Commonwealth Journal of Local Governance
|
|
Home
| Databases
| WorldLII
| Search
| Feedback
Commonwealth Journal of Local Governance |
|
Philip Amis
University of Birmingham
This commentary reviews contemporary changes in aid modalities and their impact on processes of decentralization. The main change in aid delivery and disbursement considered is towards a greater emphasis on general budget support (GBS) and sector wide approaches (SWAPs). This includes considering the broad questions of firstly, the impact of emphasising GBS on local government and governance systems, and secondly, the extent to which processes of decentralization can fit in with this new approach.
The paper is not the result of a formal research project, but reflects the findings of a wide range of consultancy activities; in particular, work for OECD in 2002 on aid effectiveness in Uganda; work for the UK Department for International Development (DFID) on aid instruments in India and Asia; evaluation of budget support in Uganda and Malawi; and work for UNDP in Rwanda on donor coordination and harmonisation.
The paper is structured as follows. The first section provides the general background and a discussion of new aid modalities. The following sections are concerned with the relationship between new aid modalities and decentralisation in general; a review of the experience of Uganda; and finally a discussion of the question of whether these new aid modalities are strengthening or weakening processes of decentralization and local government and governance.
Since the late 1990s there has been a move amongst many donors to provide budget support as a mechanism to improve the effectiveness of aid. General budget support involves the transfer of un-earmarked donor funds directly to the recipient government’s budget. This is often discussed in contradistinction to traditional project funding and to sector wide approaches.
The broad philosophy behind the move to GBS is an attempt to make the donor-recipient relationship a more mature one than under previous approaches – especially project funding. Thus where both sides agree on broad objectives – a poverty focus within a market framework – the donor need not worry about the detail and can provide the recipient government with a stable source of funding and some flexibility in expenditure.[1] This approach is based upon ideas of partnership rather than a relationship based on patronage and/or charity. A key aspect is to work with and use existing government systems. This approach also requires complementary inputs: dialogue and conditionality, harmonization and alignment, and technical assistance and capacity building.
A final and sometimes unspoken aspect of GBS is that it is a relatively easy way of disbursing aid. It is likely to take a more significant role if the G8 pledges to substantially increase aid to Sub-Saharan Africa are implemented. It also seems inevitable that these new aid modalities will be critical in meeting the Millennium Development Goals (MDGs) to which the international donor community agreed in the late 1990s. We note that a leadership role for local government in meeting the MDGs was one of the themes of the 2007 Commonwealth Local Government Conference held in Auckland, New Zealand.
To a substantial extent the move to sector and budget support was a reaction to the problems of the project-based approaches that had been well documented. These include inter alia: their time bound nature; their tendency to pay high salaries and to attract the best personnel; and most importantly their tendency to ‘honeycomb’ established institutional structures and in many cases to bypass and undermine and weaken existing government systems.[2] The result was that international assistance was systematically weakening the government systems it was supposed to be supporting. This process has been well established and documented across much of Sub-Saharan Africa.
It is worth noting that there are significant differences within the donor community in the extent to which they support GBS. Budget support is widely supported by the UK (DFID), the Netherlands, Scandinavian donors (SIDA, Norway and Danida), and by multilateral agencies, notably the European Commission and the World Bank. The United States (USAID), France and Japan are the main donors opposed to this approach, for a variety of reasons including accountability for funding to Congress (US), other political factors, and simple inertia. German donors (GTZ and KfW) are also currently in the ‘outgroup’ but may be changing their approach. More broadly, there may have been a slight change in attitudes with the 2005 Paris Declaration on Aid Effectiveness, which included a commitment to “increased use of programme-based aid…” (OECD 2005, p6). Precisely how this is implemented depends upon how it is interpreted in detail, but the impression conveyed is that all the major donors effectively signed up to a non-project way of disbursing aid.[3]
The International Development Department (IDD) of the University of Birmingham led an international consortium to carry out a Joint Evaluation of Budget Support between 2004 and 2006. This was a major, and the largest to date, evaluation of the impact of GBS, funded by 20 donors. It involved a rigorous methodology; the development of a ‘causality map’ of the relationship between inputs and outputs, outcomes and possible impacts; plus individual country studies in Burkina Faso, Malawi, Mozambique, Nicaragua, Rwanda, Uganda and Vietnam. These results are published and available on the Internet elsewhere;[4] and in this section we shall highlight some of the major findings.
The literature suggests that the main advantages of general budget support are the following: improved harmonization amongst donors; alignment with partner countries; a reduction in transaction costs; improved efficiency in public expenditure; more predictable funding; more effective state and public funding; and finally improved domestic accountability through increased focus on the government’s own accountability channels.
