12 Improving Overseas Development Assistance: The Broad Sector Approach

Laura Wallace
Published Date:
May 1997
  • ShareShare
Show Summary Details
Michael Foster

In recent years, donors have been developing sector-wide approaches to providing development assistance, as an alternative to traditional project aid. This new approach to development cooperation mainly derives from some ideas developed by Steve Denning in the Africa Department of the World Bank and is most fully described in a World Bank discussion paper by Peter Harrold and Associates.1

In my remarks today, I would like to talk first of all about the budget context for external aid and some of the problems with the traditional project approach that have led donors to think about sector-based approaches. I will also discuss some of the problems with reforming budget management, as a context and to some extent a prerequisite for sector aid. I will then go on to describe the main features of the sector approach—which is now a central element of the World Bank’s lending strategy in Africa—with an example drawn from Zambia’s health sector investment program. I will conclude with some comments on critical success factors for the sector approach and the role of conditionality in these programs.

The Budget Context for Aid

Most official development assistance is channeled via the government budget. It is fungible, which means that donor funds spent on, for example, a health project may release a government’s own resources to finance another activity entirely if the government substitutes the donor project funds for expenditure that it would otherwise have funded from its own resources. It is government budget decisions that determine how aid is used. This does not mean that more aid necessarily leads to more government spending. Governments have three main options for how they can use external assistance from donors: they can increase spending, reduce taxation, or reduce their borrowing. The last two options can provide direct benefits to the private sector (lower taxes, easier credit). Increased spending can also provide indirect benefits to the private sector, if it is used wisely to improve infrastructure and essential services.

African budgets and budget management have some characteristic weaknesses. These are not found everywhere in Africa, and many of these features are improving in many countries. Nevertheless, common weaknesses include:

  • large deficits that strangle private access to credit;

  • a low tax effort, often lacking fairness and transparency;

  • too much spent on projects and salaries, too little on operation and maintenance;

  • subsidies and spending priorities that are poorly targeted, often benefiting small minorities of the population;

  • budgets that are fragmented by dependency on aid; and

  • poor levels of accountability, with an ever-present risk of corruption.

These weaknesses mean that donors cannot be indifferent as to how aid resources are used in the economy. Certainly, a stronger domestic budget performance would help governments to use aid more productively. For this reason, donors have pressed governments to define a clear macroeconomic framework, with clear targets for the budget—including the roles of domestic and external finance, and the level and sectoral shares of spending.

Problems with Projects

In the aid-dependent countries of Africa, donors have themselves contributed significantly to the budget problems. One of the most important reasons for developing the sector approach has been an increasing recognition of the serious problems resulting from the project approach.

There is too much project aid relative to resources for operation and maintenance, with donors reluctant to pay directly for operation and maintenance. Donors do contribute indirectly through counterpart funds from balance of payments support, but there remains an imbalance between the resources available for project commitments and those available for nonsalary current costs.

There are too many donor-driven projects, making it impossible to develop and implement a coherent, consistent policy for the sector as a whole.

Local management capacity is undermined by the need to service donor projects. The proliferation of donors, of projects, and of visiting missions to discuss projects absorbs a disproportionate amount of the time of senior managers in African governments. Donor projects tend to absorb many of the best staff, while heavy use of expatriate advisers tends to reduce local ownership and inhibit the building of local skills and experience.

Donor projects may escape budget disciplines. Too many donors go directly to line ministries rather than going, as they should, to the ministries of finance and planning. Donors often fail to give sufficient reporting of commitments and disbursements on the projects that they are implementing. This lack of financial information makes it very difficult for the central economic ministries, or line ministries, to plan resource allocation in line with their priorities.

A project approach tends to lead to inconsistent standards of provision across the sector as a whole. Islands of excellence within project areas are surrounded by oceans of underprovision elsewhere. Where donor projects exist, they draw in most of the available funding and most of the management attention. Outside the areas where there happen to be projects, there is very little.

Efforts at Budget Management Reform

Governments and donors are trying to overcome some of these problems by moving away from a project approach toward general financial support to implement agreed sector policy and expenditure plans. But to make this possible at the sector level, it is also necessary to develop a framework of overall improvements in budget management. The Special Programme of Assistance to Africa (SPA) has been giving some attention to this and has produced draft guidelines that were recently discussed with African officials at a meeting in Malawi organized by the German GTZ (Gesellschaft für Technische Zusammenarbeit). There are plans to hold similar discussions with Francophone countries. I should stress that the SPA donors are not trying to impose these guidelines without first fully discussing them with our African partners and taking their views into account.

