Aid allocation decisions of bilateral donors in Ugandan context

ABSTRACT

Aid allocations choices bilateral donors make can be influenced by the institutional environment of a recipient government and the way this environment is perceived. This study examines bilateral donors’ approaches to their aid allocations in Uganda. Uganda provides an interesting case study since the government’s control over the aid funds decreased considerably over the past decade-with donors significantly tightening aid fungibility while the total aid funds Uganda received increased compared with the previous decade, 2002–2009. The donors seemed to control aid fungibility through aid modalities, sectors and/or channels; most aid was carried out as projects, in health-related sectors and through non-state actors, leaving small leverage for the Ugandan government over the aid funds. These might be the donors’ tactical responses to the cloudy political environment of Uganda by increasing their supervision over the aid funds, instead of cutting them. The largest donor, the United States, mostly shaped the aid portfolio of Uganda showing specific preferences in aid modalities, sectors and channels. Uganda should refine its effort towards an improvement of the national political context. This would increase donors’ confidence and willingness to loosen aid fungibility and lead better use of aid resources.

KEYWORDS:

• Aid allocation
• bilateral donor
• fragmentation
• governance
• official development assistance
• Uganda

1. Introduction

The aid allocations choices bilateral donors make are essentially voluntary; they can choose to expand aid to new recipient countries, modify aid portfolios in the current recipient countries or discontinue aid (Acharya, De Lima, and Moore 2006). The voluntary nature of aid allocations enables bilateral donors to adopt their own aid criteria, shaped in specific contexts of and relations to recipient countries including political affinity, economic interests, geographic proximity, cultural values, and historical ties, among others (Dreher, Minasyan, and Nunnenkamp 2015; Rogerson and Steensen 2009). Yet, the collective outcome of their aid allocations likely creates gaps and overlaps in aid distribution, and literature has recognized the issue as a significant problem for effective development assistance (Gehring et al. 2017; Rogerson and Steensen 2009).

The term ‘aid fragmentation’ is often used to describe one of the aspects of uncoordinated aid, which broadly refers to the extent of aid dispersion within or across recipient countries (Acharya, De Lima, and Moore 2006). Since aid fragmentation involves multiple donors, aid channels and aid modalities, the negative connotation of the term is generally associated with high transaction costs. On the other hand, it is argued that fragmented aid can produce positive outcomes. For instance, Gutting and Steinwand (2017) suggest a recipient country with multiple donors can be less affected by an overall aid shock since a donor’s reduction in aid may be cushioned by others. They thus argue aid fragmentation can play a role as meaningful insurance against aid shocks and political instability in recipient countries (Gutting and Steinwand 2017). Also, it may encourage donor competitions and increase negotiating power of recipient governments (Gehring et al. 2017; Johnston 2020). Yet, the overall impact of aid fragmentation probably hinges on characteristic differences in recipient countries and development sectors as well as a sheer number of donors and aid modalities involved. Gehring et al. (2017), for example, provide empirical evidence that a recipient government with high administrative capacity could absorb fragmented aid better than those with low administrative capacity (Gehring et al. 2017). Nonetheless, negative effects identifiable from uncoordinated aid should be reduced for better use of aid resources, and donor coordination is often suggested as a means to do so (Acharya, De Lima, and Moore 2006).

One of the logics behind the approval of donor coordination is that it concentrates aid, which in turn reduces transaction costs (Kimura, Mori, and Sawada 2012). However for bilateral donors, donor coordination is not cost-free in political terms and difficult to achieve due to various donor motivations with foreign aid in the first place. Bourguignon and Platteau (2015) argue intensified donor coordination costs the donors politically because it may press them to give up diplomatic levers on actions by recipient governments (Bourguignon and Platteau 2015). Besides donor coordination, budget support is another widely endorsed means to reduce negative effects of fragmented aid (Furukawa 2020). General budget support refers to donor funds that are disbursed through the financial management system of the recipient government rather than being earmarked for specific uses (Furukawa and Mikami 2014). Although budget support often comes with a set of policy conditions attached, it is considerably less restrictive than project-type assistance (Winters and Martinez 2015). In the early 2000s with an increasing interest in aid effectiveness, there was an expectation that aid portfolios of donors would gradually evolve from project-centered support, sector-based support to general budget support (Hayman 2011). However, the fungible nature of budget support makes it prone to corruption as it can be relatively easily channeled to usages other than donors’ intentions (Winters and Martinez 2015). This issue has alerted donors and studies show that poor governance of a recipient country negatively predicted the likelihood to receive aid in the form of budget support (Clist, Isopi, and Morrissey 2012; Nordtveit 2014). Those studies imply bilateral donors may evaluate their interpretation or their level of concern about the institutional quality of a recipient government when allocating aid funds.

With this backdrop, the objective of this study is to examine how bilateral donors have allocated aid, collectively and individually, to a recipient country with changing institutional environments. In doing so, Uganda is selected for two reasons. First, Uganda is unique in that it has experienced a reversed status as a recipient; in the past, Uganda was the pioneer recipient country for donor coordination, therefore rewarded with increased aid inflows and budget support; yet with its deteriorating political situations, it became the recipient country where the donors tended to tighten aid fungibility with decreased budget support (de la Cuesta et al. 2019). Second, Uganda was ranked among the top ten recipients for the most fragmented aid (Fløgstad and Hagen 2017). For these reasons, the country offers an interesting case to examine details of the aid allocation by its bilateral donors. Findings from this study can provide useful insights for relevant policies forward for the bilateral donors and the Government of Uganda (GoU). Two points should be noted in advance; first, this study does not intend to evaluate aid effectiveness of the donors as it examines the aid allocation profile at a project level; second, this study concerns a donor’s aid relation only to GoU, not to relative significance of Uganda among the recipient countries of the donor. With that, the rest of the study is organized as follows; the next section illustrates how GoU and the donors interacted and influenced one another in the past; the third section briefly describes the data and methods for analysis; the fourth section examines and discusses the aid allocation in the recent decade; and the last section concludes.

