Hedwig Riegler & Simon Scott: PSI Vehicles as ODA Generators
Updated: Oct 23
About this paper
Hedwig Riegler and Simon Scott examine six assumptions behind including non-concessional private sector instruments in ODA, the emergence of development finance institutions as ODA generators, and proposals to restore coherence between development strategy and statistical measurement. The paper discusses the rules and negotiations as they stood in June 2023.
Editorial note: The original text follows in its historical context. References to rules, proposals, figures and future events reflect the source document, rather than an update to the present day.
PSI-extending vehicles: the perfect ODA generators?
by Hedwig Riegler and Simon Scott
First and foremost, a disclaimer: nothing said here should be construed as discrediting the positive role that development finance institutions and similar vehicles can play in fostering development by extending financial instruments at, or close to, market terms. Our point is rather that ODA and so-called private sector instruments (PSI) don’t go together, due to the special function of ODA within the universe of broad development finance. Put bluntly, PSIs are designed to be non-concessional whereas the whole purpose of ODA has always been to track concessional development finance and, specifically, “donor effort”.
What follows is in three parts. First we outline what we consider to be the main mistaken ideas behind the push to include non-concessional finance in ODA. Next we discuss how and why development finance institutions and similar vehicles offering such finance came to be seen as the ideal generators of “cost-free” ODA numbers. Finally we summarise the adverse consequences of these developments and point to the need for a fresh approach.
Six flawed assumptions behind counting ODA for non-concessional flows
How did the movement arise to insert PSI – which is really just a code-word for non-concessional development finance – into ODA? The underlying reasons for this mal-development turn out to be rather subtle. At least six interrelated bad ideas slowly crept into, and came to dominate, some DAC members’ thinking about the ODA concept:
1) “If it is developmental, it is ODA”
Whenever a developmental objective, be it only a secondary one, was identified in an activity undertaken by an official agency, the assumption became that it must somehow be scored in ODA. In this mindset, if some form of non-grant development finance was discovered to be outside the ODA umbrella in the statistical reporting directives, the DAC Secretariat’s job became to find a way of scoring a “grant equivalent” for it. ODA eligibility was reduced to showing an official origin and a developmental purpose. It was forgotten that ODA is also supposed to have a concessional nature. Behind the scenes, the proponents of the new approach understood that it really required getting rid of the concessionality pillar altogether. But they could not admit this openly, or it would have been rejected by the entire development community. So they abolished concessionality sneakily and surreptitiously, using wording calculated to obscure, perhaps even from those whose approval was required, the fact that they were fundamentally changing – and weakening – the ODA definition.
2) “ODA is the only representation of financial support for development”
Deleting ODA’s concessionality requirement effectively turned it into the “one and only” development finance. This was the endpoint of a long history of neglecting other developmental flows. The original, broad concept of development finance covering not just ODA but also Other Official Flows (i.e. non-concessional official finance for development), private finance for
development at market terms, and private concessional finance (i.e. resources from foundations and non-governmental organisations) gradually sank into oblivion in both policy and statistical debates. The global quantitative target of 1% of GNI – from which the 0.7% sub-target for ODA had been derived – was gradually erased from the minds of decision-makers. Instead they started thinking of ODA as “the only game in town”, and became obsessed with meeting the 0.7% target. Their failures to do so with real money then led them to consider more and more artificial strategies to boost their ODA numbers, culminating in the effort to squeeze non-concessional instruments into ODA under the rubric of PSI.
3) “ODA statistical rules need to provide incentives to use certain instruments”
Quite a few statistical no-nos had to be committed in order to incorporate PSI into ODA. Given that no sound technical reasons were found to legitimise these moves, advocacy focused instead on policy arguments: the standard line in methodological proposals became that “these instruments are beneficial, so they need to be reflected in ODA in order to incentivise their use”. But this turns the measurement rules themselves into a policy instrument, designed to change behaviour rather than just record it. And ODA statistical rules can only change behaviour by distorting measurement, since the only incentive they can provide to use a particular instrument is to overscore the effort it involves. This makes the strategy self-contradictory. ODA scoring only provides an incentive if ODA remains credible as a measure of donor effort; but the only incentive it can offer is to score ODA in excess of the real effort, or where there is no real effort. In the long run, such a strategy is bound to fail, though it may succeed in the short run as long as the public can be gulled into believing that ODA remains the “gold standard” of aid measurement, when in fact it has been systematically debased.
4) “Statistics is a political issue, therefore statistical rules must be made by political actors based on policy considerations“
Urging that statistical rules need to “incentivise” certain actions has always appealed to DAC members’ policy officials, who have increasingly come to dominate ODA rule development in ways that have turned it into “politics not statistics” – which was openly confirmed in a recent interview by the then DAC Chair, Susanna Moorehead. The problem is that statistical development controlled by policy people and guided by political concerns means the end of statistics – the end of facts and data aiming to reflect reality, and the beginning of fabricated results and misleading propaganda. It is for this reason that both the United Nations and OECD, in their respective recommendations on good statistical practices, laid so much emphasis on the principle of ensuring the independence of statistics from political influence.
