Is it time to ‘decolonise’ the fungibility debate?

Abstract

Recent literature has established that development assistance is often fungible and that this is undesirable. In line with current efforts to ‘decolonise development studies’, we critically reflect on the underlying assumptions of this line of thinking. We establish a framework that differentiates between potential positive and negative fungibility. We hypothesise that recipient governments can redirect their own funds and achieve positive fungibility, if (1) the marginal value added in the alternative target sector/region is higher; (2) equity concerns are adequately addressed when other sectors/regions are supported; and (3) temporal delay helps to cushion instability of aid flows. There are indications that this positive fungibility might be quite prevalent. Future fungibility research should therefore no longer assume that fungibility is in itself undesirable.

Keywords:

• Aid effectiveness
• fungibility
• positive fungibility
• sustainable development goals
• development assistance
• decolonising development studies

1. Introduction

The idea that foreign aid is fungible has been discussed in development literature since the 1950s when the Chief Economist of the World Bank observed that even if money is tied to specific projects, it remains fungible.¹ Fungibility of aid can be defined as aid resources intended to finance a specific expenditure that are ultimately used to finance an entirely different expenditure. White and Dijkstra have noted that fungibility is the idea that ‘aid does not pay for the item it is accounted for but for the marginal expenditure it makes possible’.² The conventional wisdom in such cases is to assume that the alternative use of aid is less productive or socially beneficial than the intended purpose.³ This paper contributes to the current debate on fungibility by demonstrating theoretically that this conventional wisdom needs to be challenged and proposes a more nuanced framework for differentiating between potential positive and negative fungibility. There is a broader current movement within development studies – sometimes referred to as ‘decolonising development’ – in which conventional wisdoms, often originating from a colonial and/or patronising perspective, are challenged. This journal, Third World Quarterly, has been at the forefront of this movement.⁴ The negative view on fungibility, which is based on the assumption that donors know best, is also one of these concepts.

Empirical studies on foreign aid largely conclude that sector-specific development assistance is usually fungible,⁵ and a large number of these studies assume that the fungibility thus discovered is undesirable or a threat to meeting global development standards.⁶ Studies measuring fungibility have had a long-term impact on donor strategy and policy formulation regarding aid distribution and delivery. Donors have been developing and testing various methods to reduce fungibility or its perceived negative effects⁷ such as requiring matching funds, engaging in performance-based payment or reducing the consequences of fungibility by targeting aid to well-governed countries or regions. Measuring fungibility and developing aid strategies on the assumption that fungibility is by definition undesirable might be incorrect.

Not many studies have analysed the long-term positive or negative influence of fungibility from the perspective of recipient countries. Our paper tries to address this gap in the literature from a foundational theoretical perspective. We look at fungibility of development assistance from a new viewpoint and link it to the effective achievement of the sustainable development goals (SDGs). We try to add to the existing literature by developing a theoretical framework to explain the desirability of fungibility. The research question therefore is: under what circumstances is fungibility of development assistance desirable or effective when measured in terms of progress towards achieving SDGs? Indeed, as McGillivray and Morrissey have noted, ‘There is nothing inherently wrong or inappropriate about fungibility; all it indicates is that donors and recipients have differing views about how expenditures should be allocated. The presumption that donors are right need not always hold’.⁸ Building on that crucial statement, this paper tries to establish a potential positive link between SDGs and aid fungibility. We aim to address the issue of ‘positive’ fungibility from both a theoretical and a practical angle.

We start off by providing an analytical background discussing the current relevant academic debates on fungibility. In the section that follows, we first provide an overview of the four different types of fungibility encountered in the literature. We then move on to propose theoretical approaches to differentiate between positive and negative fungibility. In this paper we define fungibility as positive if the reallocation of aid (fungible aid) leads to more effective attainment of the SDGs than non-fungible aid. Based on the theoretical approaches thus defined in section 3, and considering the peculiarities of actual donor behaviour, we then analyse in section 4 the occurrence of positive fungibility in practice as hypothesised. We conclude that, in all likelihood, there is a share of fungibility that can be labelled positive.

2. Current debates on fungibility and practices to reduce it

In this section we provide an overview of the growing relevant academic debate on fungibility. The first discussion revolves around how fungibility is defined, which in turn leads to the question of how fungibility is appraised: is it seen as something positive or negative? After that, we zoom in on how fungibility is measured, as there are ongoing controversies that also impact our framework on positive fungibility. The last section focuses on the measures that are being taken to reduce fungibility.

