The term fungibility has been misunderstood and misused in development circles, so creating confusion that leads to inappropriate policy conclusions.  This post goes back to basics and tries to bring a little rigour and clarity to the question of fungibility of foreign assistance.  As we will see, the technical concept of fungibility is irrelevant to the questions of whether aid is used as intended and whether it is effective. Nontheless donors do need to consider whether and how to take account of the possible reallocation of other resources that may occur when they give aid, and whether this has implications for whether and how they give aid.

Here is the summary version – the long version is below.

• To say that an asset is fungible means technically that one unit is directly convertible into another unit of the same asset.  But in development, people have come to use the term in a different, technically incorrect, sense.  They use it to mean that the recipient may respond to aid by changing the way they use their own resources, with the risk that aid “frees up” resources to be used in unpredictable ways.  (The tangential connection to the proper idea of fungibility is that aid may partly substitute for the other resources available to the recipient community.)
• Donors have two responsibilities which relate to this.  First, they should ensure that the aid they give is used for the purposes for which it was given.  Second, they should ensure that their aid is good value for money, which requires them to take account not only of the direct effects of aid, but also the indirect effects.   These indirect effects include the overall impact on resource use in the recipient community.  Resource reallocation is just one of the indirect effects, and it unlikely to be the most important.
• These two concerns are a close analogy to the general obligations in public financial management to ensure that public money is both properly spent and good value for money.  These are distinct concerns which are better not confused.   Aid can arrive safely and be used for the purposes intended even if the recipient community has made offsetting changes in its own resource use.  Being used for the purposes intended is a separate question from whether aid is achieving its broader goals and achieving good value for money: it is this latter question which may be affected by how it impacts resource use in the recipient community.
• Properly defined, neither fungibility nor liquidity of the aid affects the extent to which recipients can make offsetting changes in resource use.  The extent of offsetting changes is determined by whether recipients are willing and able to make choices about how to use their non-aid resources. The characteristics of the aid and how it is delivered make little difference to this.
• By mislabeling the question of broader resource use as an issue “aid fungibility” we create the misleading impression that the way we give aid is an important determinant of the extent and nature of the possible impact on overall resource use.  We create the impression that by choosing the right aid instrument, we can limit or prevent these effects.  In fact, aid could be entirely non-fungible and illiquid and still have big, unpredictable effects on resource allocation.
• There is a lazy assumption that these issues are more of a concern for aid given as budget support to governments.  Aid given in kind and aid given through NGOs is subject to the same issue of impact on overall resource use as aid given in the form of budget support.
• If donors are concerned about the indirect effects of aid on the allocation of government resources, then they should be looking for ways to strengthen the resource allocation process and to exercise more influence over the recipient’s budget allocations.  Giving their aid through NGOs is likely to have the opposite effect, since donors thereby exclude themselves from the dialogue about resource allocation.  Rather than try to bypass these issues, donors should increase their engagement in improving public financial management, supporting transparency and accountability to parliament, and providing aid through government budgets so that they have a locus to influence the recipient country’s use of resources.

A more detailed explanation follows.

 

What does “fungibility” mean?  What does “liquidity” mean?

An asset is said to be “fungible” if a unit of that asset is easily interchangeable with another unit of the same asset.  So a sack of barley is fungible (one sack of barley is pretty much like any other), but a bottle of wine is not fungible (because you would not readily exchange a bottle of Margaux for a bottle of Liebfraumilch).  Crude oil is fungible, diamonds are not.  Money is fungible – a $10 bill is just as valuable as another –which is of course a desirable characteristic of your medium of exchange.  For many assets, such as money, fungibility is a strength, not a weakness; this is implied by the Latin origin of the word, fungi, meaning “to perform”.

“Liquidity” is the first cousin of fungibility. If an asset is “liquid” this means it can be easily converted into another, different asset.  So although diamonds are not fungible, they are generally quite liquid because they can be easily sold and traded for other assets.

The indirect effects of a gift

For Christmas, my partner bought me Beatles CDs.  It was a very well-chosen gift. As a big Beatles fan I wanted these CDs and if she had not bought them for me, I would have gone out after Christmas and bought them for myself.  After Christmas, I bought myself a Florence & The Machine album instead.  I probably wouldn’t have bought this as well as Beatles CDs (since I implicitly ration myself on buying CDs).  So the overall effect of my partner’s gift, taking account of how my own behaviour was affected, is that I have a Florence & The Machine album I would not otherwise have had.

What then was my partner’s gift to me?  Was it Beatles CDs or was it Florence & The Machine?

It would be obtuse to say that my partner gave me a Florence & The Machine CD for Christmas. (I am not sure that she has even heard of Florence & The Machine.)  When we listen to the Beatles CDs, we both know that we are listening to her present.  I am more emotionally attached to my Beatles CDs because I know they were a gift from her.  The common sense view is that she bought me a Beatles CD, and that it arrived safely and was a very good gift.  The fact that I subsequently diverted my own spending from the Beatles to Florence and The Machine does not in any way detract from this.

A common example of this phenomenon is a wedding list. This is a custom by which a couple getting married place in a shop a list of the goods they would like to be given as wedding presents.  It is a very helpful practice because it means that the couple gets presents they actually like; it avoids duplication; and it ensures that they get gifts which fit together well (e.g. bed linen).   It also simplifies the task for their wedding guests. When we buy a gift from a wedding list we know that the chances are very high that if we don’t buy it, someone else will.  But that does not prevent us from looking through the list to find something that we would like to give.  The couple and we may remember in years to come what we bought them.  When we buy wedding presents, we don’t think that we are really buying the marginal gift on the list.

