7 Defining the sector

In several countries, SWAps have been developed at a sub-sectoral level -focusing in many cases on district services and/or primary care². There are some advantages to this approach, particularly if it offers a manageable way of dealing with problems of financial accountability and performance monitoring - in the first instance. In other words, it may be easier to develop systems, that will allow pooling of external and government funds - frequently referred to as a district basket - as a precursor to a more comprehensive approach - as is the case in Zambia. There are, however, several potential drawbacks if SWAps remain limited to the district or primary level.

2 The same is true in the education sector where the focus is often on primary or basic education.

Social sectors, such as health, have more integrated characteristics than, say, the transport sector, which can be more easily disagregated into coherent institutional components (for example, the roads sector) In health, a focus solely on primary care may mean that intra-sectoral resource allocation - for example, between hospital and first contact care, management and service provision, and in-service training and professional education - is neglected. And yet it is these critical imbalances between major spending categories that need to be addressed through a sectoral approach. In addition, a focus on primary care can perpetuate the concentration of donor funds on those aspects of sectoral development which are more attractive to external agencies, and lead to unsustainable levels of financing for one particular level of the system.

The conclusion, therefore, is that health SWAps should ideally be concerned with the sector as a whole, and thus the entire network of public and private institutions financed, managed or regulated by the ministry of health. When this is not possible - as will be the case in many countries - the development of systems for monitoring performance and tracking funds throughout the sector will be a priority for the programme of work (see Section 15).

Multi-sectoral programmes are less common - with the Pakistan Social Action Programme (SAP) being the most prominent example. In this programme, which is somewhat different in nature from SWAps planned in other parts of the world, donor funds have been used to reimburse primary level government expenditures in four sectors (health, population, education and water supply). Thus, strictly speaking, the SAP is concerned with four sub-sectors, and in the first phase was primarily concerned with macro-policy issues - specifically with increasing the level of domestic funding for the social sector.

In support of multi-sectoral programmes, it is often argued that a sector-specific approach will fail to address the multiple determinants of ill-health. The case of the SAP in Pakistan is helpful in this respect in that, while it was successful in its objective of increasing domestic funding across four sectors, it is said to have been less effective in improving service quality in the individual sectors concerned. It is therefore important to be clear about the main purpose, and acknowledge the possible limitations, of a multi-sectoral approach. Secondly, there is no reason why a sector-wide approach should signal a return to a medically-dominated approach to health development. There is a strong case for broadening the scope of health policies. This, however, is a matter for negotiation with national governments. Lastly, there are significant advantages in working through existing national institutional and financial structures. Whilst recognising the complementarity of interventions in more than one sector, this does not have to be achieved through the creation of new multi-ministerial programmes.

Lastly, some agencies have also suggested that the sector could be defined in terms of cross-cutting themes such as gender, poverty, indigenous people or environment. It is now generally agreed, however, that in the absence of an appropriate institutional structure and national financing programme, this is not a realistic option. Rather, cross-cutting themes need to be addressed in the context of specific sectoral programmes.

8 Country context

To date much of the thinking on SWAps has focused on low-income countries in Africa. There is no reason, however, why the principles of a sector-wide approach cannot be applied in a much wider variety of contexts. Three different situations are considered below: decentralised states, middle income countries, and countries emerging from conflict. A more complete review might also consider the relevance of sector-wide approaches to those countries in the Former Soviet Union in which development assistance plays a significant role.

Decentralised states

Whilst decentralisation can take many forms, two situations merit particular attention: large, federally-organised countries, where the key issue is whether SWAps should be developed at national or state level; and smaller countries where a nation-wide approach is clearly desirable, but in which managerial and financial responsibility for different parts of the sector is divided between central and local government.

In the first case, most of the interaction between government and donors has traditionally taken place at federal level. However, apart from those institutions supported from the national budget (such as referral hospitals and research institutions), the main link between financing and implementation (and thus in terms of the original definition "a coherent sector defined by an appropriate institutional structure and financing programme") is to be found at state level. From the perspective of constituent states, the federal level often acts in the same way as an external donor, providing funds and materials which supplement those available from state revenues or direct external contributions.