2.1. Defining fungibility: multiple types of fungibility

A review of the literature has helped us in identifying four different types of fungibility: sectoral fungibility, geographic fungibility, temporal fungibility and general fungibility.⁹ We discuss these types of fungibility throughout our literature review and so we summarise them altogether as follows:

1. Sectoral fungibility is the result of aid intended for a particular sector leading to a government diverting its funds to an alternate sector, such as aid given for the education sector resulting in a government diverting its funds to the defence sector instead.¹⁰

2. Geographic fungibility implies that aid is not fungible within sectors but rather between geographical regions, as can be seen in the research done by Wagstaff on Vietnam, where aid was not fungible across the health sector but was fungible between provinces – that is, the government diverted its own funds to alternate provinces when the World Bank supported the health aid programme in one province.¹¹

3. Temporal fungibility is the fungibility resulting from the unpredictability of aid flow from the donor side, as a result of which the recipient country reduces its own spending during the time period when aid flow is smooth to compensate for the times when aid flow is unpredictable.

4. General fungibility is the result of aid given for general purposes being used for some other purpose, such as when aid is given for consumption expenditure but is instead used for investment expenditure.¹²

2.2. Appraising fungibility: nearly always a negative perception

Multiple empirical studies have been conducted to understand the link between fungibility and aid effectiveness. For example, Pack and Pack tried to empirically analyse the case of the Dominican Republic to understand aid fungibility and concluded that that there was clear evidence of categorical foreign aid resulting in a decrease in development expenditure in that sector. They observed that 88 cents per dollar was being used for debt repayment by the government, and that development expenditure decreased by 5 cents for every dollar of foreign aid in Dominican Republic during the period 1968–1986.¹³

Khailji and Zampelli developed a similar model to analyse the fungibility of foreign aid received by the Government of Pakistan. They argued that most of the donors prefer to give categorical aid (for example for dams, bridges, roads, etc.) instead of general financial assistance, because of the idea that the given aid will be used for its intended purpose. In the case of Pakistan, they concluded that US assistance to Pakistan, in the form of both military and non-military aid, was transferred completely to fungible resources, because for each rupee of aid, public spending increased by 0.26 rupees and the remaining 0.74 rupees were absorbed by private-sector consumption. It is noteworthy, however, that they did not differentiate between aid being given to private parties such as nongovernmental organisations (NGOs) or cooperation but rather assumed that all the aid passed to the government, making their results somewhat biased.¹⁴

Similar studies related to fungibility at country or project levels were conducted on Indonesia, Sri Lanka and Tanzania, and a combined study was conducted on Thailand, Honduras and Rwanda, discussing the fungibility of aid in HIV programmes.¹⁵ Most of these studies identified the presence of fungibility for development aid in the respective countries and raised concerns about the effectiveness of foreign aid being given to developing countries.

Pettersson further developed the model outlined by Pack and Pack to look at sectoral fungibility by empirically linking economic growth, poverty and aid.¹⁶ Using data from 57 countries, Pettersson estimated that 65% of total sectoral aid is used for purposes that it was not originally intended for. He compared fungible and non-fungible aid sectors and concluded that no evidence was found of the latter working better than the former when measured against economic growth. However, when looking at the effectiveness of foreign aid in poverty reduction, he did conclude that non-fungible aid sectors have helped more in improving the welfare of people compared to fungible aid, which had no empirical effect.¹⁷

Wagstaff conducted one of the first studies that did not assume fungibility to be necessarily undesirable, but rather accepted the possibility that there might actually be appropriate reasons to redirect resources after support has arrived. Wagstaff discusses the case of health projects in Vietnam funded by the World Bank where the government and bank had a mutual understanding that government funds would be diverted towards provinces that were not targeted by the Bank. This arrangement allowed the Vietnamese government to achieve a higher level of spending in both project and non-project areas.¹⁸ Henceforth, we will use the term ‘mutually agreed fungibility’ for this type of fungibility.

2.3. Curtailing fungibility: vain efforts of donors to reduce fungibility

Because many academics (and donors) hold the view that fungibility is relatively undesirable, significant efforts (or recommendations) have been made to reduce it. These efforts have often been in vain, resulting mostly in increased transactions costs for both the donors and the recipients. There have been two opposing ways in which donors have aimed to reduce fungibility and/or its consequences: (1) restricting the flexibility of domestic resources of the recipient government and (2) circumventing fungibility by loosening the definition of the intended sector/region for international funding.