2. Past interactions between donors and Uganda under chancing circumstances

Uganda has been reliant on foreign aid for its economic growth (Bwire, Lloyd, and Morrissey 2017; Kasirye and Lakal 2019). Towards the end of 1990s however, GoU realized its heavy reliance on foreign aid became more complicated by uncoordinated donor interventions. In 1997, GoU launched the Poverty Eradication Action Plan (PEAP), an overarching framework for national development plans, priorities, and strategies to guide public investments and organize donor supports (Uganda 2005). With PEAP, Uganda became the first recipient country to direct the donors to operate within the national development framework (Habraken, Schulpen, and Hoebink 2017; Kasirye and Lakal 2019). In 1999, GoU adopted the Sector-wide Approach to further improve aid coordination by aligning sectoral priorities between the government and the donors (Habraken, Schulpen, and Hoebink 2017). This approach was also supposed to address weaknesses of stand-alone projects and capture potential benefits from donors working in a joint enterprise (Riddell and Niño-Zarazúa 2016).

All those early efforts by GoU for aid coordination rewarded Uganda to become the first country that received debt relief under the Heavily Indebted Poor Countries (HIPC) Initiative and the Poverty Reduction Support Credit from the World Bank (Habraken, Schulpen, and Hoebink 2017; Nkusu 2004). This further bolstered GoU to implement PEAP, and in 2000 the Bretton Woods Institutions endorsed additional debt relief for Uganda under the enhanced HIPC Initiative (Nkusu 2004). Yet, some critics argue the debt relief simply allowed GoU to access more loans, instead of pushing the government to reform its fiscal policies (Tebetso 2020).

Uganda’s good performance in economic growth and its strong political commitment to aid coordination encouraged donors to rapidly increase their presence in the country (Bwire, Lloyd, and Morrissey 2017; Habraken, Schulpen, and Hoebink 2017). But this donor proliferation created additional needs for GoU to coordinate and align the donors to the national priorities, and to add more aid principles to PEAP. Simultaneously, Uganda urged the donors to shift their aid composition away from project-centered support but towards budget support so that the aid could be utilized more flexibly based on its development agendas (Bwire, Lloyd, and Morrissey 2017; Uganda 2005). After the 2001 endorsement of general budget support for Uganda, the budgetary aid increased its share in GoU’s total public expenditure especially for poverty reduction (Bwire, Lloyd, and Morrissey 2017; Habraken, Schulpen, and Hoebink 2017).

To further harmonize aid modalities of the donors and reduce transaction costs, the World Bank and other donors including Germany, Netherlands, Norway, Sweden, and UK signed on the Uganda Joint Assistance Strategy 2005–2009 (UJAS) in 2005. Yet the influential donors, US, Japan, and UN agencies did not join UJAS while adhering to their own assistance strategies on Uganda. However, in 2006, one year after the adoption of UJAS and the Paris Declaration, many aid activities in Uganda came to a near standstill due to the deteriorating political environment; the President, Yoweri Museveni changed the Constitution in order to run for the third term with human rights issues raised (Habraken, Schulpen, and Hoebink 2017). The suspension of the aid activities also included the budget support (Nunnenkamp, Rank, and Thiele 2016). In fact, even before the 2006 incident, UK reduced its budget support by small amounts in 2004 and 2005 and Ireland and Norway reduced their budget support by 25% in 2005 because the donor community declared the existence of large-scale corruption in GoU at the 2003 Consultative Group Meeting (Tangri and Mwenda 2006). But the 2006 incident might be too visible and loud for the donors to disregard. Even so, most budget support donors did not choose to cut the aid drastically, rather redirected their aid to sectoral supports, projects, or humanitarian assistance in the northern area of Uganda (Nunnenkamp, Rank, and Thiele 2016). This was a more cautious approach of the donors to balance between continuing development assistance and avoiding reputational risks i.e. assisting the allegedly corrupt government directly (Habraken, Schulpen, and Hoebink 2017).

Yet, some donors such as Belgium, EU, UK, and US took little action under the changing political environment because they considered Uganda strategically important for its leading military role in the African Union Peacekeeping Force in Somalia and for trade in Central Africa (DFID 2014; Habraken, Schulpen, and Hoebink 2017). Despite the disparities in the donor actions, pressure from the donor community seemed effective on GoU. The Human Rights Watch, for example, claimed GoU held a referendum on multiparty politics as a result of the threat from the donors to withhold aid although President Musevini made an irate statement that he would not yield to pressure from the donors (Hayman 2011).

When PEAP was expired, GoU introduced the National Development Plan (NDP) 2010-2015, debatably with more domestic approaches and less donor influence (Habraken, Schulpen, and Hoebink 2017; Kasirye and Lakal 2019). Critics on NDP, however, argue it lacked government coordination, a framework for the division of donor labor, and realistic plans for the priorities. At the same time, the introduction of NDP pressed the UJAS-signatory donors to reconsider UJAS, which was directly connected to the expired PEAP. In the end, the UJAS-signatory donors did not extend it and returned to their country assistance strategies to individually align aid activities to NDP (Habraken, Schulpen, and Hoebink 2017).

