1. The Department for International Development (DFID) is moving away from delivering aid through discrete projects towards channelling support through national government budgets. This move is largely supported by many of the larger UK aid agencies and by developing country governments.[2] In 200-07 it provided £461 million in budget support to 13 countries (Figure 1). DFID has set out a range of benefits it seeks from budget support, including improvements in service delivery leading to reduced poverty and strengthened government systems within developing countries.[3]

Figure 1: Countries where DFID provided budget support in 2006-07

Note 1: Rwanda normally receives general budget support but recorded no expenditure for 2006-07 as a disbursement was brought forward into financial year 2005-06.

2.  Budget support has led to increased expenditure on services to benefit the poor in six out of nine countries where budget support programmes have been evaluated.[4] It often supports developing countries in making education and health services free to allow better access for the poor.[5] There have been clear increases in the quantity of services delivered in many countries receiving budget support. For example, in Rwanda, between 2000 and 2007 the percentage of children enrolling in school increased from 72% to 95%.[6] In Ethiopia, the number of children at school doubled between 2000 and 2006.[7]

3.  However, such improvements cannot be attributed solely to budget support, but also reflect the efforts of the governments of the developing countries, other donor projects and prevailing economic conditions and growth rates. Many countries receiving budget support have increased their primary enrolment rates significantly (Figure 2), but so have some countries which have had no budget support. In addition, the modest increase in Ghana's enrolment rates and the decrease in the rate in Vietnam show that the pattern of benefits from budget support is by no means simple.

Figure 2: Progress in primary school enrolment rates in countries where DFID has a significant bilateral programme

Source: United Nations Millennium Development Goals database and C&AG's Report

4.  Donors have often focused attention on increasing the quantity of service delivery in line with the Millennium Development Goals. These Goals measure, amongst other things, progress in education through enrolment rates rather than educational attainment. As services have expanded to reach more people, the quality has often declined. For example, in Rwanda, the teacher/pupil ratio has increased from 54:1 in 2000 to 70:1 in 2006, and shortages of textbooks and reduced school hours are common as a result of the increased demand.[8] DFID has tried to address quality issues through targeted support. For example, it is working in the education sector in Zambia and Rwanda to address rising class sizes through recruitment of more teachers.

5.  Budget support aims to improve the financial management systems of developing countries so that they can spend their government budgets more efficiently and effectively. Making such changes to government systems takes time. DFID's assessments of financial systems tend to show positive improvements, for example, in adopting reform measures.[9] DFID drew attention to specific cases of reforms of financial management such as improving tax collection and increasing competition in procurement. For example, Mozambique has introduced a new procurement law and has carried out an independent census of public sector payroll.[10] However, although the overall scope and number of reforms has often been good, implementation has been slow and the reforms have often had limited impact. In Ghana, for instance, serious weaknesses such as poor budget control had persisted despite reform measures.[11] DFID reported that it had delayed budget support payments in a minority of countries, including Uganda, Ghana and Sierra Leone, where it had not been satisfied with progress in financial management reforms.[12]

6.  So far budget support has limited impact on the actual performance of financial management systems. Donor support has made little difference to the sustainable capacity of those bodies delivering services. Broad based indicators, capturing the main aspects of financial management, show that public financial outcomes have not demonstrably improved between 2000 and 2005 in countries where DFID has a country programme.[13] The lack of improvement was evident whether or not DFID provided budget support during this period. In Uganda the number of indicators where benchmarks were met actually decreased between 2001 and 2004.[14] Overall those countries which received budget support were performing at the same level as those which did not.[15]

7.  Reducing transaction costs was one of DFID's stated aims when providing budget support, but the effect on such costs has been under-researched. DFID's administration costs have generally declined relative to the amount of aid provided, although in half of countries costs in absolute terms rose over the period.[16] DFID and other donors have not yet collected information on the transaction costs incurred by developing country governments. DFID reported that setting up budget support has been more costly than expected due to the need to agree priorities and monitoring with all donors, but it expected costs to decline over the medium term. It agreed to monitor the impact of budget support on recipients' costs in the future although it noted that attribution of costs to budget support is hard since much aid is still provided through projects and other channels.[17]

8.  DFID recognises that economic growth is key to achieving its overall aim of reducing poverty. It believes that budget support can assist the developing country economy to maximise the potential for growth, for example, by promoting good macroeconomic management and encouraging increased productivity.[18] In Malawi, however, an independent evaluation found that the government had a poor grip of the macro-economic situation and that budget support actually worsened both macroeconomic performance and the steady flow of aid funds.[19] Although budget support has been provided by DFID for ten years in some countries, direct links between the support, economic growth and poverty reduction are hard to detect. Many studies on links between growth and aid in general have been inconclusive or contested.[20]

9.  Budget support is DFID's preferred way of giving aid.[21] DFID can point to areas where benefits have arisen, but not the scale of benefit attributable to budget support. It also has not established a way of assuring itself that more could not have been achieved through other types of aid.[22] Without such assurance DFID cannot provide evidence of its cost-effectiveness. This lack of evidence raises doubts over the validity of DFID's claim that budget support is the most powerful instrument for achieving broad development outcomes.[23]

10.  DFID has not focused sufficiently on analysing and comparing costs to ensure that it gets maximum impact from its funding.[24] After up to ten years experience of budget support in thirteen countries, DFID offered just one example indicating that budget support might be cost-effective. This was that the unit construction costs of government-built schools in Mozambique were only half of those built by donors. DFID has no broad base of such information. For example, the National Audit Office found that the DFID offices visited did not capture information routinely on measures of efficiency such as the unit costs of delivering services through different providers.[25] In February 2008, DFID established an Investment Committee to improve its approach to identifying the cost-effectiveness of its choices. That Committee plans to develop clear benchmarks for unit costs, average costs and rates of return and to ask DFID country offices to assess their plans against these benchmarks.[26]

3   Q 16 Back

4   Qq 33-35 Back

5   Q 32 Back

6   Q 7 Back

7   Q 29 Back

8   Qq 7, 29, 68 Back

9   Q 42 Back

10   Q 107 Back

11   C&AG's Report, Figure 12 Back

12   Q 12 Back

13   Qq 42, 106 Back

14   Q 12 Back

15   C&AG's Report, para 3.5 Back

16   Q 17 Back

17   Qq 121-123 Back

18   Q 69 Back

19   Qq 10-11 Back

20   C&AG's Report, Appendix 6, page 61 Back

21   Q69; C&AG's Report, Executive Summary, paras 1, 2.10 Back

22   Qq 72-73 Back

23   Qq 3, 5, 75 Back

24   Q 119 Back

25   C&AG's Report, para 4.14 Back

26   Q 128 Back