Restricting the possibilities for fungibility – limiting the flexibility of domestic government resources. In 1998, the World Bank published a report called Assessing Aid that highlighted the problem of aid fungibility with regards to aid effectiveness. In the report, Dollar and Pritchett concluded that, to reduce fungibility, donors need to identify the quality of governance of the country to which they are donating by looking at public expenditure and analysing the fiscal discipline of the government, priorities across sectors and technological efficiency of the recipient government. The donors should not place additional conditions on the recipient countries, but rather should focus on financing parts of projects that the government is not financing, thus ensuring very little fungibility.¹⁹ If the recipient governments are not spending any money in a specific region or on a specific theme in the first place, they cannot redirect this funding. The shortcoming of this method is that – besides imposing preferences on aid-dependent countries – donors would sometimes like to complement local funding, which is not possible in this case.

To restrict the flexibility of domestic funding, other modalities were also developed including aid on delivery and performance-based support.²⁰ In addition, pooling funds and sector-wide approaches (SWAps) have been popular in recent years; these require the recipient government to take the lead on the project being financed by the donors using its own funds and/or administrative capabilities.²¹ Some multilateral donors (including the International Monetary Fund [IMF]) have also suggested techniques like ring-fencing to ensure that the recipient government is not able to reduce its funding for the specific sector to which aid was being given. Although these approaches are promising, they require a high level of commitment and reliability among the donors themselves, which is sometimes hard to maintain.²² A last approach to reduce the flexibility of domestic resources has been to provide off-budget support, such as by giving aid to local NGOs. Ignoring local governments and creating a parallel system through NGOs also has the disadvantage that is it less sustainable in the long term.²³

Circumventing fungibility: loosening the definition of ‘intended’ sectors and regions. Some donors have opted to provide general budget support to reduce concerns of fungibility, as the donors do not target their support to a specific region or theme.²⁴ With general budget support, it is accepted that recipient governments may use the resources as they deem fit. General budget support is more often provided to governments in places that have demonstrated an acceptable level of governance or poverty reduction strategies. Despite a promising start for general budget support in the early 2000s, this aid modality is already on the decline.²⁵ It appears that political reasons have contributed to the declining popularity of this aid modality, along with changes in donor interests and an increase in concern about the behaviour of the recipient countries. To be more accountable to taxpayers, donors seem to have switched their focus over time to project aid.

In sum, it appears that even if donors choose different aid modalities to reduce or circumvent fungibility, there are various drawbacks involved. This article therefore suggests that instead of putting effort into reducing fungibility, it might be more useful to put energy into analysing what types of fungibility are positive and capitalise on that. To conclude, it is important to understand the main takeaways of the current academic debates on fungibility. This review demonstrates that there is ample debate on whether, where and how much fungibility exists. We noted, however, that there has hardly been any debate – apart from, for instance, the article by Pettersson – on the desirability of fungibility, which is precisely the aim of our article.

2.4. Motivation behind fungibility: do donors know better?

The implicit assumption of most of the research on fungibility has been that donors know what is best for achieving SDGs (or their precursors). However, is this really the case? The selection of a partner country to work with is one of the key aspects of the development policy of any donor.²⁶ Donor aid allocation could be based on the recipient’s need or performance, or it could also be based on the strategic interests of the donors themselves such as to develop a trade relationship, gain political support or fulfil self-interest – especially commercial interests.²⁷ Baydag and colleagues analysed aid patterns for the eight biggest donors of the Organisation for Economic Co-operation and Development (OECD)’s Development Assistance Committee (DAC) and found out that some donors, including the United States and Japan, are highly strategically oriented when selecting aid partners. The United States selects aid partners on the basis of security interests, while Japan prioritises aid to Asian countries to maintain its own security and prosperity.²⁸ In a similar vein, Berthélemy used an empirical methodology and found that aid could be used to strengthen commerical ties with the recipient country in addition to political alliances. Even if some donors might not have strong geopolitical motivation for aid, almost all of them have trade interests.²⁹

We can therefore conclude that donor motivation is not always focused on achieving the SDGs, which is our yardstick. Aid may sometimes be driven by strategic motives that might be more beneficial for the donors themselves than for the recipient’s developmental goals. To be able to theorise on the desirability of fungibility, it is first necessary to focus more on disentangling the concept by describing the various types of fungibility.

3. Theoretical approaches to positive fungibility

Based on the previous arguments, in this section we will theoretically hypothesise three different angles of ‘positive’ fungibility. First, we will expand on welfare effects, which are mostly grounded in micro-economic theory. Second, we will turn to redistribution and utility maximisation, which are concepts mostly used in debates on the moral economy. Third, we will focus on transaction costs, delays and government spending, which are mostly derived from institutional economics.