The Joint Evaluation was to give a positive assessment in five out of the seven countries (the exceptions were Malawi where there was a breakdown in the partnership, and Nicaragua where the process had hardly begun). This way of disbursing aid was found to have important positive and systematic effects, particularly in the field of public financial management in terms of bringing about an increase in discretionary budget funds, improved financial management, and in using government systems and budgets rather than setting up separate systems. There were also some gains in the efficiency and effectiveness of expenditure. These gains were all basically the result of working with and strengthening existing government systems. Capacity building and technical assistance (TA) were important complimentary inputs but were often not well coordinated. It was not really possible to judge the impacts of this approach on poverty reduction given the length of time needed to measure any impact, the problem of attribution, and data constraints. However, it was broadly possible to trace through increases in expenditure in service delivery ministries, in particular in health and education. The results tended to confirm an increase in terms of quantitative coverage of services rather than improvements in quality in service delivery.
In relation to aid delivery GBS was very successful in supporting donor harmonization and alignment, but it should be noted that this was also being promoted by other initiatives aimed at donor harmonization. Finally, the evaluation also suggested the value of a complementary approach in using the different aid modalities, rather than suggesting that one modality was per se superior.
On the negative side the principal observed drawback was for GBS to be unpredictable. This mainly results from the fact that GBS is a very high profile way of disbursing aid with an implicit ‘seal of approval’ of the partner country on the recipient by the donor. The result is that as a way of disbursing aid it is much more vulnerable to domestic political considerations affecting the donor country. For example, in the last few years some donors have removed support from Uganda for ‘governance reasons’, while others (DFID) have limited support to Ethiopia on human rights grounds (repressive response to student demonstrations). The starkest example is the removal of budget support to the Palestinian Authority following the recent Hamas electoral victory. This is a major concern as it undermines one of the key theoretical advantages of GBS, namely the stability of funding arrangements.
The second major area of concern was in relation to the claim that GBS increases domestic accountability: the evaluation found only limited examples of this effect taking place. Thus the idea that GBS would increase the processes through which citizens, non-government organisations and others would hold their respective governments to account were (as yet) not materialising. This has also be confirmed elsewhere (see Renzio, 2006).
There is also some evidence that there is a degree of tension between GBS as an aid modality and the operation of a competitive multi-party democratic system. Thus there are suggestions that in both Indian States and Sri Lanka opposition parties would not honour GBS agreements and that these would have to be renegotiated with a new political regime. The practicality of operating GBS in genuinely democratic political systems is therefore somewhat problematic and has not received sufficient attention. That GBS seems to work best in one-party and/or authoritarian regimes may have some validity; however the driver for this seems to depend more upon a desire for donors to talk/negotiate with technocrats than on other more sinister political motives.[5]
Finally, the Joint Evaluation examined three potential negative effects. Firstly, the suggestion that GBS would lead to an increase in fiduciary risk; secondly, that GBS would result in a substitution effect for local revenue collection; and finally, that it had an inherently pro public sector bias and might inhibit the private sector and/or growth. On all three counts the study did not find adverse effects occurring. In summary, the Joint Evaluation was ‘cautiously optimistic’ about the impact of GBS as a mechanism for disbursing aid.
So it looks very much as if there has been a major change in the mechanism of aid disbursement and that the new approach associated with GBS is here to stay. Furthermore, the processes of scaling up of aid and of some donors (eg DFID) putting limits on administrative costs are likely to significantly increase the importance of GBS as a way of delivering assistance. The remainder of this paper addresses the significance of this change for processes of decentralization and local government and governance. Or to put it in a more vulgar form: how can local government ‘get in’ on this new aid act?[6]
In relation to decentralization the first point to make is that amongst many national government officials and donor economists local government is often treated as if it were invisible. This partly reflects a pre-occupation with national policy and financial systems, but also an implicitly aspatial approach by economists. This blind spot is all the more remarkable given that in most contemporary approaches to poverty reduction primary education and health are given a starring role. In most political systems – including across the Commonwealth – local governments play an important role in delivering these services (Shah and Shan, 2006).[7]
The process of GBS with its emphasis on the critical role of central ministries of finance and the importance of public financial management (PFM) is likely to increase this centralizing tendency. That GBS supports the role of ministries of finance viz-a-viz other ministries was a common finding in the Joint Evaluation. Indeed the emphasis of putting everything ‘on budget’ is a deliberate attempt to try to strengthen the ministries of finance as the sole provider of financial resources.[8]
It is important to understand that this is partly an attempt to undermine both a project and a sector wide (or SWAP) approach. The latter often sets up a situation where interest groups (or policy networks or communities) including both ministry and donor officials are established to promote aid and expenditure in specific sectors. To caricature the situation, it is quite common for a donor official with a particular sector objective (say health or education advisors) to form an alliance with the personnel in the relevant ministry. Their objectives are to get more funds for their specific sector, often in the form of sector wide budget support and/or a SWAP. These interests are often not congruent with macro-economic considerations, general budget support and the thinking of the relevant ministry of finance.
