Hedwig Riegler - Stretching ODA by "Extending" Equity Investment, May 2023
Updated: Oct 23
About this paper
In this May 2023 guest paper, Hedwig Riegler examines proposed ODA reporting rules for equity investment and mezzanine finance. She considers grant equivalents, discount rates, comparability, transparency and the distinction between donor effort and recipient benefit.
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.
Stretching ODA by “extending” equity investment
by Hedwig Riegler
In 2016, the DAC High Level Meeting (HLM) announced that the DAC would henceforth include “private sector instruments” (PSIs) in ODA despite their nature as market-based finance. The high-level representatives may not have realised it, but this breached two firmly established principles of ODA measurement, because it required:
1) changing the ODA definition in respect of PSIs by replacing one of its definitional pillars (concessionality) by hard-to-prove additionality criteria: namely, financial, value and developmental additionality for the recipient; and
2) thereby switching the rationale of ODA measurement to recipient benefit for PSIs, whereas the measurement rationale for ODA as a whole is donor effort.
But the DAC HLM Communique 2016 also went beyond these conceptual changes by decreeing a detailed methodology for ODA accounting of PSIs. Unfortunately, that methodology was full of flaws that would inevitably lead to inconsistencies and meaningless data. The DAC’s long-drawn-out attempts to implement it have been gradually turning the “ODA metric” – a term no longer appropriate – into a “political narrative” at best.
The current proposals on how to incorporate equity investments and mezzanine finance into ODA reporting carry this dismal project even further, committing numerous statistical no nos in the process. The discussion below focuses on equity investment, but the points made apply mutatis mutandis to the equity part of the proposals on mezzanine finance.
In common parlance, one usually “makes an investment”, which carries the implication of putting money to work with the hope of not only compensating for inflation but realising a gain in real terms. But the DAC proposals discussed below instead speak of “extending an investment”, implying that some sort of gift or favour is being conferred and downplaying the idea of earning a return. However, the very methodological approach put forward contradicts this notion by proposing rules that are based on, and encourage the estimation of, expected profits.
The 2016 HLM Communiqué, despite being billed as containing mere “principles” for reporting PSI, was very specific about how to report ODA for equity investments. It offered two concrete options: ex-post reporting as the “usual” case and ex-ante reporting for specific cases where sufficient information is available upfront to estimate risks and expected profits:
Equity investments by DFIs or other vehicles in private sector entities in developing countries are counted on a grant equivalent basis ex post, i.e. they are initially counted at face value and their reflows discounted ex post, upon exit. The reflows will be discounted using differentiated discount rates by income group, applying a cap on reflows corresponding to the original investment. In specific cases and where enough information is available to estimate the reflows and risk undertaken, measurement will be on a grant equivalent basis calculated ex ante, adjusted ex post. (Section C, page 8, HLM 2016 Communiqué)
The first version [DCD/DAC/STAT(2023)3] of the currently proposed statistical rules on equities foresaw implementing this HLM provision as originally intended. For ex-post reporting (the usual case) the face value of the investment would be recorded as ODA upfront and, upon exit, the reflows (actual dividends and sale proceeds received), would be reported as negative ODA, discounted using the same discount rate as would have been used if ex-ante reporting had been selected (e.g. 12% for investments in LDCs, as given in the illustrative example in the document). Ex-ante reporting (applying only to “special cases” where expected profits are sufficiently known upfront) would instead record an estimated grant equivalent upfront and, upon exit, recalculate this figure using actual reflows and report the difference as an ex-post adjustment against the year of exit.
However, these suggestions have now been revised into an even more generous rule proposal [DCD/DAC/STAT(2023)3/REV1] that involves:
1) inflating ODA by a greater margin through an increase in the proposed discount rates from 12% to 13% for LDCs, from 10% to 10.5% for LMICs, and from 9% to 9.1% for UMICs;
2) implementing a “portfolio approach” that operates with weighted averages of multiple investments and only requires ex-post downward revisions of ex-ante figures if average discounted returns over multiple investments exceed the ex-ante estimates by more than 10%; and
3) turning the previously envisaged regular case of ex-post reporting into the exceptional, special case “for members that cannot use the ex-ante method…”. This exacerbates the intransparency inherent in the ex-ante method – first by favouring it, and second by the ruse in 2) above that it can now bundle transactions on exit so that individual investment-level details are never disclosed.
In the rest of this paper, I will try to explain how both versions of the Secretariat’s proposed methodology and both of its basic reporting approaches (ex-ante and ex-post) entail statistical flaws, and why, fundamentally, including PSIs in ODA is morally wrong.
The statistical flaws
1) Misapplied grant equivalent methodology based on misinterpretation of the meaning and function of discount rates
The discount rate used in calculations of donor effort is a representation (in per cent) of the annual loss incurred by sacrificing money in the present. It is an estimate what the loss of value might be each year through both inflation and the risk of losing all or part of one’s capital or expected returns. Any grant element calculated upfront is therefore an assumed loss, which is very likely to diverge from reality, by an amount that will depend on how close the chosen discount rate and the estimated returns came to actual developments.
