House of Commons - Public Accounts - Twenty-Seventh Report

Hansard

Archives

Research

HOC Publications

HOL Publications

Committees

Select Committee on Public Accounts Twenty-Seventh Report

2 How DFID appraises and manages budget support

11. Over the past five years, DFID has increased its use of budget support from £268 million to £461 million. It believes that budget support is particularly effective as it provides an opportunity to shape national policy and the overall allocation of the government's resources. It also allows DFID to be involved in much broader issues around public financial management and accountability.[27] Although the initial costs of providing budget support remain high due to the need to set up monitoring systems and to co-ordinate with others, budget support also allows DFID to increase the amount of its funding at little additional cost. Given the pressure on DFID to spend more funds on development without a corresponding increase in staff resources, budget support is an attractive option.[28]

12. DFID has not set clear parameters for deciding which countries are suitable for budget support.[29] The underlying principle is simply whether the benefits outweigh the risks, and this principle allows a great deal of flexibility of approach by country teams. There is no maximum level of risk which DFID is prepared to take in providing budget support. Rather, DFID takes an overall assessment of the risks and benefits of providing aid in each circumstance.[30]

13. DFID has bilateral programmes in 120 countries, although it spends nearly three quarters of its funds in 25 target countries.[31] Some 14 of those 25 countries receive budget support. So DFID does not consider 90% of the countries in which it has bilateral programmes suitable for budget support. There are various reasons for this view. Some countries are middle income countries, so any contribution DFID made through budget support would not be significant enough to have real impact. In countries with political regimes such as Zimbabwe, supporting a government budget would not be appropriate. In some other countries, DFID's involvement is limited to humanitarian assistance. Finally, in some countries such as the Democratic Republic of the Congo, DFID has assessed the risks to be too high.[32]

14. Decisions about which countries receive budget support sometimes appear arbitrary. For example, DFID provides budget support in countries such as India, Vietnam and Pakistan which are emerging countries with significant economic growth rates.[33] There are of course significant poverty challenges remaining in each of these countries but the reason why these particular countries should merit budget support more than some other countries is not clear. Similarly when considering whether the risks outweigh the benefits, it is hard to understand why DFID denied budget support to Kenya because of high perceived levels of corruption, but gives budget support to Sierra Leone and Cambodia, which have higher perceived corruption levels. In the case of Sierra Leone, DFID explained that it hoped that budget support would prevent the government from failing, which in turn would hopefully prevent a re-emergence of conflict. DFID consider conflict prevention to be cost-effective.[34] Whether budget support could have prevented conflict in Kenya remains an open question.

15. When appraising prospects for budget support, DFID requires its country teams to make an assessment of the potential benefits and risks. However, only two appraisal documents pulled together these two elements to provide a clear and explicit overall assessment to senior staff and ministers. The absence of the assessment means that it has been difficult for DFID to control the quality of individual decisions.[35] DFID accepted that it needed to improve the quantification of the benefits and risks, and has recently revised its guidance on how to assess the expected benefits of budget support.[36]

16. Corruption is a risk when providing any type of aid. There are particular risks arising from spending funding through national systems which are generally weak in key areas such as financial management and accountability. DFID stated that it manages the risks of funding not reaching its intended beneficiaries by establishing in advance clearly what results it wants the funding to achieve. It also carries out Fiduciary Risk Assessments to look at the overall financial capacity in these countries, including the budgeting, audit and accountability systems, procurement arrangements and financial reporting.[37] But such assessments generally focus on the strengths and weaknesses of these systems rather than the significance of any weaknesses for potential corruption or inefficiency.