While working in Uganda in 2002 I was able to observe a dispute between the Ministry of Finance and the Ministry of Health around such an issue. The latter was in the process of receiving sector support for health and HIV/AIDs that it had negotiated separately with specific donors. Meanwhile the Ministry of Finance was sought to prevent the transfer of funds arguing that it would jeopardize Uganda’s overall macro-economic stability, exchange rate and inflation targets.[9] This was a clear illustration of the conflict around the mechanisms through which overseas aid could be disbursed in the Ugandan context. The mandate of ministries of finance in most countries, together with the logic of GBS, supports the former taking the lead in managing external assistance. This is entirely consistent with the notion that all external assistance should use existing governmental systems.
The majority of SWAPs and/or sector plans are a vehicle for a multiple donor approach, whereby a range of donors seek to combine their efforts within a given sector. This usually involves designing a sector plan to which a multiplicity of donors are able to contribute either through ‘basket funding’ or with each donor agreeing to fund separate sections of the plan. The development of such a sector approach is thus a very important component of a general donor attempt at harmonization and alignment.
In most cases the sector approaches, depending upon local circumstances, are first developed in the health and education ministries. This seems to reflect both the importance of the two sectors in poverty reduction and also the relative ease of coordination.[10] At its simplest the negotiation is between the relevant donors in a sector plan and the ministry, which is seen as the major institution for service delivery in the sector.[11] The impression gained is that these are the two easiest sectors in which to achieve a joint sector plan and a multiple donor funded SWAP. Nevertheless the amount of time and effort required on all sides to make such arrangements work should not be underestimated – in particular amongst donor officials. This is particularly the case with personnel from the lead donor who must usually not only organise the relevant groups and/or committees, but also persuade new donors to join the partnership and energise the process.[12] All the anecdotal evidence suggests that these are not easy arrangements to establish: they require substantial personnel and resource inputs.
For many donors a commitment to donor harmonisation and such sector approaches is effectively a central directive from their headquarters, as is the case with most north European and Scandinavian donor agencies. Thus there are pressures to extend such arrangements to other sectors. The next sectors are often agriculture, water and sanitation, and law and justice, with local government and decentralization being somewhat of a laggard. The complexity of arrangements seems to depend in part upon the extent to which the core ministry is really the main actor in the sector, as is the case in education and health.
Local government would seem an obvious candidate for a sector approach, not least given its importance in poverty reduction. However for reasons we shall discuss it is somewhat problematic for the design of such programs. In Kenya, Malawi, Ghana, Mozambique and Rwanda, to name those known to the author, there are ongoing attempts to develop sector programs for local government. As we shall discuss Uganda is an exception in having completed the development of a program.
The following reasons seem to explain why it is difficult to design sector programs for local government and decentralization:
▪ The number and diversity of stakeholders
▪ Potential confusion as to who are the key stakeholders: eg the Ministry of Finance, the Ministry of Local Government or the local government bodies themselves
▪ Donors unable to agree about the most appropriate approach to decentralisation[13]
▪ Tension with the Ministry of Finance controlling agenda: who ultimately should control the funds and how they are allocated out to lower tiers of government?
▪ Weak financial and management capacity both at the Ministry of Local Government and in local government bodies themselves: in most countries local government is one of the weaker line ministries
▪ Confusion with other line ministries such as health and education that may have already devolved programs to a local level
▪ A tendency for public sector reform programs not to include local government as a national priority.
There is also a debate about the most effective system of central-local transfers to be used, depending upon specific government objectives (see Shah and Shah, 2006). However, it is clear that it is important to have a system of central-local transfers that is regular, robust and formula driven through which funds can flow easily and effectively.
In summary there is a somewhat worrying possibility that decentralization processes do not easily fit into the new aid architecture. The issues set out above will need to be addressed if GBS and local governments are both to play their respective roles in reducing poverty and achieving the MDGs. This is a complex, awkward, daunting but potentially achievable task.