The recent development process of Uganda has been driven under the Vision 2040, approved in 2007 and launched in 2013 (Kasirye and Lakal 2019). The Vision 2040 reflects part of GoU’s movement towards large-scale and state-initiated development while instrumentalizing a 5-year NDP to achieve its goal (Kasirye and Lakal 2019; Nayler 2021). Thus, donors are expected to organize their support around NDP for reduced aid duplication and increased complementarity. Three institutions of GoU are mainly in charge of leading the NDP implementation: the Ministry of Finance, Planning and Economic Development (MoFPED), the National Planning Authority (NPA), and the Office of Prime Minister. Before the establishment of NPA in 2002 as a statutory planning agency, MoFPED oversaw the development planning and the budget (Bwire, Lloyd, and Morrissey 2017). But the institutional rearrangement, or the planning authority transferred to NPA, caused complexities for initial cooperation between MoFPED and NPA to execute the first NDP. The outcome of the institutional power struggle turned out a lack of linkage between the priority areas of the first NDP and the annual budget allocated to them. The second NDP 2016–2020 is known to lean more towards infrastructure development, arguably crowding out social expenditures in relative terms. The logic behind this shift is that development of productive infrastructure including transport and energy is a key to meet the objective of the Vision 2040 whilst many donors insist that progress cannot be made without a socially balanced development (Kasirye and Lakal 2019).

In sum, the case of Uganda pointed that the aid practices of the donors were at least partly influenced by the donor perceptions about and the donor interactions with the political environments of the recipient government. The case also described the divergent donor behaviors prevented the donors from speaking with one voice for heavier pressure on GoU to improve its institutional quality. From the GoU’s standpoint, different donor behaviors may work for its advantage in some cases since the donors unlikely reach a full consensus to hold GoU accountable (Kasirye and Lakal 2019). After a brief description of the data sources and methods below, the following section explores in detail how the donors have allocated aid funds during the recent decade in Uganda.

3. Materials and methods

This study examined the aid allocation profile of the bilateral members of the Organisation for Economic Co-operation and Development (OECD) Development Assistance Committee (DAC) between 2010 and 2019 in Uganda. The aid data for the earlier years were often incomplete or missing at the project level, but the available data between 2002 and 2009 were utilized as reference points when needed. All disbursement records were extracted from the OECD statistics (OECD Statistics). The reason for utilizing disbursements, instead of commitments was that the current study was more interested in examining the actual outcomes of donor decisions on aid and the interactions with GoU i.e. disbursement than the indications of donor aid policies i.e. commitment (Nordtveit 2014). The basic unit of the analysis was an aid event or each annual aid disbursement, defined as such in the previous literature (Acharya, De Lima, and Moore 2006). For simplicity, the aid event is called an aid hereafter.

The analysis excluded aids coded as 900 in the Creditor Reporting System (CRS), following the logic in Kimura, Mori, and Sawada (2012); the 900 category specifies administrative costs of donors, spending on refugees in donor countries and promotion of development awareness, which are clearly not related to aid allocations within the recipient country. For sectoral examination, the aids were sorted by the three-digit codes ending with zero, which classify the most general sectors in the five-digit CRS purpose code system. For aid dispersion, the Herfindahl–Hirschman Index (HHI) was measured by squaring the share of specific (be it a sector or a donor) aid over the total disbursement (Kimura, Mori, and Sawada 2012). Studies on aid allocations frequently measure HHI as a proxy indicator to quantify the extent of aid dispersion; HHI closer to 0 signals higher aid dispersion or lower aid concentration while HHI closer to 1 signals lower aid dispersion or higher aid concentration (Furukawa 2020; Gutting and Steinwand 2017). The Theil index is also used for similar purposes, which is to measure relative inequality (Fløgstad and Hagen 2017). Regarding utilization of these two indices, there have been discussions on whether one is more suited than the other to measure aid fragmentation. Yet, some studies indicate HHI and Theil are highly correlated showing almost identical results (Fløgstad and Hagen 2017; Oh and Kim 2015). The aids were also grouped by the aid delivery channels and the aid types according to the OECD criteria. Each aid delivery channel is labeled with a numeric channel identification, ranging from 10000 (through Public Sector Institutions), 20,000 (Non-Governmental Organisations and Civil Society) to 90,000 (Other). Different aid types are coded in capital alphabets, from A for Budget Support, B for Core Contributions and Pooled Programmes and Funds, C for Project-type Interventions to H for Other in-Donor Expenditures.

Finally, this study divided the bilateral donors into two groups based on their contribution size: significant donors, and insignificant donors. The criterion for the division was partly referred to Gehring et al. (2017); DAC designates a group of donors as significant when they cumulatively account for over 90% of total aid inflows in a recipient country. With this, Gehring et al. (2017) denote the remaining donors or the donors cumulatively accounting for less than 10% of total aid inflows as not significant or small. It should be noted that in the following results section, some numeric results from the allocation analysis were not shown in the tables for brevity, but they were explained in the relevant contexts for discussions.

4. Results

This section examines how the bilateral donors allocated aid funds, collectively and individually, under the Uganda-specific circumstances during the recent decade. The section starts with overall findings from the allocation analysis, then moves to donor-specific findings, focusing on the significant donors. During the study period, Uganda received total UDS 10.8 billion with few drastic annual swings (Table 1). For the annual average of disbursement inflows, it increased in comparison to that of 2002–2009: USD 0.8 billion for 2002–2009 vs. USD 1.08 billion for 2010–2019, both measured in constant 2019 USD. Total 29 DAC bilateral donors were involved in supporting Uganda, nine of which cumulatively accounted for over 90% of the total disbursement (Table 2). Thus those nine donors were designated as the significant donors; they were US, UK, Japan, Germany, Denmark, Norway, Sweden, Ireland, and Netherlands in the order of the disbursement size. The remaining 20 donors jointly accounted for less than 10% of the total disbursement, thus designated as the insignificant donors.