3.1. Welfare effect

To better understand the welfare effect, we can consider a simple static microeconomic model of utility maximisation that has been discussed several times in the literature and which relies on the model developed by McGillivray and Morrissey.³⁰ To simplify our model, we assume that a developing economy consists of only two sectors: health and education. The government divides its expenditure between these two sectors only and therefore spends a fixed amount in the health and education sectors. We also assume that, at the current level of spending between the two sectors, the utility of the government is maximised.

Let us assume that the government receives a certain amount of development assistance for health (in the form of a cash transfer). Now, the government has two options: (1) it can either keep its own expenditure in the health sector the same and thus allow the overall funding of the health sector to increase by the amount of aid; or (2) it can divert some (or all) of its own funding from the health sector to the education sector, which may be relatively underfunded because it did not receive any development assistance. This would maintain the same ratio of spending between health and education. At this point, we leave the broader fiscal effects, such as the flypaper effect – that is, when a central government grant has the effect of increasing local government expenditure by more than the amount of the grant – out of the picture.³¹

By keeping the spending ratio the same as before, the recipient government tries to maintain the best allocation of its budget. This would mean that the funding of the donor is fungible, as some of the recipient government’s own funding that was intended for health spending is now redirected to education. Fungibility in this case would be positive because it allows the government to maintain its allocation efficiency and therefore maximise its utility; the only way it could avoid fungibility in this case is if it manages to increase its revenue through borrowing or taxation, but we know from the literature on aid and taxation that aid may or may not decrease total recurrent revenue (the literature presents conflicting views in this regard).³² By diverting some of its funds away from the health sector and into the education sector, the government can increase the marginal effect in the health sectors – that is, if the amount of funds in the health sector increases drastically, the marginal value of the dollar in that sector declines. Putting it in terms of SDGs, which focus on different sectors, it makes sense that – for effective achievement of the SDGs – the government diverts its funds to an alternative sector when funding in the initial sector increases, as this would increase the marginal effect of aid. Hence, fungibility allows for more effective resource allocation in terms of a comparative static analysis. We can therefore establish from our discussion that fungibility is positive when the marginal added value of the aid dollar in the alternative sector or region is higher than that in the intended sector or region. Not all fungibility is necessarily positive, of course, as the example below indicates.

Alternatively, there could also be negative fungibility. Let us assume that, between the two sectors, the government is spending more money on education to a degree that prevents achievement of SDGs. This decision could be driven by considerations of local political economy (eg to strengthen the powerbase of a certain minister or political agent). The donor decides, therefore, to give aid to the underfunded health sector. The government may decide to further divert its funding away from health (where there is a high marginal added value) and into the education sector (with low marginal added value) as a result of incoming aid, resulting in fungibility. The donor in this case was unable to identify and rectify such misallocation. This kind of fungibility would be negative in nature, as it creates sectors with lower overall welfare (compared to the optimal scenario) because no use is made of the higher marginal value added in in the education sector. Hence, the resource allocation is inefficient. We can therefore conclude that fungibility would be negative if the marginal value added in the alternate sector or region is lower than that in the intended sector or region.

3.2. Redistribution and utility maximisation

We can discuss a similar model from the point of view of redistribution as well. Let us assume that the donor gives aid for the health sector in an easy-to-reach province, but it is actually a farther-away province that is underfunded. The recipient government therefore has the same options: it can either contribute the same amount of its own funding to the health sector in the two provinces (no fungibility), or it can decide to keep the ratio of total funding to the two provinces similar (fungibility), which would mean diverting some of its own funds away from the easy-to-reach province.

With any spending option, the government must choose between the options of equality and total welfare. The debate over equality versus total welfare has been discussed quite extensively in the economic and political science literature. Equality is concerned with impartial application of rules (or resources) such that similar cases are treated in the same way, whether a citizen lives in a province that is easy or hard to reach, while an increase in total welfare may require the unequal allocation of resources to achieve greater happiness of others.³³ For instance, the government of a recipient country faces a trade-off between increasing total welfare and increasing equality between provinces. The recipient government, having a better understanding of the local situation, might prefer one province rather than another. The donors, on the other hand, might have their own intrinsic motivation for aid allocation, as discussed in section 2.4, which may not be directly relevant to welfare increase or efficiency of allocation in the recipient country, but may rather be motivated by their own security concerns or commercial interests. The donor may also be facing information asymmetry, and therefore may not have all the information related to the socio-economic factors of the recipient country necessary to make informed funding allocation decisions (we discuss this further in section 3.3). Thus, the allocation priorities for the donor and the recipient countries might differ. We can therefore conclude that fungibility resulting from such a case could be considered positive if the inequality between the alternative and intended regions can be reduced through reallocation of funds.

