Simon Scott: Abolishing Concessionality, May 2023
Updated: 7 days ago
About this paper
In this May 2023 guest contribution, Simon Scott examines the role of concessionality—the gift or subsidy element—in Official Development Assistance (ODA). He argues that changes agreed by the OECD Development Assistance Committee in 2016 opened the way for non-concessional private sector instruments to be counted as aid.
The paper traces the distinction between concessionality and additionality, then considers grant equivalents, discount rates, debt relief, guarantees and equity investment. It asks what these changes mean for ODA as a measure of donor effort.
Editorial note: The full original paper follows. Its arguments, judgments and references to proposals and rules are retained in their May 2023 historical context. Section headings have been added to help readers navigate the web version.
Click twice on the link below to see the original pdf:
Related reading: https://www.odareform.org/post/simon-scott-statistical-no-nos
Abolishing Concessionality
By Simon Scott
Concessionality and the 2016 DAC decision
The most consequential decision the DAC ever made on ODA accounting was to remove the requirement that it offer some element of gift or subsidy. They took this decision 19 February 2016, but in wording so tricky and convoluted that one wonders if they realised what they were doing.
Concessionality means giving something away. It’s essential to the idea of foreign aid. In fact, foreign aid only exists because a gift of resources may be needed where trade, lending or investment on free-market terms cannot get some important development job done.
Measuring government foreign aid as Official Development Assistance (ODA) started in the late 1960s. In 1967, developing countries launched a campaign for “a separate minimum target…for the official component of aid flows”, leading to the UN’s 0.7 per cent ODA/GNI target in 1970. The DAC was already collecting figures on all official flows, so all it had to do to measure progress on the target was to define which of them qualified as ODA or aid.
It was clear at the time that ODA had to include the notion of “giving” if it was to meet the public’s idea of aid, make the UN ODA target meaningful, and allow the DAC itself to assess “donor effort” and “burden sharing”. And for the next 47 years, it was always understood that, in the words of the DAC’s own ODA definition, “each transaction” of ODA had to be “concessional in character”.
But a funny thing happened on 19 February 2016. Without any warning to the UN or the global development community, an obscure footnote in a long technical annex to the Communique from a DAC High Level Meeting (HLM) announced in small, faint type that:
The Reporting Directives will be updated to take into account the fact that the
“concessional in character” criterion is not appropriate for assessing the ODA characteristics of PSI [private sector instruments].
This wording seems to have been deliberately designed to minimise and obscure what was really being done. It was not even clear that it embodied a decision at all, rather than just the “taking into account” of a “fact”. The only further explanation was this piece of gobbledegook:
Under the instrument-specific approach, the measurement of donor effort will be based on the system of risk-adjusted grant equivalents. However, the system will be adjusted for the fact that i) PSI are non-concessional in nature and that application of the concessionality thresholds agreed in the sovereign loan context in December 2014 could incentivise unnecessary subsidisation of finance; and ii) financing the private sector is generally riskier than the official sector, in principle necessitating a risk premium in the discount rate additional to the already agreed sovereign risk premia, based on evidence and with due consideration to not inflate ODA.
Notice again how “the system” is merely being “adjusted” for a “fact”. And the adjustment itself is presented as something that “will” happen, as if it is inevitable or already agreed.
One wonders whether a single minister or other high representative at the 2016 DAC HLM really understood the full implications of all this convoluted verbiage. Reading it, they can only have gathered that there was some special small category of “private sector instruments” which for some valid but mysterious reason could score as ODA despite being “non-concessional in nature”.
What are private sector instruments?
But what were these “private sector instruments”? From the name, and from the way the annex dedicated to them started out by trumpeting
the private sector’s key role going forward in providing the ideas, the ingenuity and the entrepreneurial energy required to find new, innovative solutions to sustainable development challenges in line with the transformative vision embodied in the Sustainable Development Goals and the consequent need for the development assistance community to engage the private sector in the 2030 agenda through Vehicles dedicated to private sector [sic]
a minister might have gathered that private sector instruments were mechanisms for engaging the private sector in development. But this is not the case. It is completely irrelevant whether the private sector is targeted – for, as another faintly typed footnote concedes,
Such instruments may also be extended to public or public-private entities.
