VoxEU Column Development

Budget versus project aid: A tradeoff between control and efficiency

Despite its many benefits, donor governments show little enthusiasm for budget aid, instead preferring to give project aid over which they have greater control. This column argues that budget aid is better than project-specific aid because it attributes full responsibility of expenditure to the recipient government, allowing voters to respond at the ballot boxes to how well the aid is used.

Budget aid has never been donors’ preferred way of giving aid. Consider Figure 1, which shows the share of total aid that donors give to the recipients’ budget rather than for any specific purpose. As can be seen, the share of aid given directly to the recipients’ coffers is declining since the mid-1990s, its 2014 value being as low as 0.75%.

Figure 1. Budget aid as a share of overall aid commitments

Notes: Budget aid as a share of total aid commitments from all 28 bilateral DAC donor countries to developing countries (OECD CRS 2016).

The low and declining importance of budget aid is much in contrast to the ‘new rhetoric on aid’, emphasising the importance of recipient ownership for the successful implementation of aid projects. In order to maximise ownership, it is hard to think of better ways to deliver foreign aid than simply transferring the money to the recipient government’s coffers, ideally without attaching conditions to it. Budget aid strengthens recipient politicians’ control over the aid, increases transparency over who is in charge, and – given that disbursements of project aid depend on a number of project-specific conditions – can reduce the volatility of aid flows (Klasen 2009). Potentially negative consequences of donor fragmentation lose importance when the recipient alone is responsible for how to spend the aid. Parallel systems in managing and evaluating projects, procurement, and accounting of aid can be unified, reducing the potentially devastating effects on the recipients’ bureaucracy (Klasen 2009). Budget aid increases accountability of the recipient governments as they alone are then responsible for delivering to their populations, rather than a diffuse mix of a number of different donors and the recipient government in concert.

Figure 1 indicates that donors are less enthusiastic than the authors of this column about the positive features of budget aid. The share of this aid is low, and falling. Arguably, the reason is political. Recipient countries abuse parts of the aid received, for example by channelling it into regions that are dear to the recipient politicians (Dreher et al. 2014). In some countries receiving aid, violations of human and democratic rights abound. Rates of economic growth might be low in spite of having received substantial amounts of aid (Dreher and Langlotz 2015), embarrassing the donor. Donor politicians and aid bureaucrats supporting the budget of recipients that spend some of their resources against the preferences of the donor country population will see support for their aid policies erode. What is more, donors seem keen to be seen giving. Project aid is more visible than aid being mingled with other resources in the recipients’ budgets. It can directly be attributed to the donor, fostering good will among the recipient population and, probably as important for the donors’ aid bureaucracy, fostering support of the donor population for giving foreign aid. What is more, donors have their own geo-strategic and commercial goals in mind when granting ‘aid’. In essence, donors prefer project aid, because their preferences diverge from those of the recipients on how to use it.

None of these reasons is a good one. A donor with the recipients’ best interest in mind does not need to be seen giving (Matthew 6: “So when you give to the needy, do not announce it with trumpets, as the hypocrites do in the synagogues and on the streets, to be honoured by men”). Truly benevolent donors would respect the preferences of (democratically elected) governments rather than imposing theirs. Aid is to some extent fungible, so that parts of the project aid end in the government’s coffer anyway. However, unlike when given as budget aid, part of the aid is stuck in projects that are of little use to the recipient – the costs of the aid to the donor might easily exceed the benefits to the recipient. What is more, if aid is given for geo-strategic reasons it can hardly be expected to affect development (Dreher et al. 2016b). In fact, it is not really ‘development’ aid when the primary motive is political, but is counted as official development aid nevertheless.

Most importantly, in order for aid to be effective, the donors’ cross country information needs to be combined with the recipient-specific knowledge about domestic institutions and culture – in short, with what ‘works’ in a specific domestic environment (Dixit 2009, Besley and Persson 2011, Marchesi et al. 2011, Easterly 2015). The recipient government knows better than the donor in which sector and region the aid is most effective. The donor has complementary knowledge about what has worked in other countries, and what has not. As we argue in a new paper, when donors and recipients do not agree on how to best use the aid they will not truthfully communicate (Dreher et al. 2016). Project aid involves donor and recipient knowledge to some extent, but uses the recipient’s knowledge to a lower degree compared to budget aid. Project aid thus represents a more ‘centralised’ type of aid, with lower ownership.

Giving aid as project aid when preferences diverge can reduce development success if communication between recipients and donors is essential to succeed, and recipient information is important. In our paper, we focus on the donors’ trade-off between loss of control when giving the aid as budget support, and loss of information when giving it as project aid, in deciding about the two types of aid. Our theory suggests that the appeal of budget aid depends on the difference in donor and recipient preferences and on the relative importance of their private information. We argue that budget aid is less effective than project aid when the recipient country’s information is less important compared to those of the donor. To the extent that donors consider their own information more valuable than the recipients’, they rely more heavily on project rather than budget aid. Arguably, their evaluation about the relative importance of donor and recipient information is biased in favour of their own information, leading to a strong emphasis on project aid and thus reducing the use of recipient country information below what would be optimal.

Overall, we suggest that budget aid should be the default mode of granting aid to democratically elected recipient governments. Some recipients will use the aid for their own personal and political benefits rather than development projects. Human and democratic rights will be violated in countries receiving aid. Inequality might increase in the short term. In the longer run however, it will become obvious that the recipient country government is in the driver’s seat, rather than receiving orders from abroad. We expect voters to be better able to attribute responsibility in such comparably transparent environments, and to ultimately punish their governments at the ballot boxes should they fail to deliver some development success. In autocracies, the differences in preferences between donor and recipients are arguably too large to grant the aid as budget support, even when recipient information is important. In such environments, we see a larger role for project aid. However, we consider it unlikely that aid can promote development in such environments. To the extent that donors have development in mind when giving the aid, rather than their own goals described above, discontinuing aid for autocracies might be the better choice in the longer run.


Besley, T and T Persson (2011), Pillars of Prosperity: The Political Economics of Development Clusters, Princeton University Press.

Dixit, A (2009), “Governance, institutions and economic activity”, American Economic Review 99, 5-24.

Dreher, A, S Langlotz, and S Marchesi (2016a), “Information Transmission and Ownership Consolidation in Aid Programs”, CEPR Discussion Paper 11443.

Dreher, A, V Eichenauer, and K Gehring (2016b), “Geopolitics, Aid and Growth: The Impact of UN Security Council Membership on the Effectiveness of Aid”, World Bank Economic Review, forthcoming.

Dreher, A, and S Langlotz (2015), “Aid and growth. New evidence using an excludable instrument”, CEPR Discussion Paper No. 10811.

Dreher, A, A Fuchs, R Hodler, B C Parks, P A Raschky, and M J Tierney (2014), “Aid on Demand: African Leaders and the Geography of China’s Foreign Assistance”, AidData Working Paper 3.

Easterly, W (2015), The Tyranny of Experts: Economists, Dictators, and the Forgotten Rights of the Poor, Basic Books, New York.

Klasen, S (2009), “Ein Plädoyer für die Budgethilfe”, KFW-Development Research.

Marchesi, S, L Sabani, and A Dreher (2011), “Read my lips: the role of information transmission in multilateral reform design”, Journal of International Economics 84: 86-98.

525 Reads