Contrary to this, the recipient government’s behaviour could sometimes lead to increased inequalities between regions while also negatively affecting welfare. For example, a recipient government may discriminate against a certain ethnic group or a specific region based on political motivations, that may lead to an increased degree of marginalisation of its people or increased levels of poverty in the region discriminated against. If the recipient government diverts its funds away from the intended region and into the alternate region with this motivation, the resulting fungibility would be negative in nature. Thus, we can conclude that fungibility is negative when the reallocation of funds increases the inequalities and degree of marginalisation between the alternate and intended regions.

Looking at fungibility from the angle of inequality and redistribution encompasses more than just the geographic element used an example here; it can also apply to various aggregations of the population, such as gender, ethnicity, profession, etc.

3.3. Institutional inefficiencies

Positive fungibility can also be viewed through the lens of public choice theory and institutional economics. The donor, for instance, has to compete for votes by indicating how well government policies and plans have served the interests of the voters when deciding on aid commitments. As a consequence, each new minister of development aid in an OECD country faces pressure to present a ‘new’ or ‘better’ policy, and needs to respond to public pressure, which leads to various aid fads.³⁴ The recipient government, on the other hand, has to juggle two (often) competing priorities: (1) how well it spent the funds supplied to it by the donor and (2) how well it caters to the interests of its voters.

We can also consider the lag theory of fiscal or monetary policy to understand the behaviour of donors and recipients. There are various lags that delay the implementation or effectiveness of a fiscal or monetary policy, including information lag, recognition lag, action lag and implementation lag.³⁵ These lags are the result of the time taken by different governmental (and nongovernmental) institutions to access and approve (or deny) the proposed policy. Similar to this, there is always a time lag between donor government commitments of funds for recipient countries and when the funds are actually disbursed, as can be seen by the data on the OECD credit reporting system.³⁶ This is especially relevant when there are temporal shocks on the donor side that can delay disbursements and affect the budget planning of the recipient government.

Thus, as the interests of the two actors – the donor as principal and the recipient as agent – conflict, this leads to delays in disbursements by the donor or sometimes a complete withdrawal from the commitment. The time lag can further increase if the donor government faces institutional inefficiencies, large bureaucracies within its government or a change in political ideologies. Thus, under such circumstances, we can consider fungibility positive if delays in internal funding by the recipient country help to stabilise flows to particular sectors or regions.

Conversely, the government may delay its own spending in the development sector or show irregular spending patterns because of diversion of funds to a non-development sector, including military expenditure (especially under a military regime), election campaigning (nepotism) or corruption. Delays in disbursements of funds in such cases would increase the temporal shocks and destabilise development projects even more, thus resulting in negative fungibility. We can therefore establish that fungibility would be negative if delaying government funding would further destabilise financial flows to a particular sector or region, making that sector prone to temporal shocks.

4. Positive fungibility in practice

Based on our theoretical framework and literature review, we try to reframe fungibility by establishing the conditions for positive fungibility. In this section the prevalence of those conditions is researched. We conclude that these conditions might quite often be present, which leads us to the provisional conclusion that fungibility might quite often be positive. Please note the caution we use here: we argue that it is possible that positive fungibility might be prevalent, not that it is prevalent, as recipients sometimes make funding choices that are not in line with the SDGs. We argue that more research is needed to detect this. This paper developed a theoretical overview addressing the conditions under which these forms of fungibility could be positive. Based on our theoretical framework, we established that the positivity of fungibility depends on:

1. The marginal added value of the aid dollar in the alternative sector or region compared to that of the intended sector or region (if the marginal added value is higher in the alternative sector or region, the fungibility could be considered positive). This is linked to the theoretical approach of the welfare effect in section 3.1.

2. The level of inequality between the alternative region and the intended region (if inequality between the alternative and the intended region can be reduced, the fungibility could be considered positive). This is linked to the theoretical approach of redistribution and utility maximisation in section 3.2, and this inequality perspective can also apply to other aggregations of the population (eg gender, occupation, etc.).

3. The degree to which delaying government funding (temporal fungibility) helps to cushion the volatility of aid funding (if delaying own funding helps to stabilise flows to particular sectors or regions, this fungibility could be considered positive). This is linked to the institutional economics approach, linked in turn to institutional efficiencies in section 3.3.

We will be discussing each of the criteria in detail in the following sections.

