Too much, too quickly

Development Policy,Peace & Security10 Oct 2011Ruerd Ruben, Antonie de Kemp, Stefan Leiderer

Budget support is a relatively new, yet already widely criticized, phenomenon. However, despite a lack of evidence of its impact on poverty reduction, it is too early to write it off. Budget support is likely to have a significant long-term impact in some partner countries.

Development cooperation has already undergone two paradigm shifts in this new millennium. Initially, donors moved from project aid to budget support, if not in practice, then at least in name. But now, a decade later, support for the instrument is gradually waning, and budget support has become highly contentious. To make matters worse, much of the discussion about budget support is infused with misinformation and misconception.

On the rise

Budget support – which channels funds directly to recipient governments instead of to projects – is not a new instrument, but one that gained momentum in the first decade of this millennium. Several parallel developments contributed to its rising prominence. The first is the empirical assessment of the conditions needed for effective aid. In 1997, Craig Burnside, professor in the department of economics at Duke University in North Carolina, and David Dollar, the US Treasury Department’s economic and financial emissary to China, concluded in their seminal paper, ‘Aid, policies and growth’, that aid has a positive effect on growth in developing countries that pursue ‘sound policies’, meaning sound fiscal, monetary and trade policies.1 Aid would not affect growth when these policies are absent.

This conclusion was reiterated in ‘Assessing aid’, an influential World Bank study published in 2000 and co-authored by Dollar.2 The conclusions – despite severe criticism in academic circles – were instrumental in promoting greater reliance on budget support as they stressed the need for stronger selection criteria. In other words, only ‘good performers’ were to be eligible for budget support. The conclusions also supported demands for stronger country ownership.3 And in a 2003 journal article called ‘Can foreign aid buy growth?’, William Easterly, professor of economics at New York University, showed how ‘aid bureaucrats’ accepted Burnside and Dollar’s ‘fragile’ conclusions and used them to design aid policies in the new millennium.

4Second, general budget support and sector budget support – which channel money to a specific government ministry for use in a specific sector – were also responses to project aid’s perceived lack of efficiency and effectiveness. This was caused by the fragmentation of aid efforts, poor coordination and a lack of ownership. Partner countries were tired of the flood of uncoordinated projects in their countries. They, and the donors, felt that pooling aid funds would result in a more effective approach with a higher likelihood of achieving sustainable results.

Budget support was supposed to generate a number of improvements: more predictable aid, lower transaction costs, stronger government systems, more efficient aid allocation and domestic accountability. In the spirit of the United Nations International Conference on Financing for Development held in Monterrey, Mexico in 2002, which yielded six guidelines for financing development, partner countries reclaimed ownership of their own development by stating their preference for general budget support, not fully realizing that new conditions could actually impair this ownership.

Third, The Millennium Development Goals (MDGs) were also instrumental to the rise of budget support, as they could only be achieved by dovetailing development assistance with government structures in the recipient countries. A more informal, but no less valid argument for the move towards budget support was the limited capacity of embassies in donor countries to manage the growing number of projects.

One instrument, many objectives

On paper, donors agree that decisions on budget support should be unanimous in order to send consistent signals to partner governments. But the reality is different. Donors do not coordinate their approaches when it comes to the objectives and conditions for budget support. Often, it is specific political situations at home that determine budget support’s lifespan, rather than a joint assessment of developments in the partner country.

Budget support was introduced to enhance donor harmonization and coordination, on the one hand, and improve the ownership and accountability of partner countries, on the other hand. Increasingly though, donors have used budget support as a way of encouraging key political and macroeconomic reforms in partner countries, or to promote better governance and democratic principles. Sometimes, budget support has served other purposes. For instance, Ghana and Mali received budget support partly because of their military contributions to peacekeeping forces in Liberia and Sierra Leone, respectively.5

The various objectives of budget support have led to the emergence of different conditions and expectations. Preconditions normally entail more ‘technical’ eligibility criteria, such as well-defined national policies and strategies for poverty reduction, stable macroeconomic frameworks and credible strategies for improving public financial management. Some donors stipulate other criteria, which require receiving countries to adhere to democratic principles and human rights, or fight corruption. These concepts are far from homogeneous and can be interpreted differently depending on the situation.