Table 1. Annual aid disbursements and aid delivery channels of donors.

Year	Total^a	Channel^b	No of aid events	Total^a	Average per aid event	% Total
2010	1011.4	0	10	2.8	0.3	0.0^c
2011	965.7	10000	5370	4329.9	0.8	40.2
2012	915.7	20000	8361	2602.1	0.3	24.2
2013	1034.9	30000	83	19.5	0.2	0.2
2014	1062.8	40000	936	1578.8	1.7	14.7
2015	997.7	50000	1536	560.5	0.4	5.2
2016	1114.6	60000	974	760.7	0.8	7.1
2017	1303.9	90000	1810	917.6	0.5	8.5
2018	1222.7
2019	1142.4
Average	1077.2

^a Measured in constant 2019 USD millions.

^b 0: not reported, 10000: public sector institutions, 20000: NGO and civil society, 30000: public-private partnerships and networks, 40000: multilateral organizations, 50000: teaching/research institutions or think-tank, 60000: private sector institutions, 90000: other.

^c Due to rounding-off.

Table 2. Aid disbursements of donors and their utilization of public institutions channel for aid delivery.

Donor	Total aid amount^b	% of total amount	Public channel Uganda^c	Avr. public channel All recipients^d	Donor	Total aid amount	% of total amount	Public channel Uganda	Avr. public channel All recipients
US^a	5061.6	47.0	33.9	37.6	Iceland	44.3	0.4	90.3	54.9
UK	1613.1	15.0	24.5	28.7	Finland	44.1	0.4	0.6	31.2
Japan	644.4	6.0	80.7	84.5	Australia	38.8	0.4	9.3	30.0
Germany	526.2	4.9	72.2	68.1	Switzerland	13.3	0.1	2.8	39.4
Denmark	470.7	4.4	64.1	44.6	Spain	10.8	0.1	1.8	39.2
Norway	459.4	4.3	42.5	32.3	Hungary	9.4	0.1	1.7	55.9
Sweden	415.7	3.9	15.7	38.3	Luxembourg	2.5	0.0^e	0.0	0.0
Ireland	320.4	3.0	34.7	31.2	Poland	1.3	0.0	15.6	61.7
Netherlands	242.5	2.3	38.2	37.7	Greece	0.6	0.0	12.5	73.5
Belgium	216.6	2.0	64.4	47.7	Slovenia	0.6	0.0	0.0	0.0
France	184.5	1.7	93.8	86.1	Rep. Czech	0.5	0.0	0.0	0.0
Korea	162.5	1.5	62.8	84.3	New Zealand	0.4	0.0	3.8	54.3
Austria	127.1	1.2	55.7	50.5	Rep. Slovak	0.2	0.0	0.0	0.0
Canada	102.0	0.9	2.2	35.1	Portugal	0.1	0.0	0.0	0.0
Italy	58.0	0.5	37.9	73.8

Source: OECD statistics & https://doi.org/10.1787/2dcf1367-en.

^a Significant donors in bold.

^b Measured in 2019 constant USD million.

^c Percentage of public institutions channel utilization of donor in Uganda.

^d Average % of public institutions channel utilization across all recipients of donor.

^e Due to rounding-off.

The bilateral donors utilized various aid delivery channels (Table 1). Approximately, 40% of the total disbursement was channeled through the public institutions whereas 60% flowed through the non-state channels. Yet, there was a difference in channel utilization for aid delivery across the study years; in 2010 about half of the total disbursement was delivered through the Ugandan public institutions, but the share was reduced considerably to 24% in 2019. For the noticeable reduction, Dietrich (2013) may offer some relatable insights; the study argues the bilateral donor decision about selection of the aid delivery channel is endogenous to the quality of state institutions in a recipient country because the state institutions could serve as a credible signal for gauging the probability of successful aid implementation (Dietrich 2013). Although one proxy indicator is seldom able to offer sufficient information on institutional quality, a change in the index for corruption control might be relevant to the shift in the donors’ channel utilization with their aid funds. The corruption index on GoU worsened from 2010 (bottom 19 percentile rank) to 2019 (bottom 12 percentile rank) while the donors’ utilization of the public institutions channel decreased (World Bank databank). Bermeo (2010) also provides relatable evidence on a relation between donors’ channel choices and governance of recipient countries; low governance quality of a recipient country positively predicted more aid being delivered through non-state channels (Bermeo 2010). Probably, bypassing the state institutions implies donors’ concerns about their aid becoming ineffective or wasted by weak state institutions. However, any empirical relation between the governance quality of GoU and the donors’ channel choices should be established by further research.

At an individual donor level, the bilateral donors showed different preferences for aid delivery channels (Table 2). Among the significant donors, Japan and Germany disbursed aid funds mainly through the public institutions, or 81% and 72% of their total disbursement, respectively. In contrast, US and UK used the public institutions channel with 34% and 25% of their total disbursement, each. The majority of other donors also tended to prefer less the public institutions than the non-state channels. Furthermore, many donors seemed to utilize the Ugandan public institutions channel less, compared with their average utilization of the public institutions channel across their recipient countries (Table 2). This comparison also hinted some donors might have a straightforward preference for a specific aid delivery channel. For instance, regardless of their channel preference in Uganda, US, UK, Japan, Germany, Ireland, and Netherlands showed similarities in their channel preference between Uganda and their overall recipient countries. Others such as Denmark, Sweden, and Norway showed distinct differences. Simultaneously, their choice of the aid delivery channel could be influenced by the nature of their preferred aid modalities. Some aid types can be more effectively delivered through certain aid channels. For instance, a project-type intervention that requires specific local knowledge may stand a higher chance to succeed by an experienced non-governmental organization (NGO) specialized in that region.