4.1. Higher marginal added value in alternative sector/theme

Based on the definitions and theoretical framework established above, we hypothesise in this paper that fungibility might be positive if the marginal added value of the aid dollar is higher in the alternative than in the intended sector, region or time period. This can be the case, for instance, because:

1. Donors are known to have their fads and their donor regions or themes, as well as their orphan regions and themes.³⁷ Donors tend to pursue their own political interests and agendas when allocating aid that may be different from the need of the recipient country.³⁸ There are ample instances of donor fads; Hilhorst and Douma, for example, documented amongst others that there was so much support for victims of sexual violence in the DRC that there was a sharp rise in invented rape cases.³⁹ Similar cases of overfunding have been documented in Aceh (Indonesia) and Sri Lanka in the wake of the tsunami.⁴⁰ Aid from international NGOs was also found to be strongly clustered.⁴¹

2. Local governments might wish to cash in on fads and redirect their own funding elsewhere.⁴² This can result in fungibility that is desirable for the recipient country as well as the donors. From an SDG perspective this could be considered positive fungibility.

4.2. Reduction of inequality levels

We hypothesised that fungibility can be positive if, by re-allocating its own resources, the recipient country can reduce its resource inequalities (the example of the easy-to-reach vs remote regions). To test the notion of ‘leaving no one behind’, Briggs used a geotagging survey to analyse the allocation of aid in Nigeria, Senegal and Uganda by all donors and found that aid is not targeted to the poorest people within a country or even within a region. This kind of allocation was not limited to multilateral donors, including the African Development Bank or World Bank, but also applied to bilateral donors.⁴³ In a similar vein, Thiele and colleagues observed that allocation of education aid by donors was not actually based on the indicators set by the United Nations; surprisingly, none of the donors analysed took primary school enrolment rates or similar factors into consideration when allocating aid for education.⁴⁴ The donors usually lack either the political will or the ability to reach the poorest in the population, and so it made sense for the recipient government to step in and ensure that the funds were distributed equitably.

A relevant example related to positive fungibility can be found in Wagstaff’s analysis of the case of Vietnam, where the World Bank encouraged the recipient country to divert its funds to other provinces that were not being funded directly by the donor: that is, mutually agreed fungibility. So, while the donor wanted to increase the welfare of the region the government reallocated its funds to increase equality between regions, as has been discussed in section 3.2 of this paper. According to Wagstaff it is important for the discussion of fungibility to look at the overall impact of aid on the country; that is, those areas that were not formally covered by the project should also observe an increase in the quality and quantity of facilities and outcome by more than they would have done in the absence of the project.⁴⁵ Thus, by making use of positive fungibility, the recipient and the donor were able to increase the overall welfare of the country. Also, in light of the guidelines of donors to ‘do no harm’ and encouragements to work in a conflict-sensitive manner, donors might hence even encourage recipient governments to re-organise their own expenditures to enable an equitable spread.

4.3. Absorption of temporal shocks

Aid donors are known to be particularly unreliable and inconsistent. The predictability of aid flows is low, and the volatility is usually high.⁴⁶ Recipient countries are usually aware of this principal–agent problem (discussed in section 3.3) and may delay their own disbursements when the option of external funding is made available. They can stockpile their own funding and wait to release it when there is a dry spell in external funding, leading to positive fungibility. From a practical point of view, this kind of fiscal behaviour requires that the recipient country has the ability to save; if no such mechanisms exist within the government, the recipient government risks borrowing to maintain its spending once aid starts to dry up.

Development assistance by donors is, in general, quite unpredictable.⁴⁷ Even for investments into crucial programmes where stability of funding is a necessity – for instance, the antiretroviral treatment for HIV patients – fluctuations exist. These fluctuations could arise because of differences in commitment and disbursement that might result from changes in political leadership on the donor side, the inability of the recipient government to meet its agreed-upon targets set by the donor or simply delays in initiation of the intended project due to bureaucratic and institutional inefficiencies on either the donor or the recipient side. If we analyse OECD data on commitments and disbursements by donors to developing countries, we observe that between 2000 and 2016, the disbursements have always been lower than the commitments except in 2006.⁴⁸ In this case it is smart for the recipient country to suspend its own funding and save for when a shortfall arises.

In 2010, for example, the newly elected Dutch government introduced a new aid policy that brought about budget cuts on official development assistance (ODA) to almost 800 million euros and reduced the number of partner countries from 33 to 15. The cuts negatively affected the budgets of social sectors in partner countries, especially in the health sectors, leading to an alleged 12,000 additional deaths in Burkina Faso.⁴⁹ If the partner countries had reduced their own funding during the time when aid was coming in and had reintroduced it when the dry spell came, it would have reduced the spending shock in the health sector.