Bilateral donors such as Germany and the Netherlands have minimum governance requirements and use their contributions like a carrot to uphold minimum standards and extend them. Other donors, such as the European Commission, inject money into the system arguing that only through these actions will recipients’ governance systems improve.

The lack of consensus among donors is also reflected in the different conditions they set for the funds they allocate, and in the overloaded performance assessment frameworks (PAF). For instance, the World Bank assesses government actions, while most bilateral donors focus on outcome variables. Some donors focus on public financial management and governance issues, while others stress the importance of progress on the MDGs. These frameworks include donor countries’ priorities and specialists’ hobby-horses, but do not necessarily reflect ownership. Sometimes donors are deliberately vague about the eligibility criteria, because this gives them more freedom to manoeuvre.6

Not a blank cheque

Imagestate Media Partners Limited – Impact Photos / Alamy

From the start, budget support has had its staunch critics, who argue that it is likely to favour corrupt regimes. These regimes, in turn, would have little incentive to alleviate poverty, since it may encourage political activism hostile to their regimes.

Djankov and colleagues argue in their 2008 article ‘The Curse of Aid’, published in the Journal of Economic Growth, that aid has a negative impact on democracy.  One of their arguments, shared by Paul Collier, author of The Bottom Billion, is that budget support is afflicted by a ‘resource curse’: the funds provide a windfall but may result in rent-seeking behaviour. Moreover, the argument goes, budget support has a negative impact on domestic accountability and reduces incentives to raise taxes. In a similar vein, Dambisa Moyo, author of Dead Aid and How the West Was Lost, advocates a complete cessation of budget support because it only enhances corruption and has a negative impact on economic development.9

Critics do not always shun the use of deceptive tactics in this debate. A notorious example is school grants in Uganda. Several authors have quoted Ritva Reinikka and Jacob Svensson’s well-known tracking survey, ‘Explaining Leakage of Public Funds’, published in 2001. It shows that only 13% of the school grants, provided by the Ugandan Ministry of Education actually made it to the schools.10 Dutch magazine HP De Tijd subsequently concluded in a 2008 article that ‘eighty-seven per cent of all the aid that Uganda received between 1991 and 1995 for basic education disappeared in [someone’s] deep pockets’.

The tracking survey, however, was not analyzing aid, but the contributions of central government to schools. Moreover, these school grants account for only a very small part of the total basic education expenditure, as they do not include wages and funds for school and classroom construction. Nor was the money necessarily stolen. It may have been used for other purposes as well, though a substantial amount of it probably was pilfered.

The extensive focus on budget support in the public debate gives the impression that it has become the main aid instrument in many countries. For instance, Dutch daily newspaper De Volkskrant stated in an article published on 15 May 2011 that ‘in the Netherlands, like in many other rich countries, budget support is the main aid modality’.

Nothing could be further from the truth. Not many countries provide substantial amounts of budget support. In Germany, it accounted for no more than 2.5% of the new bilateral commitments in 2008.11 In the Netherlands, 3.5% of total official development assistance in 2009 was provided as general budget support. In Sweden it was 6%. But the United Kingdom’s budget support contribution constitutes about 20% of its total bilateral aid, and comprises about 25% of the European Commission’s total support.

Another misconception is the idea that budget support is a blank cheque. Countries receiving budget support are always obliged to comply with a number of conditions, including:

  • Developing sound macroeconomic policies, which can be confirmed by a positive International Monetary Fund assessment of overall macroeconomic performance
  • Showing commitment to fighting poverty
  • Implementing public finance management reforms
  • Showing commitment to good governance (including the fight against corruption)

Moreover, donors are extensively involved in the internal affairs of budget support processes. They hold governments accountable for achieving the targets. And these targets, set in the PAFs, are scrutinized in annual reviews. In other words, budget support is anything but a ‘blank cheque’.

However, it is exactly because it is felt that these conditions are often not all met – exacerbated by a number of corruption cases – that budget support seems to be on the wane. Several countries, such as the Netherlands, Sweden, Finland and Germany, have tightened their rules for budget support. Governance issues, including election fraud, caused a number of countries to cancel their budget support to Nicaragua in 2008 and 2009. The Netherlands and Sweden did not disburse to Rwanda in 2009 because the country supported rebels in eastern Congo. The Netherlands suspended budget support to Tanzania that same year because of corruption, and the new Dutch government has further reduced the number of countries receiving budget support. Recently, Sweden cancelled budget support to Zambia because the government was deemed to be lax in its fight against corruption. Germany cancelled budget support to Malawi for the same reason. The United Kingdom and the European Commission, the main providers of general budget support, continue to be its main supporters, although the United Kingdom has become more critical as well.