Overall, therefore, there is a strong case for suggesting that the state, rather than the federal level is the best starting point for a sector-wide approach. Developing a sector-wide approach at state level will be considerably easier in circumstances in which the federal level ministry of health provides un-earmarked support for state health budgets. If federal contributions (like donor funds) are linked to specific activities, channelled directly to districts thus bypassing state treasuries, or earmarked for "national" programmes, the development of coherent state-level sectoral policies and spending priorities will be far more difficult. This is the situation in India, for example, where it has been recognised that for a sector-wide approach to family welfare to succeed requires that the purpose, balance and composition of federal subsidies to the states be carefully reviewed.

The situation is no different in principle in smaller countries where district-level local government has a significant role in financing and managing health services. In practice, however, two sets of problems need to be considered. Firstly, if central government provides a block grant to districts to be allocated according to their own priorities, earmarking funds for health (or other sectors) by donors, or national government, will undermine local autonomy. At the same time donors, are likely to need some assurance that their funds are being spent on the purpose for which they were originally provided. The way forward lies in negotiated agreements between central and local authorities on the proportion of funds allocated to priority sectors.

A further challenge arises from the fact that very few governments in aid receiving countries actually do provide un-earmarked block grants to local authorities. Rather, the situation is more complicated, with some funds (often for hospitals or specific national programmes) still being retained by the national ministry of health, while funds for primary care are channelled directly from the treasury to local authorities. In these circumstances - well illustrated by the situation in Tanzania and Uganda - strategy development and financial planning for the sector as a whole will be problematic. More work is therefore needed in developing an approach to sectoral development in countries where local government has an increasingly important role.

Middle income countries

Many middle income countries in South East Asia, the Middle East and Latin America have a number of characteristics in common. These include: more mature institutions, multiple agencies involved in financing, purchasing and providing health care; active private and social insurance markets; significant levels of decentralisation to provinces and/or municipalities; active, and in some cases, advanced reform programmes affecting several different types of institutions. In most of these countries only a small proportion of health expenditure comes from external agencies.

Despite the limited role of donor investment in the health sector, many countries value external intellectual inputs into policy development. This may take the form of flexible technical and financial support for policy experiments or pilots; opportunities to evaluate the advantages and disadvantages of different health care systems, through contacts with colleagues in other countries: and, in some cases, provision of practical know-how in implementing new approaches to health care financing and management. SWAps in middle income countries are thus more likely to be concerned with policy development, than with greater external involvement in financial planning, or the development of common management arrangements.

Unstable situations and countries emerging from conflict

Experience of developing sector-wide approaches in countries emerging from conflict is limited. Whilst Mozambique is perhaps the best example, attempts in other countries, such as Sierra Leone, have been disrupted by continuing instability. Clearly, there are several potential pitfalls: limited capacity within newly-formed governments to develop sectoral policy and strategy; on-going political conflicts and divisions between rival factions; acute scarcity of national financial resources, exacerbated by heavy military spending: dysfunctional management and administrative systems; and the need to provide direct humanitarian assistance to those most at risk of disease or death, often involving a plethora of external agencies.

There is no reason in principle, however, why a sector-wide approach should not provide the basis for sectoral development - once basic humanitarian needs have been met. Rather than re-establishing planning and management systems around the needs of separate projects, the progressive establishment of systems which allow donors to support a common development programme, led by government, has much to recommend it.

9 Health outcomes and priority health programmes

Sector-wide approaches are concerned with improving health status - bringing together work on health systems and health outcomes. Achieving this overall objective will involve the introduction of new technologies and practices, and the protection of funding for interventions of proven effectiveness. It is the latter which causes particular problems. Should separate provision be made for external investment in categorical programmes with major public health importance - such as malaria, TB, reproductive health or HIV/AIDS?

In principle, the issue is relatively straightforward. If major public health problems are given due prominence in sectoral policies, and receive an adequate allocation of financial, human and material resources within an overall sectoral spending programme - then no separate line of funding should be needed. Furthermore, there is no reason why the provision of technical advice on new technologies or approaches should require the establishment of a separate programme. Difficulties arise only when external and national investors implicitly or explicitly disagree about priorities - such that, in the judgement of donors or their technical advisers, funding of the sector as a whole would result in insufficient resources being made available for tackling major causes of ill health.