In sum, we can redefine fungibility as potentially positive and potentially negative by taking the SDGs as a yardstick. Potential positive fungibility is therefore the type of fungibility that helps the recipient countries avoid (or absorb) temporal shocks, reduce inequality between regions or other aggregations of the population, and/or create complementarities between external and internal funding through marginal value added towards the achievement of SDGs. However, it would be noteworthy to point out the main limitation of our analysis, which is the lack of empirical evidence for the three cases we have made above, as there has been almost no research in the literature on positive fungibility. We acknowledge that recipient behaviour is also not always focused on maximising public welfare or development goals, but in this section we try to demonstrate that the conditions for positive fungibility are often present.

5. Discussion and conclusion: time to decolonise fungibility?

This exploratory study aims to evoke a different angle on the debate concerning fungibility and aid effectiveness. We do not argue that all fungibility is positive – far from it. Negative fungibility probably exists (as discussed in section 3), but positive fungibility also probably exists, and has not yet been well documented. Further investigation and greater elaboration are needed due to the limited level of discussion in the literature about SDGs and fungibility. Empirical data need to be collected to analytically test the concept of positive fungibility theorised in this paper. To generate insights into fungibility, especially if a deeper understanding into the reasons for fungibility is desired, a time-consuming combination of qualitative and quantitative research is needed.⁵⁰ Practical application and testing of the concepts discussed in this paper will therefore require substantially more time and effort.

With this paper, we hope to contribute to a shift in the debate on fungibility: it is time to look at fungibility differently. Instead of seeing it as something exclusively negative, scholars are invited to understand that fungibility might be positive. We hope to use this paper to encourage a more inclusive and healthy debate on fungibility and development. In follow-up research, we will therefore make use of the hypothesis that fungibility in development assistance may sometimes be necessary to achieve long-term positive progress towards SDGs. Developing countries can have a greater understanding of their socio-economic needs than donors do and may therefore be able to allocate their resources more effectively by investing in another development sector, instead of complementing the development assistance they are receiving for a specific sector.

Within the broader field of development studies, the call to ‘decolonise development’ is steadily increasing.⁵¹ According to the critics of development studies, assumptions constructed during colonial times regarding racial and civilisational hierarchies still form much of the rationale about the world and determine what is worth studying. In this thinking, the donor often gets the benefit of the doubt and recipients do not. Is fungibility one of those topics in which implicit biases and assumptions still prevail? Is there a need to decolonise fungibility? There are various indications that this might be the case. Looking at the authors cited in this study, a small minority come from developing countries, which is one indicator for a need to decolonise a subfield.⁵² When we analysed texts on fungibility it was striking how often it was implicitly assumed that fungibility, which could also be framed as an act of agency by the recipient government, was something that needed to be rooted out. Decolonising fungibility is not a mere academic interest: quite some effort is currently being taken to reduce fungibility and develop aid modalities to this effect. What emanates from this paper, however, is that there are potential drawbacks to these efforts to curtail fungibility. Instead of aiming to reduce fungibility lock, stock and barrel, the topic should be dealt with in a more pragmatic and evidence-based way.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Notes on contributors

Zunera Rana

Zunera Rana (1990) is a research and teaching assistant at Rhine Waal University of Applied Sciences in Germany and a PhD candidate at Radboud University in Nijmegen. Originally from Pakistan, Zunera did her bachelor’s in economics and finance at Lahore School of Economics, Pakistan, and her master’s in economics at Rhine Waal University of Applied Sciences. Her PhD research is focused on aid effectiveness and fungibility of development assistance.

Dirk-Jan Koch

Dirk-Jan Koch (1980) is serving as the Chief Science Officer of the Netherlands Ministry of Foreign Affairs and is the Special Professor of International Trade & Development Cooperation at Radboud University in Nijmegen. Dirk-Jan finished his PhD (cum laude) at the Radboud University in 2009 focusing on geographic choices of international development NGOs. He was based for 5 years in the DRC and 2 years in Kenya, where he worked, amongst other positions, as a regional director at Search for Common Ground, as a professor at the Catholic University of Kinshasa and as a diplomat of the Netherlands to the DRC.

Notes

1 Dollar and Pritchett, Assessing Aid.

2 White and Dijkstra, Programme Aid and Development.

3 Pettersson, “Foreign Sectoral Aid Fungibility.”

4 Bourgois, “Decolonising Drug Studies in an Era of Predatory Accumulation”; Capan, “Decolonising International Relations?”; and Spiegel et al., “Decolonising Online Development Studies?”