What the evidence tells us

Given the controversies, more evidence of budget support’s effectiveness would be useful. There are several reasons why evidence is lacking. First, it is a fairly recent type of aid, so there is little data available. Second, the fact that the objectives of budget support are often not well defined and not necessarily coherent makes rigorous evaluation difficult.

This difficulty is confounded by the fact that each individual situation is different.12 And third, choosing an appropriate counterfactual – which examines what the situation would have been without the intervention – is controversial.

But what do the findings tell, even if preliminary?13

Cross-country comparisons that analyse whether budget support works better than other types of aid tell conflicting stories. Some studies suggest that project aid has a more significant impact on economic growth than general budget support.14 Others, however, found the opposite was true, provided that donors and recipients shared the same preferences on budget allocations, and the support was relatively small in relation to the country’s total budget.15

To kick-start the new EU ‘Discussion Papers’ series, Jonathan Beynon and Andra Dusu analyzed the link between budget support and the progress of the Millennium Development Goals (MDGs) in their 2010 paper, Budget Support and MDG Performance. They concluded that countries receiving a relatively large amount of general budget support deliver significantly better results on the MDGs for primary education, gender equity in education, child mortality reduction and access to safe drinking water and sanitation compared to countries receiving less.16

In general, country case studies of budget support focus on the extent to which it creates an ‘enabling environment’ for better government policies, instead of assessing the contribution that has been made towards economic growth or poverty reduction. Admittedly, the case study approach also makes it difficult to attribute economic growth to thebudget support (as there is no explicit counterfactual).

Several studies suggest that budget support has a modestly positive influence on donor harmonization, alignment, ownership and efficient allocation of public expenditure.17 However, in some instances, it has also led to a costly duplication of procedures. For instance, the budget support dialogue does not necessarily replace the dialogue between donors and recipient ministries at the sector level, but it generates new consultation and monitoring structures without necessarily abolishing the old ones.18 In practice, donors use the budget support dialogue structures to discuss specific sector issues. This creates tensions between governments, who seek targets they can easily achieve, and donors, who push for more ambitious goals.19

Critics of budget support have suggested that it risks crowding out domestic taxation. As long as the money comes from abroad, there is no need to tax the citizens (voters) in recipient countries. So far, however, studies have not found evidence for this thesis.20 Geske Dijkstra, associate professor of economics at Erasmus University Rotterdam, and Arturo Grigsby, a Nicaraguan economist, concluded in their 2010 Evaluation of General Budget Support to Nicaragua 2005-2008 that – mainly because of late and unpredictable disbursements – Enrique Bolaños’ administration in Nicaragua mainly used budget support for the repayment of domestic debts, thereby contributing to macro-economic stability.21

One of the most cited impacts of budget support is its positive effect on public financial management and domestic accountability. Budget support gives donors a legitimate argument for insisting on improvements in these areas. There appears to be a broad consensus that budget support has a positive impact on the improvement of public financial management and therefore creates more transparency and domestic accountability.22 However, in general, progress is much slower than anticipated and most results are evident in areas with ‘low-hanging fruit’, such as budget classification systems and reporting.23

So far, country case studies have not conclusively revealed a correlation between budget support and corruption. In fact, corruption indicators have improved slightly in several countries receiving budget support. It appears likely that budget support has helped to reveal cases of fraud and strengthened transparency in government systems. Karel van Kesteren argues in his 2010 book Verloren in Wanorde (Lost in Disorder) that project approaches with a multitude of funding sources present higher fiduciary risks than budget support. 24 Recent corruption cases involving funds from the Global Fund to Fight AIDS, Tuberculosis and Malaria, prove that corruption is not limited to budget support.

In line with Beynon and Dusu’s findings, most case studies conclude that budget support has improved people’s access to services, for instance in health and education, but in most cases this was not accompanied by better service.25 These studies strongly emphasize the need for increasing resources – such as building classrooms and improving health facilities – but often overlook the problems of delivery channels and incentives for front-line service providers.26 As a consequence, rural areas face enormous shortages, and it is difficult to recruit and retain local service providers.