Among the non-state channels, NGOs are of particular importance to the bilateral donors (Dietrich 2013). They provide the donors with an alternative or complementary aid channel due to their issue-focused specialization and expertise for local situations. In Uganda, NGOs have been thriving with over 3800 NGOs in operation (Johnston 2020). Early in 1989, the increasing presence of NGOs pushed GoU to enact the NGO Registration Law and establish the National NGO Board in order to register all NGOs. Again in 2016, GoU enacted the NGO Act 2016 and established the National Bureau for NGOs, a semi-autonomous body under the Ministry of Internal Affairs. The 2016 Act mandates the Bureau to not only register NGOs, but also to regulate, coordinate, inspect, monitor, and oversee all operations by NGOs in Uganda (NBNGO website). This institutional development solely for NGOs may echo substantial roles NGOs have been playing in the country.

Among the aid types, the most preferred type by the donors was the project-type interventions (Table 3). Chenery and Strout (1966) mention the project system to transfer assistance has the virtue of simplicity when the aid transfer system should include a basis for determining the amount of transfer, and the specification of the form of resources to be furnished and a basis for controlling their use (Chenery and Strout 1966). In line with controlling aid use, the project-type intervention more likely guarantees a higher level of donor autonomy over resource management and implementation terms because the donors can control project details including monitoring schemes (Kasirye and Lakal 2019). The project-type interventions might also help the donors justify aid spending for their public since the outcome of the project is generally more visible and less complicated for such justification than other types.

Table 3. Details of aid types and sectors in project-type.

Aid type		No of aid events	Amount^a	% of total amount	Sectors in C type^c	Amount	% in C type amount
A^b	A01	10	158.5	1.5	130	3274.2	40.5
	A02	57	234.0	2.2
	A total	67	398.5	3.7
B		3066	1803.2	16.7	120	893.5	11.0
C		11819	8091.2	75.1	310	634.7	7.8
D		3403	421	3.9	150	546.3	6.8
E		703	57.1	0.5	230	448.3	5.5
F		7	0.4	0	720	394.5	4.9
G		15	0.4	0	110	371.4	4.6
Total		19080	10771.8	100 %

Of all 22 sectors covered with C type, top sectors based on disbursement were shown.

Source: OECD statistics.

^a Measured in 2019 constant USD million.

^b A: budget support (A01: general budget support, A02: sectoral budget support), B: core contributions and pooled programmes and funds, C: project-type interventions, D: experts and other technical assistance, E: scholarships and student costs in donor countries, F: debt relief, G: administrative costs not included elsewhere.

^c 110: education, 120: health, 130: population policies/programmes & reproductive health, 150: government & civil society, 230: energy, 310: agriculture, forestry, fishing, 720: emergency response.

Dreher, Minasyan, and Nunnenkamp (2015) argue donors prefer donor-managed projects when they have little confidence in capacity of the recipient government to use aid constructively. In a word, a lower level of donor trust may be associated with a higher share of project-specific support (Dreher, Minasyan, and Nunnenkamp 2015). Furthermore, Winters and Martinez (2015) indicate concentration of specific kinds of projects depends on governance quality of the recipient country; for instance, donors tend to consider large infrastructure projects more susceptible to corruption than short-term small projects that are tied to predictable outcomes (Winters and Martinez 2015). In Uganda, the majority or over 50% of the project disbursement was allocated to two health-relevant sectors, the 120 (health) and the 130 (population policies/programmes & reproductive health) whereas infrastructure types such as transport and energy were comparatively minor (Table 3). However, when the donor projects are not effectively coordinated, the uncoordinated projects may backfire with project overlaps and negatively affect administrative capacity of the recipient government. Ugandan officials note that the project-based interventions often ran parallel to the ones by the local governments, and the interventions undermined the local governments by frequently absorbing local administrative resources and crowding out their own initiatives (Kasirye and Lakal 2019).

Following the project-type, the core/pooled funds type was the next choice of the donors (Table 3). With the core/pooled funds type, the most supported sector was the 720 (emergency response), which received 52% of the total disbursement of the type. This sectoral choice seemed somewhat predictable because the emergency responses need collective efforts for larger and swifter impacts. However, the majority or 60% of the total amount of the core/pooled funds type went to the B03 (sub-type of B: contributions to specific-purpose programs and funds), over which the bilateral donors still held some control for directing aid funds to the programs and funds based on their strategic preferences.

For the budget support type, which had been much discussed and strongly endorsed by the donors in Uganda in the earlier years, it accounted for less than 4% of the total disbursement during the recent decade (Table 3). This was a drastic decrease relative to the level of 2002–2009, during which GoU received over 10% of the total aid as budget support. Among the two types of budget support-general and sectoral, Germany, Norway, and UK included the general budget support in their aid profile and UK was the largest provider accounting for over 80% of the total general budget support. Studies state the general budget support is more politically motivated and an important tool for the formal colonial powers to maintain political influence (Dreher, Nunnenkamp, and Thiele 2008; Dreher, Minasyan, and Nunnenkamp 2015; Nordtveit 2014). Nordtveit (2014) in particular provides relatable evidence that a colonial link between the donor and the recipient showed a strong prediction for receiving the general budget support and for the volume of it (Nordtveit 2014). Given the formal colonial connection between UK and Uganda, those studies might offer some ground for UK’s strong stance for general budget support to GoU. Nonetheless, the general budget support completely ceased after 2012; UK suspended it owing to the 2012 corruption scandal in the Office of Prime Minister and Germany also terminated it after 2012 (DFID a). The sectoral budget support, on the other hand, continued to flow in during the study period from ten donors. But the size of the sectoral budget support decreased substantially from USD 69 million in 2010 to USD 3 million in 2019. Both the termination of general budget support and the considerable decline in sectoral budget support signal the budget support donors became more risk-averse with their aid funds, considering the situations with the Ugandan governance.