In defence of separate investment for priority programmes, supporters argue that without protected funding, public health programmes - especially those that benefit the poor - will be the first to be squeezed when revenues decline, or when other parts of the sector (such as hospitals) overspend. Donor-funded special programmes offer one way of ensuring that reasonable levels of funding are sustained.

The problem, however, is that separate funding for priority programmes can perpetuate the concentration of external funds on donor-defined priorities, and thus distort overall resource allocation within the sector. Second, external funding of categorical programmes tends to lead to the establishment of separate or parallel systems for managing and financing the programmes concerned - with negative consequences for managerial decentralisation, resource planning and overall systems development.

Suggesting that separate financing for major public health programmes has no place in a sector-wide approach is currently unrealistic. Furthermore, donor funds for disease control (particularly from global initiatives such as polio eradication) may not be made available for other purposes - and an overly purist position risks reducing overall levels of external funding. What is required therefore is a pragmatic approach which minimises the negative effects on sector-wide planning and management. Elements of good practice might include the following:

· Negotiation between governments and donors about the proportion of funds allocated to public health programmes will be critical in the early stages of designing a sector-wide approach. Separate funding for priority health programmes should not be regarded as the default. Agreement on which areas of expenditure should be protected in the face of resource shortfalls will be central to budget negotiations, and government mechanisms for ring-fencing funds should be used by preference. Earmarking by donors and the establishment of separate programmes should only be used as a last resort.

· In circumstances where separate funding is required, it will be important to pay careful attention to its institutional consequences. In particular, there will be a need to consider the problems associated maintaining separate budget lines, staffing establishments, and information systems. The need to introduce new technologies or practices, and to back these with the provision of drugs, equipment or technical advice, does not in itself justify the establishment of separate or special programmes.

· In determining how resources should be allocated, evidence-based approaches (such as burden of disease studies and cost-effectiveness analyses) will be important, but must be reconciled with the need to maintain adequate funding for other parts of the health system. Decisions will be influenced by technical experts, but their role should be to help governments use limited resources effectively, not to act as lobby group for special interests.

· It is important not to conflate the need to assess the effectiveness of public health programmes with the need to monitor overall sectoral performance. Individual programmes will each have their own detailed information requirements, determined by health service managers and their technical advisers. Indicators of sectoral performance, on the other hand, will necessarily be far more selective. They will include targets in relation to an agreed set of health outcomes, the achievement of which will depend on the effective performance of a range of health programmes. They will also include targets in relation to the development of information and other management systems. This aspect of sectoral performance is concerned, not with the achievements of individual programmes, but asks whether systems are in place which make such monitoring possible.

10 Poverty and the health of poor people

Making a contribution to reducing the causes and effects of poverty is a fundamental principle underlying the development assistance provided by most donor agencies. The international community as a whole, through the Development Assistance Committee of the OECD, has recently defined a set of indicators by which the success of the global development effort can be judged³ Foremost among these is the goal of reducing by at least one-half the proportion of people living in extreme poverty in developing countries by 2015. Increasingly, therefore, signatories to this agreement are concerned that development assistance, irrespective of its particular sectoral focus, contributes to poverty reduction⁴.

3 Shaping the 21^st Century: the contribution of development assistance. OECD Paris. DCD/DAC(96)15/final (May 1996).

⁴ The DAC targets also include two social development objectives of direct relevance to the health sector: a reduction by two-thirds in the mortality rates for infants and children under five and a reduction by three fourths in maternal morality, all by 2015; and access through the primary health care system to reproductive health services for all individuals of appropriate ages, as soon as possible and no later than the year 2015.

How then to ensure that sector-wide approaches to health development address the needs of the poor and help to reduce poverty? In posing this question, it is important to recognise that it reflects a growing suspicion in some quarters that SWAps are inherently "statist", centralising, top-down, solely concerned with upstream policy issues and the supply of services, rather than mechanisms that help the poor articulate demand for better health care.