The same is true in aid. Donors can sensibly and truthfully identify the items for which their aid is being used, even if they know that , as a consequence of the aid, the recipient community is likely to change its own behaviour, including changing how it allocates its resources.  It is mischaracterization of this to say that this means that the aid has been diverted or misused.  The aid has arrived and been used exactly as intended.

That does not mean that donors should not be concerned about the indirect effects of aid, including what happens to the allocation of resources.  The indirect effects can be substantial and sometimes difficult to predict. For example, if we give food aid in kind, this may have the direct effect of providing food to a particular community, but it might have the indirect effect of undermining the livelihoods of local food growers.  There is evidence from some school feeding programmes that while the direct effect is that the schoolchildren get fed at school, in many cases the family compensates for this at home by reallocating food to less-fortunate siblings.   Rigorously analyzing and understanding all these indirect effects is an important part of understanding the overall impact of aid.   But recognizing that there are uncertain broader effects does not mean that we can no longer honestly say how our own aid has been used.

The (non-)relationship between fungibility and resource reallocation

I recently bought my partner a painting which we had seen in an art gallery in Ethiopia.  I knew that she liked this particular painting and if I had not bought it, she would have bought it herself.  Because I bought it for her, she had some money left over which she used to buy an Ethiopian ornament.  But a painting is not fungible: it is one of a kind.  Even though I gave her a non-fungible painting, she was able to adjust her own expenditure so that the overall impact is that we have an extra ornament in the house.

When we give aid, the recipient community may (consciously or unconsciously) reallocate their own resources, and this may or may not be a good thing.  But whether and how this happens is not at all related to whether the aid itself is “fungible” or “liquid”, nor whether it is provided in money or in kind.

In the example above, families were found to adapt to school feeding programmes by reallocating food within the family at home.  This would be true irrespective of whether the school feeding is financed by the government, through budget support, or directly delivered by an NGO, and irrespective of whether aid was provided in the form of money or food.  It occurs because families are able to reallocate food resources within the household.

Take another example: when countries are provided with aid to build a new road, they may choose to spend more of their own money in a different sector (e.g. on schools) or in a different part of the country.  This can be true whether the support is provided as budget support, in a sectoral programme or in the form of project aid.  Even if the road is built from imported materials by imported labourers, it is still possible for the country to use their own resources differently when they see that aid arrive.

In other words – and this is the crucial point – it is not the characteristics of the aid that determines these effects, but whether the recipient community wants and is able to reallocate its own resources in response.  Yet because our thinking has been influenced by the misleading notion of “fungibility of aid”, the policy debate tends to focus on changes to the way that aid is given as a mechanism to ensure that aid is spent as we intended.

The implications for donors, their fiduciary duties and aid effectiveness

Donors should be asking themselves two questions:

1. Has the aid we have given arrived and been used as expected? Being able to answer this is part of donors’ fiduciary responsibility to ensure that public money has been used for the purposes intended.
2. What are the wider consequences of this aid, including possible changes to how the beneficiary community uses their own resources?  Being able to answer this is important for ensuring that aid is used well, and it is part of donors’ fiduciary responsibility to ensure that aid is achieving good value for money.

A problem arises when these two distinct questions get mixed up.  We can be satisfied that our aid has arrived and been spent as we intended even if the beneficiary community has made offsetting adjustments to its own resource use.  (This is equivalent to knowing that the painting I bought for my partner really is hanging on the wall, even though I also know that the consequence of my having bought it is that she has bought an ornament.)  This interpretation of ensuring that our money has been used as intended is consistent with common sense and everyday usage, and it is consistent with donors’ fiduciary responsibilities.

This does not mean that we should be satisfied merely with knowing that our money has been spent as planned.  Donors have both a social and a fiduciary responsibility to ensure that aid achieves good value for money.  We should be interested – much more than we have been in the past – in the broader, long-run consequences of our actions.

There are many possible ways in which aid can have indirect effects which are not sufficiently taken into account in assessing whether to give aid, how to give aid, and for what purposes.  There may be effects on institutional capacity, exchange rates, relative prices, firms, domestic politics, power relationships and social attitudes.   The likelihood of domestic resource reallocation as a result of receiving aid is one among many indirect effects that ought to concern us when we provide foreign assistance, though it is probably not the most important of them.

Should we influence recipient country resource allocation, and if so, how?

In a recent blog post I have set out the reasons why I am less concerned than other people seem to be about the fact that recipient countries adjust their budget allocations in response to aid, so I will not repeat that discussion here.  Alanna Shaikh and Laura Freschi have interesting things to say on this issue as well.  I would add only that I am not naively assuming that every developing country is making rational, well-informed decisions in the interests of its poorest people; merely that it is better that they should be making resource allocation decisions than donors trying to do it from afar.

Notwithstanding my reservations, suppose that a donor decides to get involved in  resource allocation decisions in a recipient country, for example because the donor is concerned that budget allocations do not sufficiently reflect particular priorities or do not meet the needs of the poor.  For the reasons set out above, changing the way aid is given is not going to make the slightest difference to this.  It does not matter if the aid arrives as budget support, aid in kind or support to hundreds of tiny NGOs.  If the donor is concerned about resource allocation then what matters is the desire and ability of the recipient community to reallocate their own resources when the aid arrives.

In these circumstances, the donor should be considering how to improve the domestic system for resource allocation, how to make it more transparent and accountable and (if they think it appropriate) how they can be part of a dialogue with the relevant authorities about how resources are used.   If this is their goal, then giving their aid through NGOs is likely to have the opposite effect: the reallocation of aid will continue, but donors will have no locus for a dialogue about resource allocation in the budget.  If you want to influence budget allocations, you had better be giving aid to and through the budget.  Rather than try to bypass these issues, donors should engage more in improving public financial management, supporting transparency and accountability to parliament, and provide aid through government budgets so that they have a locus to influence the recipient country’s use of resources.