For the sectoral aid, the donors supported total 23 sectors, classified by the most general three-digit CRS codes ending with zero (Table 4). Of the 23 sectors, the 130 received the largest disbursement, distantly followed by the 120. Despite 23 donors involved, the 130 showed the largest HHI among the sectors, indicating high aid concentration. The reason for the sector’s high concentration was that over 90% of the sectoral total disbursement came from just one donor, US. The donor prioritized the 130 by investing about 60% of its total aid in the sector and no other significant donors showed such high sectoral concentration. Since US was by far the largest donor, its sectoral priority also shaped the overall sectoral aid profile in Uganda.

Table 4. Details of donor sectoral aid.

Sector^a	Total amount^b	% total amount	No of aid events	Average per aid event	No of donors	HHI^c
110	614.8	5.7	2571	0.2	27	0.1212
120	1133.4	10.5	2362	0.5	24	0.3304
130	3435.3	31.9	1766	1.9	23	0.8141
140	494.8	4.6	757	0.7	21	0.1763
150	999.1	9.3	3078	0.3	22	0.1114
160	350.2	3.3	938	0.4	23	0.2808
210	288.2	2.7	247	1.2	8	0.4728
220	18.0	0.2	117	0.2	16	0.2233
230	463.6	4.3	425	1.1	14	0.1870
240	121.5	1.1	327	0.4	17	0.3514
250	33.5	0.3	210	0.2	17	0.2105
310	819.4	7.6	2377	0.3	23	0.2219
320	115.0	1.1	547	0.2	20	0.1907
330	104.2	1.0	241	0.4	15	0.4081
410	139.8	1.3	553	0.3	17	0.3119
430	330.2	3.1	1457	0.2	26	0.1093
510	158.5	1.5	10	15.9	3	–^d
520	193.1	1.8	225	0.9	11	0.3980
530	0.03	0.0^e	1	0.0	1	–
600	0.4	0.0	7	0.1	1	–
720	930.1	8.6	757	1.2	23	0.2774
730	8.9	0.1	40	0.2	10	0.3135
740	19.8	0.2	67	0.3	15	0.1987

Source: OECD Statistics.

^a 110: education, 120: health, 130: population policies/programmes & reproductive health, 140: water supply & sanitation, 150: government & civil society, 160: other social infrastructure & services, 210: transport & storage, 220: communications, 230: energy, 240: banking & financial services, 250: business & other services, 310: agriculture, forestry, fishing, 320: industry, mining, construction, 330: trade policies & regulations, 410: general environment protection, 430: other multisector, 510: general budget support, 520: development food assistance, 530: other commodity assistance, 600: action relating to debt, 720: emergency response, 730: reconstruction relief & rehabilitation, 740: disaster prevention & preparedness.

^b Measured in 2019 constant USD.

^c HHI calculated by number of donors in sector.

^d Sectors with fewer than five donors were excluded for HHI due to excessively high HHI.

^e Due to rounding-off.

Regarding the number of donors involved, the sector 110 (education) attracted the highest number of donors or 27 without dominating donors (Table 4). This resulted in the very small HHI, signaling high fragmentation in the 110. The education sector generally observes a larger number of donors than other sectors, thus it can be more prone to aid fragmentation (Gehring et al. 2017). Indeed, Frot and Javier (2010) indicate the education sector was the most fragmented sector in their study (Frot and Javier 2010). The Ugandan case seemed congruent with the general characterization of the education sector. Similar to the education, the 120 (health) is considered to draw a larger number of donors. This may be due to the characteristic universality/neutrality of health and education for development, therefore unlikely to create contradictions to donor aid policies and diplomatic relations. In the health sector of Uganda, US also dominated, providing over half of the sectoral total disbursement, followed by UK accounting for 21%. Even though the 120 involved 24 donors, the strong presence of US and UK made its HHI (0.33) above the median (0.28).

Following the 120, the sector 150 (government & civil society) received the third largest disbursement and it was one of the most fragmented sectors (Table 4). With 22 donors supporting the 150, there appeared to be no dominating donor; instead, the top donors of the sector contributed similarly, ranging from 19% of the sectoral total disbursement by UK, 14% by Sweden, 12% by US to 11% by each Denmark and Norway. Recently, Ridell (2021) suggests in a broad context there was little evidence that foreign aid including democracy-building aid had a meaningful impact in Uganda on stopping democratic backslides (Ridell 2021). While Ridell’s study may serve as a good starting point, further research would be required for aid effectiveness related to the 150.

A notable point in the sectoral aid allocation may be that the sector 310 (agriculture, forestry, and fishing) received a rather small share or 8% of the total disbursement. It was despite the fact that Uganda essentially depends on agriculture for its overall economy and agriculture is the single most important income source by supplying jobs to 80% of the Ugandan workforce (GIZ website). Three donors, US, Denmark, and Netherlands in that order, provided 70% of the total 310 disbursement. Reasons for the lack of stronger support to the 310 are unclear. At least, the agricultural support through the 43040 (other multisector: rural development) can be excluded because the rural development accounted for only 0.7% of the total disbursement. Perhaps, the donors might view the 310 with sufficient involvement of multilateral organizations, and/or non-DAC donors such as China. In the Sub-Saharan regional context, Ssozi, Asongu, and Amavilah (2017) indicate more agricultural aid went where government institutions were increasingly effective (Ssozi, Asongu, and Amavilah 2017). However, evidential links between the donors’ agricultural support and the institutional quality of GoU requires separate research. The following section focuses more on individual donors.