17. By its nature, corruption is often hidden and hard to quantify, but DFID does not make full use of available information.[38] For example, it rarely makes any quantified estimates of the impact of weaknesses in systems on possible resources wasted or used for corrupt purposes. In some countries, DFID or others have commissioned Public Expenditure Tracking Surveys to track money through from centre to local level. Such surveys can identify some aspects of corruption or problems with poor financial reporting. Surveys conducted in some DFID target countries suggest that an average of approximately 7% of funds cannot be traced. Such surveys are not always carried out, however, or if they have been, are not always used by DFID in risk assessments.[39] DFID also only makes limited use of state audit reports, which also give an insight into weaknesses in record keeping, inefficiencies and corruption.[40]

18. DFID has not estimated how much funding through developing governments is wasted or used for corrupt purposes, but the estimates of others are worrying. For example, in Tanzania and Uganda other bodies have estimated that 20% of procurement expenditure is lost through corruption.[41] In 2006-07, DFID provided £90 million and £40 million respectively in budget support funding in these countries.[42] More recently, an external audit of the Bank of Tanzania showed that US$100 million had been misused,[43] and in February 2008, Tanzania's prime minister and entire cabinet resigned after being implicated in corruption over an electricity contract.[44] DFID has not suspended budget support as a result of such incidents of corruption, only because of concerns about a lack of progress on key reforms.[45]

19. In some cases additional safeguards have been introduced to ensure that funding reaches the most appropriate areas. For example, in Ethiopia funding to government included controls to make sure that aid was spent on poverty reduction. It did this by ensuring money would be disbursed to regional government for use in specified sectors, introducing checks on expenditure and processes at local level and by funding civil society organisations to increase scrutiny by Ethiopians of how the budget was spent.[46] DFID may ask for more scrutiny work if it has concerns in a particular area. For example, in Mozambique DFID agreed with the government that, while providing funding for the roads sector, there would be a procurement plan for roads and an ex-post audit every year.[47] Overall, however, country teams have not introduced sufficient safeguards to ensure that UK taxpayers' money has been spent well.[48] DFID agreed that it needs to be more proactive in its safeguards and has recently issued new guidance.

20. DFID's monitoring and management of its budget support programmes is limited by gaps in data on the services delivered, and on outputs such as the number of schools built or teachers trained. DFID has played a key role in trying to build statistical capacity in developing countries and encourage others to collect better data. It is active in 10 of the countries where it provides budget support to improve available data and monitoring. DFID believes that it is important to adopt a multilateral approach to improving statistics and has also worked with the World Bank to encourage developing countries to outline national strategies to improve their statistical systems, providing £40 million of funding.[49]

21. In the meantime, however, DFID finds itself operating in countries without systems capable of providing a full view of progress. DFID does not insist on a minimum level of information on the use of government funds before providing budget support. In countries such as Rwanda and Zambia, systems for measuring expenditure were so weak that DFID could not be confident about whether expenditure trends were benefiting the poorest. The Zambian government was moving to a new system for tracking expenditure at the time. Many countries receiving budget support from DFID have poor audit trails, which make it difficult to gain assurance that funds have been used for the purposes intended.[50]

22. Being able to measure the results of budget support is critical to making judgements about where, when and how to use it. DFID has not been systematic in its approach to monitoring budget support. It claims that this is not an issue of substance but of format, but accepts that improvements are required to make its budget support programmes comply with good practice.[51] Despite this view, DFID's monitoring is often weak in key areas. For example, 15% of indicators had no time-bound targets and 22% had no baselines.[52] DFID normally sets out what it wants its budget support programmes to achieve, but four out of six internal reviews of budget support programmes considered progress against only half or fewer of its objectives.[53] In India DFID did not have access to key financial management information collected by the World Bank, and did not carry out its own monitoring, although access to such information is now much better.[54] DFID expects to remedy soon the weaknesses in the design of monitoring frameworks and matching of objectives to indicators for new budget support programmes. It will take more time to improve existing programmes since such changes will need to be agreed with other donors and developing country governments.[55]

23. Monitoring risks is also important. But only three quarters of the Annual Statements of Progress which seek to identify major changes in risk were completed as required.[56]Only a sample of such statements is reviewed by headquarters which does not encourage full compliance or encourage country teams to give them adequate emphasis.[57]

27 Q 16; C&AG's Report, para 1 Back

28 Q 59 ; C&AG's Report, para 6.17 Back

29 Qq 102, 105 Back

30 Q 41 Back

31 Qq 5, 75, 132 Back

32 Q 132 Back

33 Q 101 Back

34 Qq 14, 36-38 Back

35 Qq 43-45 Back

36 Q 45 Back

37 Qq 18, 62 Back

38 Q 21 Back