Inevitably, some elements of sectoral policy making will indeed be top-down. Governments have to make choices - among many competing claims - about how limited resources are allocated and rationed. The main concern, however, is the nature and impact of these choices, not the fact that they have to be made. External investors will be concerned that the national policies they are being asked to finance reflect a concern for the poor, and that sectoral strategies are designed to ensure greater equity - both in terms of outcome and provision.

Once again, therefore, the choice facing donors is whether they should channel development assistance as directly as possible to those perceived to be most vulnerable or, through their involvement in the negotiation of sector policies and strategies, attempt to influence the way resources are allocated, in ways that favour the poor.

Reconciling a poverty-focused agenda for development assistance with a sector-wide approach to health development will not be straightforward. Experience suggests a number of practical issues which will have to be considered:

· Collective responsibility for sectoral development requires that donors have to acknowledge the pressures facing the national governments with whom they are working. Whilst policy and resource allocation can and should become more closely geared towards the needs of the poor, governments cannot ignore political realities or their responsibilities toward the rest of the population. Involvement in SWAps means that donors in the partnership will be concerned to influence overall spending decisions, rather than taking the easier option of assuming responsibility for only funding programmes targeted at specific groups.

· For this to be possible, and to influence sectoral policies in favour of the poor, requires a more sophisticated understanding of the impact of different aspects health policy. In other words, there is a need to know what to negotiate about. Of the many possible factors that can have an influence on the health of the poor and the prevalence of absolute poverty (such as targeting of subsidies for hospital or primary care, rural versus urban spending, distribution of health professionals, user fees, insurance arrangements), which are the most important? A particular concern in countries where the bulk of health spending takes place in the private sector, will be to ensure that policies adequately address the way that governments manage the private provision of health care - to avoid exacerbating inequity.

· The proportion of funding allocated to primary care and/or rural districts is often taken as a crude indicator of a concern for the poor. However, the idea that funds to primary care equals impact on poverty is too simplistic. Firstly, it assumes that district populations are homogeneous, and fails to acknowledge that some groups have greater access to services than others. Secondly, it justifies the impact on poverty on the basis of first principles alone, rather than hard evidence.

· In relation to the first problem, sector-wide approaches have to be concerned with mechanisms by which different groups in the population influence the form or content of services. There is no a priori reason why a sector-wide approach should lead to greater central control, and the establishment of decentralised systems needs to be a key component of the collaborative programme of work.

· In the same way that health services can only make a contribution to better health, so development assistance in the health sector can only make a contribution to the reduction of poverty, A better understanding of the relationship between health care provision and poverty, as well as methods for assessing the effectiveness of the health sector's contribution to its alleviation, are urgently required.

Lastly, a question being raised in some donor agencies is whether development assistance - in health and other sectors - should focus primarily on those countries which are judged to be making a serious effort to reduce poverty and inequity. This concern is echoed in the DAC document on "Shaping the 21^st Century" in discussing countries "in which civil conflict and bad governance have set back development for generations". The difficulty of actually defining what constitutes a serious effort notwithstanding, it is likely that donors will continue to provide assistance to the poor even in countries with inadequate policies. In these cases, a sector-wide approach may have limited application, and a twin-track, or more targeted approach to development assistance may be more appropriate.

11 Ownership and incentives

The viability and success of sector-wide approaches will depend on the degree of political support they receive from the major players involved. This in turn will be influenced by how interest groups within governments, donor agencies and civil society are affected by the introduction of SWAps.

Donor agencies

Almost all of the technical agencies, development banks and bilaterals consulted during the preparation of this document support the idea of sector-wide approaches in principle. In contrast to projects, SWAps are seen as a means by which donors can disburse funds more rapidly and, at the same time, increase the sustainability of their investments. A reduction in the managerial load imposed by multiple projects opens the door for greater involvement in strategic analysis and planning. It is equally evident, however, that underlying this broad consensus there are many differences of opinion, both within and between agencies, concerning their involvement in SWAps.

