Ask the right questions
01/04/2009 – by Geoff Handley
Budget support is coming under fire. A number of national audit offices and parliamentary committees have recently published critical reports, including those in the UK, Sweden and Germany. These reports typically discuss fiduciary risks (usually defined as the risk of waste, inefficiency and corruption) and the lack of visible and attributable results.
The emphasis on accountability to donor auditors and parliaments, however, risks distorting the efficacy of budget support programmes if accountability processes do not ask the right questions. As Penrose (2008) argues, “efforts to satisfy real or perceived concerns of voters and politicians can lead to ‘audit tyranny’”. All too often, aid programmes are driven more by donors’ risk aversion and wish for clear results than by concerns for what would best serve the recipient country (Cant et al, 2008; Fölscher et al, 2008; Burall et al, 2009).
This could be addressed, at least in part, by improving the quality of legislative scrutiny of aid in donor countries. Fölscher et al. (2008) acknowledge that there is a fundamental mismatch between the long time horizons required for budget support to deliver results and the short-term focus of politicians. The authors argue that parliamentary assessment of budget support should not only consider risks, but also assess development opportunities. Moreover, parliaments should discuss the alternatives to budget support, rather than debating the pros and cons of this instrument in isolation.
A matter of effectiveness
The Paris Declaration (PD) does not take a clear stand on modalities, so long as aid is provided as part of a vaguely defined “programme-based” approach (PBA). The use of developing countries’ own institutions and procedures (“country systems”) is quite rightly seen as an important consideration for all aid modalities, not just budget support. By definition, budget support uses national public-finance management (PFM) systems. But other aid modalities – such as common funds and projects – can be designed and implemented in ways that use country systems too (Mokoro, 2008).
But some aid modalities are less consistent with the spirit of the PD in practice. For example, whenever projects or joint “basket” funds bypass country systems – as they often do – they tend to divert attention from improving those systems. Though sector-specific basket funds are often presented as transitional arrangements, they tend to become stumbling blocks, by creating self-perpetuating incentives (Kizilbash and Williamson, 2007).
There is also evidence that implementation of so-called vertical programmes through country systems can cause problems. Consider two examples from Mozambique:
- The Global Fund to Fight AIDS Tuberculosis and Malaria (GFATM) channelled funds through the health-sector basket, but because of unpredictable disbursements, the government was forced to ask other donors to frontload their payments.
- Funding provided to the education sector under the Fast Track Initiative Catalytic Fund (FTI-CF) could not use Mozambique’s procurement system, and because it was channelled through the sector basket it prevented other donors from doing so too.
In both cases, these problems undermined the dialogue between government and donors, detracting attention from enhancing the country systems and delivering services (Handley, 2008).
In order to avoid such difficulties, donors should use aid modalities that bank on country systems from the start. Obviously, their own rules must allow them to do so. The important caveat is that programmes must be designed and implemented with great care. The devil is often in the detail.
In principle, budget support looks more consistent with the spirit of the PD. In practice, however, different donors use quite different approaches. Evidence from Tanzania and Ghana suggests that the World Bank and the IMF have made progress on dropping ex-ante conditions. In contrast, bilateral donors and the European Commission continue to rely on detailed matrices of policy actions and outcomes. According to Booth (2008, p.3), “… micro-management of country policies through conditionality is still the order of the day”.
Penrose (2008) argues that the financial impacts of budget support are often poorly understood. The conventional wisdom is that budget support is merged with tax revenue to partially finance recurrent government spending, helping to pay nurses’ and teachers’ salaries, for instance. In reality, however, the governments of stable aid-dependent countries insist on covering such expenditure from domestic revenues. The reasons are that aid flows are too unpredictable, and that governments want to prevent undue donor influence.
As a result, budget support basically serves to finance public investment. This is a reason for aid having done more to build schools than to reduce pupil-to-teacher ratios. Following this line of argument, the efficiency and effectiveness of budget support depends on the quality and quantity of public investment. Policy dialogue and programme design should accordingly focus on public investment. Instead, budget-support conditions tend to focus on increasing recurrent spending and overly complicated public-finance reforms.
Accountability vs neopatrimonial politics
Budget support is meant to boost institutional development and accountability in recipient countries. Skeptics disagree, however, arguing that this aid modality helps to sustain some of the most negative aspects of informal, neopatrimonial politics. In the case of Uganda, for example, Barkan et al. (2005) state that the high fungibility of budget support cements the weak accountability and corrupt practices of “big man” politics more than project aid.
In response, Lister et al. (2006) observe that other aid modalities are also fungible, and that project aid is not immune to corruption and diversion. They also point to sensible safeguards built into Uganda’s budget system. Indeed, Barkan’s critique arguably applies to all aid to Uganda, and it could be a strength of budget support that it explicitly admits to fungibility, thus focussing attention on public-finance management systems.
Certainly, it is an unfounded, though common assumption that fiduciary risks can be avoided by not using national PFM systems. A recent evaluation of Danida’s support to education in Mozambique illustrates how funds channelled outside the national system were misappropriated (Visser-Valfrey and Cumbi, 2008; p. 29).
All told, the crucial issue is to address developing-country needs rather than to meet donor requirements. One way to achieve that would be to improve the quality of policy dialogue between donors and recipients. “Ownership” does not mean that donors should disengage from policy debates (Booth, 2008). Policy dialogue should go along with capacity building among civil servants and parliamentarians in both donor and recipient countries. Moreover, aid-agency officials’ skills must be brought in line with their responsibilities (Penrose, 2008). Attention could also be paid to encouraging the bodies that oversee aid agencies to focus on holding them to account for aid effectiveness and long-term development effectiveness rather than fiduciary risk and short-term results (Fölscher et al, 2008; Burall et al, 2009).
Ultimately, budget support and other aid modalities must serve genuine recipient-country ownership. The fundamental task is to use the best means to achieve a set of country-specific objectives (over which aid will often have limited traction). This in turn means developing a more sophisticated understanding of the political, macroeconomic and fiscal role played by budget support – and by aid in general. Donor agency systems and skills should be revised accordingly.