So all the blather about the private sector’s key role, ingenuity, entrepreneurial energy, innovative solutions and so on was actually irrelevant to the definition of “private sector instruments”. What then, are they? Parsing the annex very carefully, one can piece together that they are
a range of financial instruments and arrangements (including equity, credit enhancements and guarantees)
used by
special public sector vehicles
that
have similar objectives and working methods to those of private sector operators.
Okay… but again, what does it all mean? The key to deciphering it was to focus on the “objectives” of “private sector operators”. Such operators are motivated by self-interest; their objective is to make transactions that are to their own benefit. That’s the opposite of concessionality, and in reality “PSI” is a code-word to cover official use of any non-concessional financial instrument. And inserting such instruments into ODA abolishes its concessional nature, the basis for its creation.
Whether they realized it or not, the DAC had set itself on a new course. Ever since 2016 it has effectively been using PSI to smuggle non-concessional instruments into a measure whose raison d’être is its concessionality.
The theoretical justification for this has been to claim that “additionality” can substitute for concessionality. The line has been that if some official finance is provided for a development activity for which private finance is not available, or stimulates the missing private finance, or affords a recipient some non-financial advantage, then ODA should be scored.
Additionality, grant equivalents and donor effort
But “additionality” won’t fly. First, it is clearly very different from concessionality, and no serious reason has ever been advanced to regard it as the qualifying “characteristic of operations for PSI to be in line with the concessional in character criterion of ODA”. Additionality is also defined in terms of recipient benefit, whereas concessionality, and ODA in general, are measures of donor effort.
On a practical level, the main method the DAC has used to claim ODA on non-concessional transactions has been through abuse of grant equivalent calculations. True grant equivalents are nothing but a measure of concessionality: they are simply estimates of subsidy or gift value. But their calculation requires the selection of a discount rate, and this effectively acts as a “slider” that allows one to produce any answer at will. Using an excessive discount rate will make a transaction show a “grant equivalent” above its true gift value, and this is what the DAC does.
The DAC has been using inflated discount rates to score ODA on loans to governments since 2018. Here, the rates work by “over-discounting” repayment obligations to give the illusion that loans are more generous than they really are. This allows DAC members to score inflated grant equivalents on all loans, and to score some ODA even on non-concessional loans.
But a grosser misuse of grant equivalent methodology came with the DAC’s 2020 rules on Reporting on Debt Relief in the Grant Equivalent System, which massively over-score the value of debt relief. Here the discount rates work in reverse: they inflate the value of the loan both as it stood originally, and after relief, and then score ODA for the inflated difference between the two. The discount rates include risk margins even when inflating past values, which is statistical nonsense.
The mathematical “horror story” of the debt relief rules conceals the basic fact that debt relief involves no “donor cost” that has not already been accounted for, either through ODA scoring, or in cash. If the loan was ODA, the risk was scored upfront in its ODA “grant equivalent”. If it was a non-ODA loan, the lender or borrower has typically paid extra to insure or guarantee repayments, and it is normally from these funds that the official sector will meet any claims in case of default.
Further abuse of grant equivalent methodology is now planned by counting official loan guarantees themselves as ODA, even where these are offered at market prices, and even though the agencies selling them pocket fees on every guarantee and only rarely have to pay out for default. So donors will soon be scoring of ODA on transactions on which they make a profit almost every time.
The only logical end-point of this process is for the DAC to find a way of scoring grant-equivalent ODA for official involvement in every type of non-concessional financial instrument. It has already tried to invent grant equivalents for equity investment, and if it now goes ahead with scoring loan guarantees as ODA, why not move on to investment guarantees, credit derivatives, bond purchases, or any other non-concessional instruments that can be connected with developing countries?
Unless this process is stopped, ODA will lose all traces of concessionality, and will cease to have any meaning as a measure of donors’ aid efforts. Would developing countries then demand that everyone go back to the drawing board, and start the whole process of establishing a “separate minimum target” for “the official component of aid flows” all over again? Only time will tell.
Original document image
Comments