5) “The underlying measurement rationale for ODA (“donor effort”) is interchangeable with a completely different one (“recipient benefit”)”
Even now, “donor effort” is still regularly advanced in DAC papers as the fundamental basis of ODA measurement. Indeed, the measurement rationale has to be “donor effort” measured in currency units, as this is the only way DAC members can justify including in ODA purely domestic expenditures such as those on in-donor refugee costs or “development awareness” programmes. However, in practice, they have switched to vague rationales based on “additional” recipient benefit in order to justify embarking on fanciful “grant equivalent” calculations resulting in ODA amounts for PSI transactions that bear no relation to the true donor financial effort involved. Changing the underlying
measurement rationale for one class of flows radically affects the results of measurement, and adding the donor effort in some flows to the recipient benefit in others produces meaningless totals.
6) “It’s OK to add or subtract numbers of a completely different nature”
Summing up items based on different measurement rationales is only one of several examples where the DAC has put fundamentally incompatible figures into a single equation to arrive at an ODA number. The provisional reporting rules for PSIs agreed in 2018 added the “grant equivalents” of loans to governments to the net flows of loans to the private sector, yielding an incoherent total for “ODA loans”. The proposed rules on equity investment subtract grant equivalents calculated for the year of purchase from ODA in the year of sale, which makes no more sense than taking a base 10 number away from a base 5 number. And all of DAC’s existing and proposed “caps” are artificial limits that obey a completely different logic from that which produces the figures they override. In fact, ODA now does not just mix apples and pears; it is reporting a veritable fruit salad of incompatible items, yet pretending they are all comparable.
The emergence of DFIs and similar vehicles as the perfect ODA generators
The main push by political-level delegates in OECD/DAC fora to include private sector instruments in ODA started with the Fourth High Level Meeting on Aid Effectiveness decisions to emphasize the role of the private sector in development co-operation and the 2012 DAC High Level Meeting commission to “modernise ODA statistics”, both of which were interpreted as giving the green light for bringing into ODA financial instruments that had always been excluded on account of their non-concessional character.
In a way this push was not new. For many years there had been a strong desire by political-level decision-makers to harvest ODA from activities that were:
- domestic expenses that official budgets would be covering anyway (imputed students’ costs and in-donor-country refugee costs were two prominent examples incorporated in ODA in the 1980s); or, even better,
- activities that were self-financing after an initial capital injection, and thus had the potential to produce a continuous, unending flow of cost-free ODA.
As an example, Austria’s programme became notorious for its small ODA core budget and huge shares of dubious and erratic ODA components. This resulted in largely unpredictable total ODA levels. Quite a few years ago one of us (Riegler) witnessed a top Austrian official at a closed-door DAC Peer Review session, frankly admitting that the fluctuations and unpredictability in Austria’s reported numbers were due to Austria taking a “budget-saving” approach to meeting its ODA growth targets. But even then, Austria was far from being the only DAC member taking this route to achieving respectable ODA levels.
New ways of extending the “budget-saving” approach started to open up by the early 2010s. Development finance institutions (DFIs) and similar vehicles were prime candidates for generating cost-free ODA. Their finance could usually pass as sufficiently developmental, and this aspect was played up in crafting their mandates. But concessionality was clearly an obstacle. The cost of DFIs to
governments lay only in their organisational set-up and initial capitalisation: after a few years of operation, the vehicles were usually making, rather than costing, money.
The trick, as already mentioned in point 1 above, was to abandon ODA’s concessionality requirement as surreptitiously as possible. After that hurdle was cleared, the very lack of concessionality in DFIs’ and similar vehicles’ operations became the key to making them into perfect “ODA generators”. Once their operations qualified as ODA, then as long as they remained financially viable, their business model of recycling returns to maintain and expand their portfolios offered the possibility of scoring ODA cost-free. In fact, governments could hope not only to limit their contributions to start-up capital, but even to gradually recoup their initial investment through DFIs’ dividends or other forms of repayment.
But suppressing concessionality was not, on its own, sufficient to establish DFIs as “perpetual motion” ODA machines. To be sure, it would allow DFIs’ outflows to score as ODA upfront. But since ODA had always been on a “net capital flow” basis, the corresponding reflows would then also have to be scored – as negative ODA receipts. And since the investments were not concessional, those receipts would likely outweigh the original disbursements, yielding negative ODA overall.
The solution to this problem was found in 2015-6. The 2014 High Level Meeting had opened the door to it by introducing “grant equivalent” measurement of loans. The big advantage of grant equivalents is that they only score positive figures. They are only reported if they are positive to start with, and no negative amounts are recorded later for the actual reflows, since these – discounted – have already been factored into the grant equivalent reported upfront.
