countries (GAVI 2004b).