Last, but certainly not least, even while several studies conclude that budget support has effectively contributed to national strategies for poverty reduction, there is still no conclusive evidence that it has an impact on poverty. In general, it appears that expectations were unrealistic given the limited resources and the limited time frame.

Less rhetoric, more pragmatism

The European Commission’s Green Paper on the future of budget support, published in 2010, successfully stirred up the debate on budget support.27 If this debate has made anything clear, it is the lack of consensus on the rationale, objectives and expectations of general budget support. This lack of consensus has important political implications as it undermines the effectiveness of the instrument and complicates the evaluation of the actual success or failure of budget support.

It appears as if budget support is neither a panacea nor a bottomless pit. Evidence shows that while the instrument is unable to bring about huge short-term changes, it is effective in achieving more modest – albeit significant – long-term objectives. One of the problems with budget support is that people expect too much too quickly with insufficient funds.28 Aid is provided to countries that have many core weaknesses, such as non-functioning markets, shortages of skilled labour, an underdeveloped infrastructure and malfunctioning institutions.29 Moreover, in many countries, budget support is of limited importance in relation to total aid, the government’s own budget and the size of the population.

There are also limits to the conditions that can be realistically imposed. Budget support cannot turn autocratic regimes into champions of democracy. Dijkstra and Grigsby conclude that in the case of Nicaragua, donors were overly optimistic to think that they could change deep institutional and political structures in the country. Neglecting underlying political structures poses a danger of creating a virtual reality.30 Nevertheless, this does not mean that budget support cannot be effective in these cases.

Budget support can be effective in achieving modest objectives, provided that donors manage to act collectively. However, experience shows that donors still largely act unilaterally, especially in times of crisis (when corruption cases are revealed or when reforms stagnate). In these instances it becomes clear that domestic considerations are more important than a unified assessment. In 2009 in Zambia, after the EC increased its support to the health sector, Sweden and the Netherlands froze their support because of a corruption scandal. That same year, the Netherlands was the only donor in Tanzania to withheld support, also the result of a corruption case. As a result, recipient countries receive confusing and contradictory signals, which undermines the instrument’s effectiveness.

As early as 2006, the Policy and Operations Evaluation Department of the Netherlands Ministry of Foreign Affairs concluded that ‘the difficulties identified in reaching the poor by giving support to the central government’s sector policy are no justification for falling back on project aid.’ This is no different for general budget support. There is a case to be made for less rhetoric and more realism and pragmatism in finding the right modality mix for enhancing development effectiveness.

Budget support is ineffective when there is no country ownership and when donors and recipient countries do not align their priorities and strategies. However, when these conditions are met, budget support helps to enhance transparency and public finance management, as it gives donors a strong argument for intervening in these areas. Moreover, it creates an opportunity to raise sector issues at a higher level. Budget support can be effective especially in combination with technical assistance, project support and a sector-wide approach. The discussion should not be centred on what is the right modality, but rather on what is the right mix of modalities in specific circumstances.

The authors would like to thank Heidi Tavakoli, research fellow at the Overseas Development Institute’s Centre for Aid and Public Expenditure, for comments on an earlier draft.


