With the focus in international aid very much on efficiency and outcome, any model that appears to guarantee both is bound to pique the interest of donors, and the Cash on Delivery Aid (COD Aid) approach developed by Nancy Birdsall and others at the Center for Global Development is certainly doing so. Although first outlined around two years ago, the idea has picked up momentum recently, chiming as it does with the demand for a bigger bang for each buck as public spending is slashed in response to the global financial crisis.
The approach sees aid disbursements based on recipients meeting pre-agreed outcomes for that particular programme or sector. So, for example, funding for education might be based upon agreed annual increases in enrolment rates, or enrolment of girls, or achievement levels. Donors would focus solely on the targets, leaving the recipient free to devise its policy or strategic plan for achieving the outcome.
COD Aid appears, at first sight, an obvious strategy – why hasn’t this been thought of before? Shouldn’t continued aid payments always be reliant on good progress being made? After all, isn’t that what monitoring and evaluation is for? And indeed, COD Aid is not entirely original in its linking of aid disbursement to the achievement of results. But in linking to a specific objective or target, it is a much more focused and comprehensively outlined approach, with the potential, so its proponents argue, for considerable gains in output.
Its champions – and there are many (especially within DFID) – highlight a number of expected gains from such an approach, just three of which will be dealt with here: it should increase recipient ownership over programmes and policy; it should make aid more regular and sustainable; and it should achieve higher gains with its focus on measurable outcomes. Is COD Aid really that good?
Improving recipient ownership
Central to the claims for COD Aid is the assertion that it will improve recipient ownership over programmes and policy. Having mutually agreed the outcome against which continued disbursements will be measured, the recipient is free to devise the strategy for meeting those outcomes without external interference. In theory, the aid could be spent on an entirely different area from the programme, provided the outcomes are still met. So if a government managed to increase female primary school enrolment, or attainment, levels to the pre-agreed levels, using the aid to purchase a new presidential jet would not be a problem. Whilst extreme, the example is used to suggest that, as with Direct Budgetary Support (DBS), COD Aid is not about tying the hands of the recipient and forcing it to spend on externally-defined areas, but enabling the government to take the measures it sees fit to achieving the ends to which it has committed itself.
COD Aid is not designed to replace DBS (which is linked to improving planning and implementation strategy for government), but to achieve outcomes in specific programme areas. As such, it may be an improvement on ownership within programme financing. However, I would suggest it is misleading to see COD Aid as a tool for improving ownership in a real sense. Outcome-oriented systems impose targets on recipient governments, and those targets surely undermine any meaningful sense of ownership. We need only look to the example of target-driven policy in the UK health service to see how, far from giving ownership to providers (or recipients in the language of COD Aid), it pulls power back to the centre, reducing the scope for local initiative in setting priorities, and creates a technocratic target-focus (to the exclusion of all else) approach. Yes, in theory the outcomes are ‘mutually agreed’. But pro-poor policies and programmes are already supposed to be mutually agreed, based upon civil society engagement, and reflecting local needs and assessments of structural and social barriers. In practice, issues of ownership remain highly problematic, and there is no real sense of how COD Aid offers a substantial improvement in this area.
Ultimately, COD Aid is a tool for donors – to better justify and defend their aid spending; to monitor spending and policy in recipient governments without being heavy handed or being seen to be overly prescriptive. This does not undermine the case for COD Aid – and it is important to note that its proponents are not describing it as a panacea, or a replacement for all aid channels, but as an additional tool that might achieve particular gains in certain areas. Moreover, they call for pilot projects to test the claims made for it. However, it seems that basing claims for it upon the language of ownership is deeply problematic.
The then Under-Secretary of State in the Swedish Ministry of Foreign Affairs, Gun-Britt Andersson, argued ‘ownership’ does not readily translate into Swedish, seeing the two words ‘responsibility’ and ‘anchoring’ as the closest to what is usually understood by the term. As my colleague Prof Christopher Cramer has written:
Anchoring especially conveys the idea that responsibilities come with ownership and that development might entail a dynamic process through which the project or programme truly ‘becomes’ one’s own (i.e. belonging to the developing country) and then loses the obligation inherent in an asymmetric donor-recipient relationship (Cramer, 2002).
Can a target-driven model, in which the power over the setting of those targets lies ultimately with the donor, really lead to such anchoring? Ultimately this depends on how outcomes are set, and the role of the recipient in this. But the fear must be, with donors concerned to demonstrate the efficiency of their aid programmes, and the increased prominence of results-based programming, the anchor will remain very firmly tied to donors.
Improving aid stability
On first glance, a model that proposes to reduce or even withdraw aid if the outcomes are not achieved would appear to be deeply destabilising, rather than working to reduce the recognised problem of aid volatility (and the knock on impact on the ability of recipient governments to plan over the long-term). Surely the chasing of targets, and the numerous factors that might impede the successful achievement of them that lie outside the control of the implementing agency, are incompatible with the hoped for stabilising of aid regimes.