5 Pack and Pack, “Foreign Aid and the Question of Fungibility”; Lu et al., “Public Financing of Health in Developing Countries”; Dieleman, Graves and Hanlon, “Fungibility of Health Aid”; Sijpe, “Is Foreign Aid Fungible?”

6 Lu et al., “Public Financing of Health in Developing Countries.”

7 Harper, “Fungibility of Aid.”

8 McGillivray and Morrissey, “Aid Fungibility in Assessing Aid,” 419.

9 McGillivray and Morrissey, ibid., also identify ‘unintended fungibility’ resulting from erroneous inference. They argue that sometimes there is an appearance of fungibility only because the allocation of spending is observed instead of analysis of the budgetary process. Because it is not a proper ‘type’ of fungibility per se, but rather a misidentification of something else that is erroneously labelled fungibility, it is excluded from this discussion.

10 Khailji and Zampelli, “Fungibility of US Assistance.”

11 Wagstaff, “Fungibility and the Impact of Development Assistance.”

12 Morrissey, “Fungibility, Prior Action and Eligibility”; Leiderer, “Fungibility and the Choice of Aid Modalities.”

13 Pack and Pack, “Foreign Aid and the Question of Fungibility.”

14 Khailji and Zampelli, “Fungibility of US Assistance.”

15 On Sri Lanka see Pack and Pack, “Foreign Aid and the Question of Fungibility”; on Indonesia see Pack and Pack, “Is Foreign Aid Fungible?”; on Tanzania see Martínez-Álvarez et al., “Is Development Assistance for Health Fungible?”; and Thailand, Honduras and Rwanda see Garg et al., “Study Raises Questions about Measurement.”

16 Pettersson, “Foreign Sectoral Aid Fungibility.”

17 Ibid.

18 Wagstaff, “Fungibility and the Impact of Development Assistance.”

19 Dollar and Pritchett, Assessing Aid.

20 Leiderer, “Fungibility and the Choice of Aid Modalities.”

21 Ohno and Niiya, “Good Donorship.”

22 Swedlund, Development Dance.

23 Herfkens and Bains, Reaching Our Development Goals.

24 Clist, Isopi, and Morrissey, “Selectivity on Aid Modality.”

25 S. Koch et al., “Rise and Demise of European Budget Support.”

26 Baydag, Klingebiel, and Marschall, Shaping the Patterns of Aid Allocation.

27 Ibid; and Berthélemy, “Aid Allocation: Comparing Donors’ Behaviours.”

28 Baydag, Klingebiel, and Marschall, Shaping the Patterns of Aid Allocation.

29 Berthélemy, “Bilateral Donors’ Interest.”

30 McGillivray and Morrissey, “Aid Illusion.”

31 Ibid; and Barnett, “Preference Revelation and Public Goods.”

32 McGillivray and Morrissey, “Aid Illusion.”

33 Rawls, Theory of Justice.

34 Hillier, “Insurance Hits Peak Hype.”

35 Jovanovski and Muric, “Phenomenon of Lag.”

36 OECD, Recent Trends in Official Development Assistance.

37 Wagstaff, “Fungibility and the Impact of Development Assistance.”

38 Thiele, Nunnenkamp, and Dreher, “Do Donors Target Aid”; Dreher, Nunnenkamp, and Thiele, “Does US Aid Buy UN General Assembly Votes?”; and Alesina and Dollar, “Who Gives Aid to Whom and Why?”

39 Hilhorst and Douma, Fond de Commerce?

40 Telford, Cosgrave, and Houghton, Joint Evaluation of the International Response.

41 D.-J. Koch, Aid from International NGOs.

42 Wagstaff, “Fungibility and the Impact of Development Assistance.”

43 Briggs, “Does Foreign Aid Target the Poorest?”; and Briggs, “Leaving No One behind?”

44 Thiele, Nunnenkamp, and Dreher, “Do Donors Target Aid.”

45 Wagstaff, “Fungibility and the Impact of Development Assistance.”

46 Ooms et al., “Crowding Out.”

47 Bulíř and Hamann, “Volatility of Development Aid.”

48 OECD, Recent Trends in Official Development Assistance.

49 Ministry of Foreign Affairs, Gap Left behind.

50 Martínez-Álvarez et al., “Is Development Assistance for Health Fungible?”

51 El Kadi, “Decolonising Development Studies.”

52 Briggs and Weathers, “Gender and Location.”