Among the significant donors, the contribution to the total disbursement varied widely from 47% by US to 2% by Netherlands (Table 2). And the number of the sectors supported by these donors differed, not necessarily proportional to their disbursement size (Table 5). As previously mentioned, the largest donor, US prioritized two health-relevant sectors; the 120 and the 130 together accounted for over 70% of US’ total disbursement. This prioritization was also explicit on its aid documents on Uganda (US Foreign Assistance website). Additionally, not only did the donor’s focus on those two sectors determine the aid profile of Uganda, but the focus made US assistance the least fragmented with the largest HHI among the significant donors (Table 5).

Table 5. Details of individual donor aid.

Donor^a	HHI^b	No. of sectors	Average per aid	No. of agencies^c	Donor	HHI	No. of sectors	Average per aid	No. of agencies
United States	0.4004	18	1.09	14	Iceland	0.4184	11	0.62	2
United Kingdom	0.1011	21	1.14	10	Finland	0.1701	17	0.10	3
Japan	0.1493	19	0.41	5	Australia	0.3255	16	0.11	1
Germany	0.1173	21	0.41	20	Switzerland	0.2088	13	0.09	3
Denmark	0.2014	18	0.76	2	Spain	0.2865	14	0.08	11
Norway	0.1614	20	0.33	5	Hungary	0.5563	5	0.52	5
Sweden	0.1710	18	0.38	8	Luxembourg	0.2814	7	0.06	1
Ireland	0.1581	18	0.25	2	Poland	0.3059	8	0.04	3
Netherlands	0.3156	12	1.12	1	Greece	0.7124	2	0.04	2
Belgium	0.1952	16	0.46	5	Slovenia	0.2348	5	0.03	1
France	0.4312	16	0.72	9	Rep. Czech	0.6793	5	0.06	2
Korea	0.1479	19	0.17	4	New Zealand	0.9266	2	0.04	1
Austria	0.2686	17	0.17	12	Rep. Slovak	0.5268	4	0.01	1
Canada	0.2546	19	0.09	5	Portugal	1.0000	1	0.04	1
Italy	0.3111	15	0.10	7

^a From largest donors based on total disbursement.

^b Donor HHI calculated by number of sectors supported.

^c Excluding miscellaneous.

Upon the independence of Uganda in 1962, US began its development assistance for the country. Yet in the wake of the military coup in 1971 and the subsequent internal unrests, US suspended its assistance until 1981 and resumed it afterwards (USAID website). The relative long-term engagement of US with Uganda and its current status as the largest donor seem to fit US in the lead donor position, defined by Steinwand (2015). Steinwand (2015) argues that lead donorship requires a long-standing tie between a donor country and a recipient country, in which the donor continues to be the biggest aid provider. And lead donorship is exclusive in that few other donors vie for the role as the top aid provider to that recipient country. Furthermore, Steinwand (2015) claims the presence of a lead donor helps coordinate behaviors of other donors and reduce donor fragmentation (Steinwand 2015). In the Ugandan context, however, the presence of US did not particularly appear to help coordinate behaviors of other donors. As discussed in the previous section, US maintains its own assistance strategy rather than actively joining the aid coordination bodies with other donors. Also, its lack of enthusiasm for donor coordination could be observed with its utilization of the core/pooled funds type; US disbursed 8% of its total disbursement to the core/pooled funds whereas the collective share of this aid type of all donors was 17%. Moreover, all the 8% by US was disbursed to the B03 (contributions to specific-purpose funds) where US could exercise a degree of control over its funds. Even so, the robust presence of US might influence aid practices of other donors in a way that the aid allocation decision by US nudges other donors to avoid the sectors US prioritizes. The UK document, for instance, states ‘UK will further reduce bilateral support earmarked to HIV/AIDS in 2015/16, an issue that receives significant funding from other donors’ (DFID 2014, 6). Although the UK document does not specify who other donors are, US has long assisted Uganda in HIV/AIDS eradication.

The current study aggregated the disbursement data at the donor country level. But this aggregation could potentially underestimate a donor’s contribution to the administrative complexities for GoU if an excessive number of donor agencies participated in aid. Acharya, De Lima, and Moore (2006) argue an immediate consequence of proliferated donor agencies is a significant increase in transaction costs for recipient governments through their engagement with donor agencies (Acharya, De Lima, and Moore 2006). From the GoU’s standpoint, each aid agency of the donor country becomes a potential negotiating partner although it depends on a level of autonomy of that donor agency in aid decision-making. Thus, the number of participating donor agencies may turn out to be an additional source for GoU’s aid management complexities.

The number of participating donor agencies varied considerably from 1 to 20 agencies (Table 5). Yet most donors with the multiple agencies had a representative that implemented a large share of disbursement, such as the development agency or the Ministry of Foreign Affairs. US disbursed aid funds through 14 agencies with the US Agency for International Development (USAID) being the primary conduit. The agency executed over 65% of total US disbursement, followed by the Department of Health and Human Services with 28%. Among other agencies, the Millennium Challenge Corporation (MCC) is worth mentioning; MCC is specifically charged to assist recipient governments that are deemed to meet sufficient quality of good governance (Winters and Martinez 2015). In a sense, the very little fund disbursed via MCC or 0.04% of total US disbursement may offer a glimpse of how US viewed the status of Ugandan governance.