As part of the Joint Evaluation a special study was commissioned into the relationship between GBS and decentralization in Uganda (see Annex 6 by Jesper Steffensen in Lister et al, 2006). The main conclusions were as follows.
▪ GBS strongly facilitated an increase in funding to local government and related service delivery functions that would not have happened with other aid modalities. The combination of GBS with a Poverty Action Fund (PAF), ring fencing of funds, some SWAPs and inter-governmental fiscal transfers gave both sector ministries and donors sufficient confidence to channel funds to local government and service delivery.
▪ This was supported by capacity building and harmonisation and alignment with Ugandan government procedures and processes.
▪ However, there have been problems with local government autonomy and flexibility in financial management, with concerns being raised about sustainability and local revenue collection.[14]
▪ There was also a tendency to increase upward accountability, often associated with conditions attached to SWAP and PAF funds. It is pertinent to note that these conditionalities are in fact to a large extent a function of those funds being debt relief. Conditions are applied to reassure Northern taxpayers that debt ‘write offs’ are being used for poverty alleviation.
The central question that needs to be posed is: do these new mechanisms of aid disbursement strengthen or weaken local systems of government? Or to put it another way: is decentralization compatible with the new aid modalities, increased funding flows and related commitments to meeting the MDGs? This is both a very important and disquieting question.
One interpretation suggests that increasing funding from the centre strengthens upward accountability to line ministries with a resultant ‘hollowing out’ of local government systems and functions. This was apparent in Uganda in the early 2000s; it made more sense for local government officials to spend time seeking to access funds from their respective line ministries than either collecting local taxation or working to coordinate activities within the local government system. This tendency seems particularly likely in new aid disbursement systems that have a strong sector approach. That the SWAP approach may be harmful to processes of decentralization is a relatively familiar argument and concern amongst reflective practitioners in the field.
An alternative view – at least in theory - is that GBS should strengthen local systems of accountability as the funds are dispersed through government systems. That GBS by strengthening and government systems would support local accountability was not confirmed by the Joint Evaluation. However, two caveats on those findings may be appropriate. Firstly, the new funding arrangements may not have had sufficient time to ‘bed-down’ with politicians, citizens and others learning new roles and responsibilities. Secondly, it is worth noting that other ways of disbursing aid – especially projects – have lines of accountability that are internal to the project rather than to local democratic systems. In reality most forms of dispensing aid are not democratically accountable.
The Ugandan experience and overall success seems to pose a more complicated and fundamental question: Is it fair and/or realistic to expect local government to be sustainable and self-financing in a national economy which is itself highly aid dependent? This question was stimulated by discussions with Danida officials who were concerned that local and district governments were too dependent upon external funding and therefore not sustainable. The discussion would then note that this reliance on external funds could be and was replicated at the national level. In a situation of commitment to MDGs and poverty reduction in an aid-dependent country, increased transfers to local government from the centre can easily undermine decentralization processes.
Is the apparent tension between new aid modalities – GBS – and local government/decentralisation largely one between short run and long run objectives, or does it reflect a more fundamental contradiction? There is no simple answer. It is interesting to compare the neighbouring countries of Kenya and Uganda: in the former the central government is now effectively independent of external assistance, while the latter is highly aid-dependent. The impact of changes in aid disbursement arrangements will clearly have a greater impact in Uganda than Kenya.
A tentative conclusion is that is in aid-dependent countries the new ways of disbursing aid are likely to weaken and/or undermine local government, whilst the same instruments in non aid-dependent countries are likely to strengthen local governments. In making such arrangements work the key variable remains local government capacity, and capacity building remains the principal intervention required in the sector.
One of the advantages of providing GBS to local government is that it not as politically visible as such assistance is to a sovereign government. It is thus somewhat less likely to suffer from the problems of uncertainty and unpredictability associated with the transfer of funds directly to national governments as discussed earlier. Furthermore local governments in general do not have the potentially repressive functions that central governments control. Human rights issues are rarely directly associated with local governments.
It is possible to make the case that such new aid modalities could also be used to directly fund larger municipal governments. Within Sub-Saharan Africa there is no question that the larger urban areas (for example Johannesburg and perhaps Nairobi) have higher GDPs than many countries that have received GBS (for example Malawi). There are of course questions about whether national governments would allow such transfers.[15] Nevertheless it is worth noting that DFID’s relatively successful slum improvement projects in India were effectively using a very similar method by providing funding directly to local governments, albeit with conditions over expenditure and a process of monitoring (Amis, 2001).