Revision of grant equivalents calculated upfront by grant equivalents determined ex-post will compare the losses assumed upfront with actual receipts up to the time of exit. In such comparisons, it is absolutely nonsensical to discount the actual receipts ex-post with the discount rate applied for
the upfront calculation and reduce them to their present value in the year the investment was made, as both recent DAC documents propose. This is a no-no for several reasons. Just to take one, a large share of the discount rate applied upfront – risk – does not apply to the past. Risk is about expectations, so it applies to the future only. By the time the ex-post calculation is made, those expectations have either materialised or not, expected reflows are known either to have been received or to be outstanding, and risk has become a dead issue – as has been pointed out clearly in other contributions to this website. Correct ex-post re-calculations of grant equivalents would allow only for changes in the value of money in determining the present value of amounts received – not for an uncertainty about returns that no longer exists when the calculation is made.
Beyond this and other technical flaws, it needs to be emphasised that no serious justification has ever been offered for the basic methodology that was somehow cooked up for rubber-stamping by the 2016 HLM. If ex-post corrections of recorded equity outflows were considered a legitimate method, then why not prefer a pure ex-post treatment: claiming no ODA upfront for the equity purchase (since nothing has yet been lost) and only recording ODA if there actually should be a loss in the end – which, in the case of equity investment in the first-loss tranche of a vehicle, C-shares, could easily happen? Why not just record all equity investments upfront as Other Official Flows (so as to capture the outflows when they happen, since this is key information on finance for development) and do an ex-post ODA calculation to determine actual losses on exit by subtracting known reflows from outflows, with all flows discounted to their correct present value without any risk margins? This would be a far more plausible method for estimating any loss the provider had suffered with an investment – and far more appropriate, since the bulk of equity investments are assessed for commercial viability and are therefore not even intended to make a loss in the first place (with the possible exception of first-loss shares).
There is no case for rejecting this alternative on grounds that it would not comply with the grant equivalent concept. The majority of ODA is still grants (i.e. = 100% grant equivalents) and another parallel case can be found in long-term finance: the ODA forgiveness grant resulting from debt cancellation of non-ODA loans. In this case, the capital disbursements were previously recorded in non-ODA categories, as Other Official Flows or – the majority – Officially Supported Export Credits, and if outstanding debt on these is cancelled, the forgiven principal plus forgiven interest including late/penalty interest results in the forgiveness grant that is now recorded in full as ODA (forgiveness grant=grant equivalent). This is an ex-post approach upon bilateral agreement of debt forgiveness that calculates the balance-of-payments value of the loss the finance provider has suffered by forgiving the outstanding debt. The DAC apparently had no reservations about this as being acceptable grant-equivalent methodology when they agreed the new statistical directives on debt relief in July 2020. While inclusion of this type of debt forgiveness in ODA is wrong for other reasons (among other things, because the nominal debt is not the real loss), the principle of leaving any ODA reporting until losses manifest themselves was perfectly legitimate and could have provided a far better approach to ex-post grant equivalent reporting of losses on equities.
2) Discount rates plucked from the air
No technical reasons are given for either the exaggerated discount rates in the original proposal or their further increase in the revised proposal now up for discussion by the DAC’s statistics working party. Double risk margins (country risk margins plus “equity surcharge”) are included in both papers without any indication of what data or assumptions they were based on. The only hint as to the
rationale behind their choice is the repeatedly expressed policy goal to incentivise the assumption of risk by rewarding it. This reveals the irreconcilable contradiction between endeavours to capture reality as accurately as possible on the one hand and attempts to “massage” measurement parameters to encourage desired behaviours on the other. But besides the pretence of doing measurement while abusing statistical rules as a policy tool, the exorbitant discount rates chosen also serve to fabricate the impression of offering subsidies that, by WTO rules, should not and in reality do not exist.
3) Compromising data comparability
Offering a choice of radically different alternative reporting methods seems to have become a standard ploy in order to achieve political consensus among DAC members on statistical rules. Already the 2016 HLM Communiqué, which allowed for reporting of PSI-ODA using either the institutional or the instrument-specific method – two completely different financing streams captured at different points of measurement – opened the door for this statistical no-no.
Both the original and the revised proposals on equities follow this trend by allowing either ex-ante or ex-post reporting. The original version gave priority to the ex-post approach, whereas the latest version favours the ex-ante method, and even allows no ex-post correction to be made if the discrepancy (wrongly calculated using the present values of the year of investment) is less than 10%. But no firm rule guiding donors on which method to choose is given in either document.
The two methods do not deliver comparable data, because the ex-ante method calculates and reports grant equivalents upfront, based on assumptions, whereas the ex-post method reports the original outflow at full face value upfront, and then makes an adjustment with discounted actual reflows ex-post – maybe ten years later. Therefore, in a single reporting year, a mix of grant equivalent data (for members using the ex-ante method) and flow data (investment disbursements at face value for members using the ex-post method) will be seen, and the total for “equity investment” will be meaningless.