What are the main features of these guidelines? First, governments should be asked to prepare medium-term budget frameworks, rolled forward annually, and to discuss the frameworks and their priorities with the donors. This does not mean that donors will try to dictate priorities, but they do have a legitimate need to be consulted. Donors need to be satisfied that the substantial donor flows are balanced by sufficient efforts to raise domestic resources, that they actually lead to additional development spending, and that spending priorities reflect development goals and an efficient balance between capital and the recurrent budgets.

Second, the guidelines ask governments to also prepare action plans for achieving acceptable standards of budget management and accountability. It is recognized that financial management systems are weak in many countries and that one cannot overcome those problems overnight. A coherent medium-term approach is needed. At present, efforts to improve financial management are often fragmented, driven by donor technical assistance, and with significant waste and duplication.

Third, the guidelines envisage that progress in implementing both the medium-term budget framework and the action plan for improving budget management should be monitored by governments, as well as reported on and discussed both with donors locally and at country Consultative Group meetings.

Finally, donors on their side also have a responsibility to respond to these efforts to improve the budget management process. Where governments embark upon a credible process of reform, donors should respond by only supporting programs and projects that are included within the budget framework, working toward channeling their assistance via the budget, and making use of the governments’ own financial management procedures. The broad sector approach provides, in principle, a convenient instrument for doing this. Conversely, where the budget management is not being reformed, it will be difficult to introduce sector programs.

Main Features of the Broad Sector Approach

The main features of the broad sector approach are, first, that there should be a medium-term expenditure framework that determines the size of the resource envelope for the sector. It is essential that the resources to be made available to the sector program from both government and donors are determined by careful consideration of overall national priorities.

Second, the government should lead a process for drawing up a comprehensive program for the sector and agreeing to it with the donors. The program needs to set out clearly:

  • the policy framework for the sector;

  • objectives and performance measures over the period of the sector program;

  • the institutional responsibilities for implementing the program, together with an assessment of capacity of relevant institutions and proposals for strengthening them as necessary;

  • financial arrangements, covering budget preparation and agreement, arrangements for appraisal and approval of subprojects and components forming part of the strategy, procurement and disbursement procedures, and accounting and audit arrangements;

  • management arrangements, including reporting and monitoring; and

  • details of the expenditure program on both the capital and the recurrent side, including identifying the recurrent cost implications of the capital investment budget. In practice, the sector program is likely to require that a rolling budget be prepared, specified in detail for the current budget year, and rolled forward annually.

Another key feature of the approach is that all significant donors to the sector should provide their support to the sector as part of the sector program, preferably using common procedures.

Zambia’s Health Sector Investment Program

The Zambia Health Sector Investment Program was one of the first of the broad sector programs to be established, and it is useful to compare it with the blueprint.

Zambia’s sector investment program has a very clear policy framework in the form of the strategic health plan. There is strong leadership and commitment to the plan at the political and official level. There is a very clear focus on decentralization, with a bigger proportion of the health sector budget going to support district health services rather than supporting the central hospitals. All of the main donors are essentially committed and supportive of the program.

However, there are a number of divergences from the blueprint. First, there is no real medium-term expenditure framework for the health sector. Zambia is struggling to restore macroeconomic stability to the economy and is operating on an annual cash budget. This makes it extremely hard to judge the level of health services that can be sustained in the medium term. Moreover, there is no clear procedure for adjusting the sector program in the light of developments in the macroeconomy.

Common donor disbursement procedures are limited so far to district services, and even the support for district services is not fully integrated within the government budget. Donors are putting in resources through common procedures for part of the program, but not for the program as a whole. Some of that funding is in practice additional to the budget, and there is no clear procedure for adjusting donor commitments according to overall budget developments. As a consequence, there is a serious danger that the sector program may prevent rather than assist sound prioritization of the budget as a whole.

The World Bank has argued against significant use of long-term technical assistance in sector programs, on the grounds that this reduces ownership and prevents the building of local capacity. However, this is a somewhat controversial issue between the World Bank and the bilateral donors, with the bilateral donors more inclined to support the use of technical assistance, provided it is within a context of strong government leadership.

A number of steering committees have been established to implement the Zambia health program, with government and donor representation. Donors who are not locally represented have tended not to be represented on these committees—for example, the World Bank does not participate. It is not clear to what extent this has so far been a problem, but it may suggest that the level of monitoring needed by such complex programs is easier to achieve with strong representation in the field.