Finally, despite the evidence not being clear we are left with two worrying questions. Firstly, can decentralized governance and accountability be integrated into new ways of disbursing aid? And secondly, is the new aid architecture itself in danger of undermining the decentralization process and emerging systems of local government?”
Amis, P (2001) “Rethinking UK Aid: lessons from an impact assessment study of
DFIDs Slum Improvement projects” (2001) in Environment and Urbanization
Vol. 13, No1. pp101-103
Amis, (2002) Uganda: A country case study for OECD Study of Donor Practices 2002-
2003 Birmingham: International Development Department, University of
Birmingham
HMSO, (1997) Eliminating World Poverty: A challenge for the 21st Century A White
Paper on International Development
IDD and Associates (2006) Evaluation of General Budget Support: Synthesis report
Birmingham: International Development Department, University of Birmingham
Lister et al (2006) Joint Evaluation of General Budget Support: Uganda Country
Report Stephen Lister, Wilson Baryabanoha, Jesper Steffensen and Tim
Willamson. Glasgow. DFID.
OECD (2005) Paris Declaration on Aid Effectiveness: Ownership, Harmonisation,
Alignment, Results and Mutual Accountability.
Renzio, Paulo de (2006) Aid, Budgets and Accountability: A survey article
Development Policy Review 24 (6): 627-645
Shah, A and Shah, S (2006) The New Vision of Local Governance and the Evolving
Roles of Local Governments in Shan, A (ed) Local Governance in Developing
Countries World Bank
UNRISD Research and Policy Brief 3 (2004) Technocratic Policy Making and
Democratic Accountability
[1] The philosophy behind GBS is very clearly outlined in the UK ‘New Labour’ government’s first White Paper (1997) on development assistance: Eliminating World Poverty – A challenge for the 21st Century, HMSO, London.
[2] See Amis (2002) for a discussion of this evolution from projects through a sector wide approach to budget support in the Uganda context.
[3] However, differences remained apparent in the Joint Evaluation of Budget Support carried out in 2004-06 (IDD and Associates 2006).
[4] All the GBS studies can be obtained from the DFID website http://www.gov.uk/aboutdfid/performance/evaluation-news.asp or from the OECD/DAC website www.oecd.org/dac/evaluation
[5] See UNRISD Research and Policy Brief 3 (2004) Technocratic Policy Making and Democratic Accountability.
[6] This may seem a crude way of raising the issue but it has often been aired to the author by various interested government and donor officials.
[7] There are some indications that the primary education and health plus market-led economic growth approach to poverty reduction may be beginning to be challenged. It is likely that the future may see a stronger emphasis on the provision of infrastructure for economic growth. The Report of the Commission for Africa 2005 can be read as suggesting this shift.
[8] Providing aid on budget was one of the indicators (indicator 5) highlighted in the 2005 Paris Declaration on Aid Effectiveness; (OECD, 2005, p5 and p9).
[9] There was a heated debate between the two ministries, even using international experts to support their respective positions. Unfortunately I was not there to see the result but am inclined to agree with the IMF country representative who noted that they were both ‘overstating’ their respective cases.
[10] Broadly speaking the donors who are interested in new aid modalities are also those with a keener focus explicitly on poverty reduction. This is perhaps not accidental as it is the impact on poverty that is often used to justify such an approach to the taxpayers of the North.
[11] This comfortable assumption may not be as clear as it seems. In many countries, for example Malawi, the public sector is not the main provider of health and education services. These are primarily provided by non-state actors, in particular faith-based NGOs who account for more than 50 % of the delivery of both education and health services. Nevertheless it can still be argued that the public sector is the main organizing agency for the two sectors.
[12] There is an issue in many donor agencies about how to account for time spent in this way.
[13] It is tempting to suggest that each donor has a tendency to seek to replicate its own country’s system of local government, more so than in other sectors. Thus there are more competing ‘models’ and examples of ‘best practice’ in local government. Perhaps the differences are at a more fundamental level and it is thus harder to get agreement than in, say, education and health. This observation is partly based on discussions with Danida, DFID, GTZ and USAID representatives whilst undertaking consultancy missions.
[14] The Ugandan government has ceased to require local governments from collecting their own revenues. Central local transfers have effectively substituted for this politically unpopular revenue source.
[15] This mirrors the debate in India about whether GBS could be provided directly to State Governments. Both the UK (DFID) and the Netherlands have provided such assistance to the States of Andra Pradesh and Kerala respectively.
AustLII:
Copyright Policy
|
Disclaimers
|
Privacy Policy
|
Feedback
URL: http://www.austlii.edu.au/au/journals/ComJlLocGov/2008/8.html