Better still, grant equivalent calculations are adjustable. The number they spit out depends on the discount rate you select. Set it high enough, and you can discount the returns on any transaction – no matter how profitable – to an amount below that of the outlays, leaving a “grant equivalent” that you can score as ODA.
Accordingly, at the February 2016 High Level Meeting, not only was ODA’s fundamental concessionality requirement quietly abolished, but a huge work programme was also set out to extend “grant equivalent” reporting to a slew of non-concessional instruments offered by DFIs and similar vehicles. And the HLM even mandated that the DFIs themselves had to be consulted about what discount rates to apply, thus handing them a decisive influence over the ODA scoring of their own activities.
The aim was to stampede DAC members into approving generous methodologies to score “ODA grant equivalents” at the next Senior Level Meeting a few months later for essentially everything DFIs do. This daring move failed, as did two further years of highly politicised negotiations. At the end of 2018, however, the DAC reached a “compromise” which would allow reporting of non-concessional DFI investments as ODA on the old capital-flow basis. This was obviously a second-best solution, as it would require the returns to be recorded as negative. So it was only approved as an interim measure until grant equivalent parameters – especially the discount rates – could be agreed.
As part of the compromise, DAC members who did not want to go along with even this infringement of ODA’s concessionality principle were allowed to only report their actual contributions to DFIs as ODA, not the DFIs’ outflows. However, this set up an anomalous situation in which donors could effectively report ODA twice: once for their seed funding to set up the DFI or other vehicle (the
“institutional” reporting option) and then again from the vehicle’s outflows (the “instrument” option), since the rules allowed members to switch from one option to the other.
The DAC recognised this potential double-counting and promised to fix it as part of the delayed move to grant-equivalent reporting of all DFI activities. They are now (June 2023) debating a proposed “lock-in period” to block the switch to measuring a DFI’s “grant equivalent” outflows until the ODA that could have been reported under this method matches what has been reported as the DFI’s inflows.
This would be fine if the two methods could be expected to produce the same answer for “donor effort” over the long term, as the DAC claims. But they can’t. Net inflows to DFIs are real fiscal spending, whereas DFIs’ outflows will be massively overscored by using high discount rates and caps to hide their real returns. As explained here, the result will be that, once the switch is made, the DAC’s grant equivalent methodologies will allow DFIs and similar vehicles to generate self-financed ODA from recycled resources until the end of their existence (which may never happen, if they operate well). The term “self-financed” can in this context be looked at from either the providers’ or the recipients’ perspective: the providers will be scoring ODA “self-financed” from profits on earlier transactions, while the recipients will be “self-financing” their ODA inflows through their and other developing countries’ payments back to providers.
How to restore coherence between development strategy and statistical recording
Now, from a developmental perspective, at least some aspects of the business model of “budget-saving” development may be desirable, to the extent that they can provide a greater volume of efficacious development funding. However, the appropriate measurement and recognition of non-concessional or minimally concessional funding should have been achieved through an open debate and consultations, first with ODA recipient countries, then with the United Nations as the principal of all aid and resource flow targets, and finally with stakeholders in the broader development community. Instead, the DAC “donors’ club” unilaterally rammed through a fundamental change in the statistical concept of foreign aid.
The result is a thicket of logical and mathematical contradictions, all deriving from the nonsensical idea of scoring grant equivalents for non-concessional transactions. And along the way, the original ODA concept of a gift of resources and a transfer of wealth from rich to poor countries has gone up in smoke.
Such an undermining of the ODA concept is bound to change donor behaviour in counter-productive ways. Instead of setting realistic ODA targets at national level, and potentially renegotiating the iconic 0.7% target to reflect the rise in the number of middle-income and high-income countries, the easy way out will be to pretend to meet targets by fabricating ODA out of cost-free activities. This in turn will disincentivise the grant aid that many poor countries still need, forcing them to accept non-concessional resources that add to their debt service and other external payment obligations.
The effort to turn ODA into a “universal measure of finance for development” seems to us to have been doomed from the start. The DAC’s stubborn persistence with it has damaged the usefulness of ODA data and the broader credibility of official statistics – as well as wasted years of DAC members’ time on a fundamentally flawed and ultimately dishonest project.
The only cure seems to be to (i) restore the original role and function of ODA as official and concessional finance with development as its main objective, and to measure it correctly following this definition and (ii) simultaneously, upgrade the role and visibility of non-ODA finance for development, since only a well-coordinated and appropriately scored suite of different forms of finance from various sources will be able to provide sufficient and efficient financing for development. By contrast, continuing to try to jam all development finance into ODA by fakery and dissembling about donor effort, concessionality, grant equivalents and all the rest of it will just take the DAC further down a blind alley.
Original post image
Comments