· Whereas concerns about accountability are common to all, some bilateral agencies are prevented by their governments from providing recurrent cost support as part of development assistance. There is also an acute awareness of the political risks of being associated with corrupt or unproductive spending

· In several agencies, sector specialists have been under pressure to focus on a limited number of technical areas within the health sector. In contrast, economists in the same organisations are looking at the health sector as one which is ripe for more broad-based support. Whilst increasing specialisation is a logical response to shrinking aid budgets and the need to clearly demonstrate results, it has important implications for moving toward a sector-wide approach. Most particularly, it lends itself to the development of projects rather than sector-wide programmes, and concentrates expertise in a few technical areas⁵.

· Despite the best intentions, several agencies acknowledge that the pressure to maintain levels of expenditure encourages managers to continue with business as usual. On a more positive note, however, several of the donors consulted had established task groups to consider the technical and managerial implications of moving toward a sector-wide approach.

5 It is also significant that many bilaterals have chosen to focus on the same technical areas (notably reproductive health, communicable disease and district level health care). Their mandates therefore overlap with each other and with some of the UN specialized agencies.

Ownership and national governments

Government ownership is seen as the sine qua non of a sector-wide approach. There is, however, a real danger with the growing enthusiasm for SWAps on the part of donors, that external pressures will be used to urge governments to take the lead, in situations where there is only limited capacity and interest in so doing. Countries that have made the most progress with sector-wide approaches (such as Ghana and Zambia), are those where government has most obviously taken the initiative and donors have had to respond.

Why should governments be interested in a sector-wide approach? The incentives are not always clear-cut. There is no certainty, for example, that involvement will guarantee increased levels of external investment. Secondly, despite statements about "government being in the driver's seat", SWAps inevitably result in greater external scrutiny and discussion of issues which were previously the sole preserve of national authorities.

SWAps can increase overall national ownership of sectoral development -particularly through greater control over technical assistance and policy advice. Nevertheless, it is important to recognise that they will also have the effect of shifting power within and between ministries. This will influence the degree to which the sectoral programme is "owned" by different parts of government.

Within ministries of health, a sector-wide approach is likely to strengthen the hand of senior policy makers - particularly those that are perceived as "reformers" - but reduce the influence of others. SWAps are also likely to decrease the power of individual project managers and other donor-funded fiefdoms in the organisation. Despite the obvious advantages of such a move, those who lose the benefits associated with control over project funding are unlikely to become enthusiastic supporters of the new approach.

In the face of uncertain domestic revenues, ministries of finance may resist pressure from donors to provide longer-term and more definite commitments to specified levels of sectoral funding. On the other hand, ministries of health that are accustomed to using unrealistic annual budgets as a way of bidding for funds, may be reluctant to submit more accurate estimates of resource needs. Particularly if they are worried that inclusion of donor commitments tempts the ministry of finance to cut their domestic funding.

A more transparent process of resource allocation may bring to light other practices - such as the last minute inclusion of capital projects in the budget -which have previously gone unremarked by external agencies. The involvement of donors in more detailed budget scrutiny may help ministry of health officials deal with such external pressures, but is unlikely to recommend the sector-wide approach to those who might otherwise have benefited from the transactions.

Ownership and civil society

Organisations in civil society need to be seen as partners in a sector-wide approach. However, it is useful to distinguish different aspects of the relationship, and identify different processes of interaction, which have a bearing on the issue of ownership.

In relation to the development of sectoral policies and strategies, it is important to distinguish between consultation by governments seeking inputs into the design of programmes, from public relations and communication strategies designed to explain government policies to beneficiaries. Both have an important role, and the latter is frequently neglected.

The process of consultation will necessarily involve those that stand to benefit from and use health services, those that have previously been excluded from access, and representatives of organisations outside government involved in the provision, financing or regulation of health care. In the latter group particularly a wide range of interests will be represented and it is important that "ownership" of policies is not achieved at the expense of fudging all controversial issues.

Secondly, the process should stress the need to make choices, if an unrealistic wish-list is to be avoided. Consultation can refine, but not substitute for decision making. In this regard, the involvement of the public in helping governments make decisions about the rationing of health care resources is not just an issue for low income countries. There is now a growing body of experience from the industrialised world as well.

Lastly, the development of mechanisms which enable members of civil society to influence the way in which health services are run - on a day-to-day basis - are arguably just as important in securing ownership of a sectoral programme.