There is also little if any merit or benefit in the proposal to achieve nominal comparability over the long run by using ex-post figures to make adjustments to ex-ante figures, because ODA data are analysed and compared predominantly on an annual basis. The best-known and most-used statistical products focus on annual figures, often in comparison to the previous year, and only a handful of researchers take a look at long time series when comparing members’ ODA performance. Moreover, if they do, the nominal ex-post adjustments to ex-ante figures will occur years later, so comparability will immediately disappear when researchers switch from nominal figures to the constant price series that they actually need to make meaningful long-term aggregates or comparisons.
And finally, this choice of different methods for reporting leads to unequal credit and inequitable treatment of donors. For example, users of the instrument approach are allowed to cap equity reflows at the amount of the original investment, whereas users of the institutional approach need to report all reflows on their governments' equity investments in PSI-extending vehicles as negative ODA (ref. footnote 4 in both proposals).
4) Compromising transparency and data interpretability
Already in the original proposal, key data needed to verify reporting using the ex-ante approach was not going to be disclosed to the public. Now Annex A of the latest document carries this
intransparency further by promoting ex-ante reporting as the “regular” method and outlining a portfolio approach that works with weighted averages that will avoid the need to provide any useful information to the public about the progress and returns of individual investments. “Keeping in mind commercial sensitivities and legal restrictions related to engaging with the private sector” is the only justification offered. These privacy considerations can be seen as an indication that PSIs have more in common with commercial activities than with ODA projects, for which full transparency has traditionally been the rule.
Inequitable treatment of DAC members also shows in the proposed rules on data provision: while users of the ex-ante method are allowed to aggregate data by portfolio or sub-portfolio in their reporting, users of the ex-post method are required to provide the DAC Secretariat with individual-investment-level data (see paragraph 17 of the document).
More importantly, however, as regards disclosure of data to the public, there is simply no transparency on key parameters for ODA calculation (e.g. maturity, expected returns, actual returns, sales). As paragraph 24 stipulates, only aggregates across multiple investments will be disclosed under either reporting method, making it impossible for the public to follow the progress or outcomes of individual investments.
Reporting guidance on Covid vaccine donations seem to have set a precedent for approaches that lump together various transactions, thus making it impossible for statistical reporters to do any plausibility checks on aggregate data they receive from extending agencies and, even more so, for consumers of ODA data to use and interpret these data appropriately. Interpretation of ODA data by external users has become impossible on account of the excessive, unnecessary complexity and diversity of reporting rules and approaches which leaves the meaning of data beyond anybody’s grasp. In fact, there no longer is any meaning to be found due to arbitrary and illogical methodological choices.
The moral wrong
Generally, governments set up PSI-extending vehicles by giving them a financial kick-start, but then tasking them to operate on a self-sustaining basis over the longer term. The idea is that the vehicles will use their initial capital injection to raise funds and extend long-term finance that will generate profits sufficient to maintain, and even expand, their product portfolios and finance their running costs. Thus, they operate similarly to revolving funds.
There is nothing wrong with this basic model of making mutually beneficial investments in developing countries – provided private finance is not crowded out as a result.
But there is a problem with the covert philosophy of PSI-inclusion in ODA, namely its aim to make ODA programmes self-financing, at least in part. PSIs are being evaluated using exorbitant discount rates in order to simulate an ODA-reportable “loss” while, in reality, a profit is made. The unwritten rule that, in order to obtain ODA eligibility, any earnings by vehicles or funds undertaking development activities need to be re-invested in development, has laid the basis for PSI vehicles to become “ODA generators”. Their market-based finance will earn profits that they will not report as ODA receipts but instead recycle into new projects to score as fresh ODA outflows. Put bluntly, in this
way a growing share of future ODA is going to be financed by developing countries themselves through “ODA-projects” that channel unacknowledged profits back to provider countries.
To make this less obvious in the case of equities, provision has been made to hide true profits by capping actual receipts at the amount of the original investment. The reason given in the papers is simply “avoiding negative ODA” – but this is nonsense in terms of statistical measurement.
The push to score ODA where no real donor financial effort exists betrays both the original idea of providing ODA to countries of the Global South to fight poverty and gradually narrow the gap between rich and poor countries, and its corresponding rationale of measuring donor effort, i.e. the real budgetary effort by rich OECD/DAC countries for the purpose of development.
Conclusion
It is possible that the unequal and unfair treatment of donors using the different methodological choices now being proposed will resolve itself in the medium term. Data already show a clear trend towards use of the instrument-specific reporting approach: hardly any members use the institutional approach. The same might happen if a trend develops towards either the ex-ante or ex-post methods. The offer now of methodological choices may be intended just to facilitate consensus in the DAC on rule-making, the hope being that donors’ reporting will eventually converge on one methodology, so that the alternative can be dropped. However, by the time this process has run its course, ODA data on equities and other PSIs will be a royal mess beyond repair, and the needless complexities will have overwhelmed many donors’ as well as the DAC Secretariat’s administrative capacities.
In any case, regardless of whether donors coalesce around particular options, all of the currently proposed new rules are methodologically, conceptually and morally indefensible, and should be rejected.
Original post image
Comments