    1. Burnside, C. and Dollar, D. (2000) Aid, policies and growth. American Economic Review 90(4): 847-868.
    2. World Bank (I998) Assessing Aid: What Works, What Doesn’t and Why? Washington DC, World Bank.
    3. Koeberle, S., Stavreski, Z. and Walliser, J. (eds) (2006) Budget Support as More Effective Aid? Washington DC, World Bank. For an overview of the debate see: Faust, J. and Leiderer, S. (2008) Zur Effektivität und politischen Ökonomie der Entwicklungszusammenarbeit. Politische Vierteljahresschrift 49(1):129-152.
    4. Easterly, W. (2003) Can foreign aid buy growth? The Journal of Economic Perspectives 17(3): 23-48.
    5. Policy and Operations Evaluation Department of the Netherlands Ministry of Foreign Affairs (IOB) (2008) The Netherlands’ Africa Policy 1998-2006: Evaluation of Bilateral Cooperation. IOB Evaluation 308.
    6. Molenaers, N. Cepiunskas, L. and Jacobs, B. (2010) Budget Support and Policy/Political Dialogue, Donor Practices in Handling (Political) Crises. Antwerp, Institute of Development Policy and Management.
    7. See: Easterly, W. (2003) Can foreign aid buy growth? The Journal of Economic Perspectives 17(3): 23-48.
    8. Djankov, S., Montalvo, J. and Reynal-Querol, M. (2008) The curse of aid. Journal of Economic Growth 13(3): 169–94.
    9. Moyo, D. (2009) Dead Aid: Why Aid Is Not Working and How There is Another Way for Africa. Penguin.
    10. Reinikka, R. and Svensson, J. (2001) Explaining Leakage of Public Funds. World Bank Policy Research Working Paper No. 2709. See also: Reinikka R. and Svensson, J. (2004) The Power of Information: Evidence from a Newspaper Campaign to Reduce Capture. World Bank Policy Research Working Paper No. 3239.
    11. Leiderer, S. (2010) Budget Support as an Aid Instrument: Neither Pandemonium nor Panacea. Bonn, Deutsches Institut für Entwicklungspolitik (DIE) / German Development Institute, Briefing Paper 9.
    12. See also IDD and Associates (2006) Joint Evaluation of General Budget Support. Birmingham, IDD and Associates
    13. A number of evaluation departments of bilateral donors have been working closely together to develop a common methodology for the evaluation of budget support. This initiative was conceived and is chaired by the Evaluation Unit of EuropeAid Co-operation of the European Commission. So far, the evaluation departments have been working on three budget support evaluations: Tunisia (European Commission), Mali (European Commission with Canada) and Zambia (the Netherlands, Germany and Sweden). The EC has published the Tunisia report. The other two evaluations are being finalized. The results are in line with the findings of other evaluations.
    14. See Ouattra, B. and Strobl, E. (2008) Aid, policy and growth: Does aid modality matter? Review of World Economics 144(2) 347-365 and Dalgaard C-J. (2010) Evaluating Aid Effectiveness in the Aggregate: A Critical Assessment of the Evidence. Copenhagen, Danida, Evaluation Study.
    15. Cordella, T. and Dell’Ariccia, G. (2003) Budget Support Versus Project Aid. Washington DC, International Monetary Fund Working Paper WP/03/88.
    16. Beynon, J. and Dusu, A. (2010) Budget Support and MDG Performance. Revised draft.
    17. See IDD and Associates (2006); Koeberle et al. (2006); Lawson, A. Boadi, G., Ghartey, A., Ghartey, A., Killick, T., Kizilbash Agha, Z. and Williamson, T. (2007) Joint Evaluation of Multi-Donor Budget support to Ghana. London, ODI; Independent Evaluation Group – World Bank (2009) Poverty Reduction Support Credits: An Evaluation of World Bank Support. Washington DC, World Bank; Williamson, T. and Dom, C. (2009) Sector Budget Support in Practice: Synthesis Report. London/Oxford, Overseas Development Institute/Mokoro.
    18. Policy and Operations Evaluation Department of the Netherlands Ministry of Foreign Affairs (IOB) (2006) . The Hague, IOB Evaluation 301.From Project Aid Towards Sector Support
    19. Lawson et al. (2007).
    20. IDD and Associaties (2006).
    21. Policy and Operations Evaluation Department of the Netherlands Ministry of Foreign Affairs (IOB) (2010). Evaluation of General Budget Support to Nicaragua 2005-2008. The Hague, IOB Evaluation 329.
    22. See IOB (2010).
    23. See IEG (2009).
    24. Kesteren, K. van (2010) Verloren in Wanorde. KIT Publishers.
    25. See Williamson, T. and Dom, C. (2008) and Independent Evaluation Group – World Bank (2009).
    26. Williamson and Dom (2008).
    27. European Commission (2010) The Future of EU Budget Support to Third Countries. Green paper from the Commission to the Council, the European Parliament, the European Economic and Social Committee and the Committee of the Regions, Brussels, 586 final.
    28. Arndt, C., Jones, S. and Tarp, F. (2009) Aid and Growth: Have We Come Full Circle? University of Copenhagen Discussion Paper No. 09-22.
    29. Riddell, R.C. (2007) Does Foreign Aid Really Work? Oxford University Press, p. 177.
    30. Dijkstra, G. (2005) The PRSP approach and the illusion of improved aid effectiveness: Lessons from Bolivia, Honduras and Nicaragua. Development Policy Review 23(4): 443-464.