Research Brief
The Fungibility Problem

Budget Support, Aid on Delivery or Project Aid?

One persistent concern raised during discussions of whether aid effectively promotes the goals of donors is fungibility, that is; the possibility that aid is used in ways not intended by donors when disbursing the funds. It could be used to lower taxes, to fund projects in a different sector, or simply to line the pockets of corrupt officials. In the WIDER Working Paper ‘Fungibility and the Choice of Aid Modalities: The Red Herring Revisited’ Stefan Leiderer looks at three different aid modalities and assesses the impact each has on aid fungibility.

Project aid, budget support, and aid on delivery

The Paris Declaration on Aid Effectiveness (2005) addressed the concern that international development assistance has often been compromised due to unintended side effects caused by the way aid had traditionally been provided. Traditional project-based aid has been criticized on a number of grounds; focussing on individual projects leads to high transaction costs and often only generates short-term improvements in specific sectors. Furthermore, as Leiderer points out in his Working Paper, project-based aid has been criticized for not giving recipient countries enough ownership over the development process, and for following donor rather than recipient priorities. In addition to this, the parallel systems for managing aid resources established by the donors have the potential effect of undermining the recipient country’s own administrative systems.

In Herat, Afghanistan, women line up to collect bags of split chick pea, wheat, and cooking oil being distributed by the UN World Food Programme (WFP). © UN Photo/Eric Kanalstein
In Herat, Afghanistan, women line up to collect bags of split chick pea, wheat, and cooking oil being distributed by the UN World Food Programme (WFP). © UN Photo/Eric Kanalstein

In an attempt to remedy these problems, donors in recent times have increasingly moved towards programme-based approaches (PBAs) which emphasize recipient country ownership and donor coordination. The most consistent form of PBA is general budget support (GBS). GBS involves supporting national development and poverty reduction strategies in developing countries by contributing money directly to the recipient country’s treasury. GBS is intended to improve the aid process in a number of ways. In particular GBS should lead to a reduction in transaction costs, an increase in the predictability of funding, and lead to the kind of government-wide policy that cannot be achieved by stand-alone aid projects. Leiderer points out that in theory GBS should also increase government accountability, improve the recipient country’s public financial management, and encourage a greater focus on mid-term effectiveness by focussing on national development objectives rather than donor driven priorities.

Regardless of the potential benefits of GBS the move in that direction has garnered criticism from the aid community and the general public in donor countries. Opponents argue that aid given in the form of GBS is fungible and as such likely to end up in the pockets of corrupt government officials or to be spent on things unwarranted by the public in donor countries. This criticism led to the search for aid instruments that could ensure that recipients used the provided resources in line with donor preferences while at the same time avoiding the problems associated with project-based aid.

One such instrument is aid on delivery (AoD). AoD is a form of aid that is disbursed proportionally to the achievement of pre-defined goals by the recipient country. This is intended to allow donors to fund expenditure on their priorities without having to get involved in implementation. The general view amongst policy makers seems to be that AoD does provide donors with much more influence over how aid is spent, however there is little empirical research on this issue. Leiderer attempts to fill this gap by assessing whether the risk of fungibility does significantly differ between these different aid modalities and by suggesting ways that donors can deal with this. He does this by employing a formal model of the government’s allocation decision.


The first question Leiderer seeks to answer is whether aid delivered in the form of GBS is really more fungible than project-based aid. Initially it may seem logical to suppose that this is the case as project aid is spent directly on donor priorities whereas GBS goes into the recipient country’s budget to do with as they wish. However Leiderer suggests that this may not always be the case. If project aid is spent in sectors where otherwise the government would have to spend money then the government can simply reallocate their spending from that area in order to achieve their ideal resource distribution. In fact Leiderer argues that project aid will only be less fungible than GBS if the recipient country is highly dependent on aid or has a low initial commitment to donor priorities.

The World Bank-supported National Water and Sanitation Programme will bring better water and sanitation services to rural parts of Azerbaijan which currently face water shortage. © Allison Kwesell / World Bank
The World Bank-supported National Water and Sanitation Programme will bring better water and sanitation services to rural parts of Azerbaijan which currently face water shortage. © Allison Kwesell / World Bank

Turning to AoD it appears that a similar argument may follow. While recipients only receive AoD after achieving objectives the donors have prioritized, they could simply use the money they receive to spend on their own priorities to return the overall allocation of resources to the one preferred by the recipient government rather than by the donors. However AoD does reduce the opportunity costs associated with investing in donor priorities and as such is more effective in encouraging such spending in countries with low and moderate aid dependency.

However, all the above assumes that donors have perfect information of the recipient governments own commitment to donor priorities. If this is hidden from donors, as is likely in reality, then the recipient government can ensure their own preferred outcome by misstating their commitment. However the way in which recipient governments are incentivized to misstate their commitment differs according to the modality on offer.

  • If donors offer GBS then recipient governments have an incentive to overstate their commitment as donors will give GBS only if they believe the governments priorities are in line with their own.
  • If AoD is offered then recipient governments have an incentive to understate their commitment as donors feel the need to incentivise investment in their priorities.

In his paper Leiderer lays out a model which shows that:

  1. If donors coordinate their actions they can use these countervailing incentives to eliminate part of the fungibility problem.
  2. Donors simply need to offer a mix of GBS and AoD and decide on a level of indicated government commitment at which they will convert either the entire aid budget or just the GBS tranche to project aid.
  3. By mixing modalities in this way donors can incentivize recipient governments to self-select into the most appropriate aid modality.

Leiderer concludes that this model provides a strong argument in favour of PBAs over project aid. If aid is delivered as an appropriate mix of GBS and AoD then the problem of fungibility is no greater than it is with project-based aid, and PBAs have the advantage of reducing transaction costs and giving recipient countries greater ownership over the development process. Leiderer notes that one clear policy implication of this is that if donors want to avoid the problem of fungibility they need to focus on the coordination and harmonization of their various development programmes.