The second largest donor, UK was different from US in that it did not prioritize a particular sector. Instead, UK spread its aid funds across 21 sectors with some emphasis on the 720 (17% of total UK disbursement), the 120 (15%), the 150 (12%), and the 160 (11%: other social infrastructure & services). This rendered UK’s HHI the lowest among the significant donors i.e. the least concentrated aid sector-wise. According to the UK document, its relative emphasis on the 720 was linked to that GoU hosted millions of refugees from DR Congo and South Sudan (DFID 2014). To support the 720, UK channeled its aid funds exclusively through a multilateral organization; the World Food Programme received 98% of UK’s total 720 disbursement. While the donor terminated the budget support with the 2012 massive corruption scandal, UK did not seem to reduce its aid to GoU by stating that UK would re-allocate its aid funds to other aid modalities (DFID 2014). The historic/colonial relation between the two countries probably kept UK maintaining a certain volume of aid and spreading its aid across many sectors, which might help expose its presence to the wider Ugandan public. Related to spreading aid, Winters and Martinez (2015) suggest evidence that donors likely diversify their aid flows when providing aid to former colonies (Winters and Martinez 2015).

Other two significant donors, Japan and Germany focused more on infrastructure sectors; the top sectors of Japan were the 210 (transport & storage) and the 230 (energy) and those of Germany were the 140 (water supply & sanitation) and the 230. The Japanese aid document states the donor emphasizes economic infrastructure, energy, agriculture, and health/water to assist development of Uganda (JICA 2020). And the German aid document indicates Germany has the three priority areas agreed with GoU: renewable energies and energy efficiency, rural development and food security, and water and sanitation (GIZ website). Of the insignificant donors, France showed similarities with Japan and Germany; France disbursed over 90% of its aid funds to the 140 and the 230, resulting in the highly concentrated aid measured in HHI (Table 5). The propensity towards the infrastructure sectors of France, Germany and Japan could be part of the reason for their principal utilization of the public institutions channel to deliver their aid (Table 2); the infrastructure sectors more likely require direct engagement with GoU for negotiations and collaborations than other sectors.

In sum, while remaining skeptical about the institutional strength of GoU, the bilateral donors increased the overall aid disbursement to Uganda during the recent decade, compared with the period of 2002-2009. The majority of the bilateral aid to Uganda was carried out as the project-type, in the social sectors and through the non-state actors, which left small leverage for GoU’s control over the aid funds. The donors might have chosen tactical responses to the cloudy political environment in Uganda by tightening supervision over their aid funds, instead of cutting the aid itself. In line with this, Nordtveit (2014) asserts the DAC donors may not be selective on governance when deciding total bilateral aid, but they could be when allocating aid (Nordtveit 2014). The aid portfolio of Uganda during the past decade was mostly shaped by the largest donor, US that showed specific preferences for certain aid modalities and appeared rather reluctant to take a lead role for coordinated aid activities. All other donors seemed to lean towards going their own way as well, which was attributable to placing Uganda in one of the recipient countries with the most fragmented aid.

5. Conclusions

Uganda demonstrated a unique case that allowed witnessing changes in donor behaviors under the unclear institutional circumstances of the recipient country. In the early 2000s, Uganda was the pioneer recipient country for donor coordination and the early supporter of the 2005 Paris Declaration for aid effectiveness. Accordingly, the donors rewarded Uganda by increasing GoU’s leverage over the aid funds as well as the aid itself. However, disagreements between the donors and GoU on the politically sensitive issues including the allegedly flawed election and the corruption soon shifted course. Part of the donor community reacted to their disagreements by pressuring GoU with aid suspension and tightening aid fungibility. Tangri and Mwenda (2006), however, indicate some donors may be more reluctant to raise governance issues and impose pressure on the recipient government because it would signal questionable legitimacy of their foreign aid spending (Tangri and Mwenda 2006). Despite the reversed status, Uganda continued to receive stable or increasing amounts of aid from the DAC donors during the recent decade. To some donors, a decision to cut aid to GoU may not be a viable option, given their strategic interests in and historical ties to the country.

The collective aid profile pointed the donors appeared to take more cautious and risk-averse approaches to ensure aid efficacy under the weakening institutional environment through indirect engagement with GoU. Even so, the donors might have baseline preferences for the state-to-state aid engagement because direct relations with the recipient government could strengthen ties in general and carry payoffs in non-developmental issues (Dietrich 2013). But the current level of donor confidence in GoU seems to hold them from returning to their early 2000s level of state-to-state engagement.

The confidence issue on GoU is not limited to the donors, however. Findley et al. (2017) indicates in a study that the Ugandan public preferred the donor-funded programs over the government-supported programs because they saw GoU suffering from corruption and clientelism. Thus, the Ugandan public perceived the donor-funded programs might benefit them more and the aid conditionality could prevent the political elites from capturing aid for their own interests (Findley et al. 2017). Perhaps further studies are needed to generalize Findley’s argument, but it suggests GoU examine confidence issues with the donors and the public to facilitate its development.

Lastly, the emergence of non-traditional donors, especially China, may change the aid-relation dynamics between the DAC bilateral donors and GoU. The entry of China into the aid business is typically thought of as undermining the aid conditionality shaped mostly by the traditional donors. China tends to use aid to secure economic interests and natural resources with few political strings attached. Thus, growing Chinese presence in the Ugandan aid arena is sure to influence the aid relations between GoU and the traditional donors (Kasirye and Lakal 2019).

Disclosure statement

No potential conflict of interest was reported by the author(s).