Supporters of COD Aid argue that once the outcomes have been agreed, and the method for monitoring those outcomes, donors (whether public or private) can pool their resources without imposing new structural demands upon the recipient. It would also be easier for other donors to join without requiring renegotiation or imposing additional channels for reporting, monitoring and evaluation. The recipient government knows that aid flows are linked to the outcome, not to changes in donor policies and priorities, and can therefore focus on achieving those outcomes without fear that the aid flows might stop with a change in government.
Critics argue that the volatility engendered by the need to hit the targeted outcomes would be more destabilising. The record of governments in rich industrialised countries in meeting their own stated outcomes is far from untrammelled success, yet they face no sanctions (albeit the sanction of being voted out of office if the gap between promises and reality is too wide). Moreover, given the complexity of structural poverty and instability, outcomes may be missed for reasons that no government could be expected to have anticipated, or could be expected to have the capacity to deal with. In reality, it is unlikely that donors would simply withdraw aid if outcomes were not met once, or there was good cause for missing the targets. Under one strategy being proposed by the Clinton Health Access Initiative for malaria control in low-endemic settings, governments would receive funding provided malaria levels remained below a set ceiling. However, should that ceiling be breached, funding would not be withdrawn, but greater control exercised by donors over policy and strategy until performance against the agreed outcomes was again maintained.
This does, to some extent, address the potentially destabilising effect of withdrawing aid following failure to achieve the agreed outcomes. Switching from COD Aid payments to controlled input financing in order to manage the transition back to performance-related aid would provide a safety-net of sorts. But it does lead to further questions as to how much COD Aid really does enhance ownership. With the ever present threat of donors stepping in to reassert controls (or withdraw funding) recipient governments are likely to be less willing to try risky innovative approaches for fear of the consequences of failure. The serried ranks of consultants and donor organisations will, in all likelihood, continue to ‘advise’ on the best approach to meeting outcomes.
This raises a question about who benefits most from COD Aid. Is it really recipients, for whom other mechanisms for ensuring aid stability and programme ownership might well be preferable? Or, despite the rhetoric, is this primarily a tool for donors – allowing for greater control over the purposes for which its aid is allocated, and for measuring the impact of its aid?
Improving aid gains
Perhaps the main benefit of COD Aid is its focus on measurable outcomes. The progress of governments in increasing the number of women attending ante-natal clinics, number of children of a certain age vaccinated, overall prevalence of malaria rates would be easily monitored, and the efficiency of a particular aid channel therefore measurable. It should also, if it works well, improve recipient monitoring and reporting mechanisms – an important gain in settings where a dearth of official records can contribute to the exclusion of individuals from vital welfare services.
But the benefit of results-oriented programmes are not clear-cut. Do they provide citizens of the recipient state with a set of measures against which they can measure the performance of their government (with the effect of potentially increasing accountability); or do they actually constrain governments from addressing the demands and wishes of the population in their determination to meet externally-set outcomes? Results-based management can constrain, creating a strait-jacket of its own and limiting scope for innovation and flexibility.
So a particular outcome might be met, and it might have positive results. But it may nevertheless fail to address underlying structural causes of poverty and inequality, and constraints in escaping these. The record of donors in linking their programmes to wider aims and objectives is not good (see, for example, my post on procurement policies), and COD Aid, in itself, does not offer a protection against more of the same.
So what of COD Aid? The problem lies less with the idea, than the purpose and use to which the idea is put by politicians and donors. A new mechanism for delivering aid, albeit as yet relatively untested, is to be welcomed. And it certainly has the potential for gains in specific areas that are particularly suitable to a result-based management approach. Certain health programmes, for example, could see substantial gains from COD Aid strategies. But some of the claims made for it do not stand up to scrutiny. Its claims for ownership are questionable, at the least, as is the assertion that it offers a more stable mechanism for aid flows: it could improve stability, but it could also undermine it – COD Aid itself does not guarantee to reduce volatility, but they way it is implemented (which is surely true for many aid mechanisms?). It is based upon the notion that results-oriented aid is better and more efficient. But this, too, is problematic: in some cases and sectors it might be, in others less so.
Nancy Birdsall, and the others who contributed to the framing of this idea, are not claiming this to be the magic bullet of aid. But the translation of COD Aid through politicians and donors is making some of the claims appear more certain, more obvious, more simple than its creators would have. It is an interesting idea, one that deserves to be tested and rolled out where proved effective. But it needs to be one tool amongst many in the efforts to reduce poverty, instability, inequality and inequity. Few would disagree, I hope. But politicians especially, in their constructions of simple narratives, need to be wary of making unsupportable claims, or seeing measurable outcomes as the sole holy grail of development financing.