Key Factors for Success

What are the key success factors that emerge from the experience to date? First, strong leadership from the government side, including political commitment to tackle the key problems of the sector. This typically needs to encompass institutional reform, appropriate staffing levels and remuneration, and clear prioritization of resources. In the absence of leadership willing to grapple with these issues, or of a donor community willing to insist that they be addressed, the sector approach can end up perpetuating unsuccessful policies and approaches. In the absence of accompanying reforms, the approach could become little more than an instrument whereby the donors push heavier resources into their favorite sectors, typically health and education, leaving the country with bloated social sectors that are still spending resources inefficiently, unsustainably, and with little effect.

Second, macroeconomic stability and predictable budgets are fundamental to judging sustainability, and it is doubtful that sector programs should be attempted in their absence.

Third, securing adequate commitment and motivation of those required to change their behavior requires not only political leadership but may also need to be supported by an effective process of civil service reform and capacity building. In the Ghana civil service reform program, the intention is to prepare performance improvement programs for individual departments and ministries. Sequencing of this civil service reform program is intended to give early priority to those sectors that are going to be the subject of sector investment programs to ensure that problems of capacity and motivation are addressed.

Fourth, common disbursement arrangements are impossible unless donors have confidence in the accounting arrangements. It is probably realistic to proceed cautiously, adopting common disbursement arrangements on those parts of the program where accountability is reasonably strong and where costs can be saved. There are probably strong economies of scale for using common procedures to support local level services, and the Zambia health experience suggests that the accounting problems can be addressed successfully. At the other extreme, it may be rather less harmful to continue to rely upon donor expertise in major offshore procurement, especially where local capacity is lacking.

Fifth, the approach needs donors’ willingness to merge their efforts and accept lower visibility of their own funds. This will be helped by establishing attractive and viable real world objectives. Donors will have less need to put their Union Jack or their Tricoleur on the side of a clinic or a school if they can say instead that their finance has contributed to a program that has raised national educational attainment or has reduced the infant mortality rate by some measurable amount.

Finally, we need to recognize that it takes a long time to put the sector programs together, and donors should not force the pace on the government or on other donors. If a donor arrives with a planeload of advisors for a three-week mission and expects to sign a loan at the end of it, there will not be much local ownership and it will not be a very soundly based program.

Role of Conditionality

The sector program approach is part of an evolving consensus on new approaches to conditionality. It is increasingly recognized that conditions are more likely to be complied with when the government agrees that the proposed action is necessary. Conditionality usually fails when conditions are imposed as a cost of obtaining a loan or grant. This is especially true when conditionality is applied to difficult institutional and management reforms of the type that are typical of the sector approach, and where meaningful conditions are also much more difficult to formulate.

There is a hierarchy of conditions. The overall level of aid to a country increasingly depends on a general assessment of governance and economic and social policies. At the macroeconomic level, balance of payments assistance is dependent on adherence to a reform program agreed with the Bretton Woods institutions. Sectoral programs come below this level. They are usually conditional upon implementing agreed sector policies and programs, but without explicit macroeconomic conditions. Project aid typically has only project-level conditions.

The level of commitments at each level within this hierarchy probably depends on conditions at the higher level being satisfied so that, for example, a new project commitment would not normally be entered into in the absence of a satisfactory sector policy. Sector program commitments may well depend on a macroeconomic framework being in place. However, once that commitment has been made, the implementation of a sector program ought to depend only on sector-level conditions. Support for a medium-term sector program ought not to be interrupted because the IMF program has gone off track, for example, unless one of the consequences has been that sector commitments are also not being met.

Sector programs should be thought of as being more concerned with partnership than conditionality. They should set out what all partners agree to do in order to develop the sector:

  • agreed objectives;

  • resources and who will provide them;

  • institutional responsibility;

  • the policy framework in which the program will operate;

  • the procedure for monitoring and course correction; and

  • the accounting and financial management arrangements.

All partners accept obligations: the government to implement the program, and donors to provide resources if the program is implemented as agreed. Both sides make commitments.


The traditional donor approach to aid, via balance of payments support and project aid, obscures how additional resources are actually used by governments. Donor resources are fungible, and overall budget management should become an increasingly important factor in allocating aid resources. Countries that set out attractive medium-term budget frameworks and make progress toward achieving good standards of budget management are well placed to mobilize donor support for their priority programs. In these circumstances, sector investment programs are a particularly appropriate instrument for donor support. However, great care is needed to ensure that the sector approach does not raise aid dependence. Domestic resource mobilization needs to increase over time in Africa to permit a rising share of donor resources to be used for development, not recurrent costs.

Peter Harrold and Associates, The Broad Sector Approach to Investment Lending: Sector Investment Programs, World Bank Discussion Paper No. 302 (Washington: World Bank, Africa Technical Department, 1995).

    Other Resources Citing This Publication