Hedwig Riegler

TOSSD – envisaged as a new statistical measure to broaden both the range of financing for development and the range of providers reporting their contributions – inspired high hopes that it might add real value to the world of quantitative information on resource provision to developing countries. TOSSD data are intended to complement ODA1, and in this role as a complementary measure TOSSD could indeed enrich the picture of development finance. ODA has often been mistaken as “the one and only” finance for development: setting the 0.7% ODA/GNI target in 1970 may have produced the erroneous general perception that it was the only relevant development finance. The resulting neglect – including by OECD/DAC – of ample and adequate communication on other resources provided for development have led non-ODA statistics to fall practically into oblivion.

Given this context, TOSSD – if designed as a truly complementary measure (in the mathematical sense) – could provide major additional benefits to the existing world of data on development finance. The decision to develop this new measure jointly in a task force that includes representatives of all stakeholders, not just providers, is also innovative and highly commendable.  It provides TOSSD with a platform that can both tailor the measure to a broad range of needs, and promote its widespread political acceptance.

However, the present state of evolution of the measure raises some serious concerns. The TOSSD Reporting Instructions (5 June 2019 edition; Preamble, para. 6) suggest the creation of a new parallel reporting system with large unclarified overlaps with the existing OECD/DAC system on both ODA and non-ODA resource flows. Unless the TOSSD reporting arrangements are streamlined and rationalised, the system will be in danger of not being populated sufficiently with reliable data and consequently of going largely unused. There are also several technical aspects of the concept which lack solidity and require substantial further work. As it claims to be a “statistical measure”, TOSSD should live up to this name – and not be framed as a political narrative with some illustrative figures that do not meet statistical standards.

Unfortunately, TOSSD development is proceeding in the context of an increasingly urgent credibility crisis regarding development finance information, above all in OECD/DAC statistics.  Several recent decisions on ODA measurement, or in some cases non-decisions (e.g. on how to deal with PSI – private sector instruments), have weakened the meaningfulness, robustness, and credibility of data produced by the DAC. For the first time, recently published 2018 DAC ODA data merge items of contradictory nature into the headline ODA figure (e.g. flows + grant equivalents; PSI data based on the institutional approach + PSI data based on the instrument-specific approach). See the note by Simon Scott on recent developments of ODA measurement here and his subsequent exchange with the DAC Chair here. The bodies working on developing the TOSSD measure should not go down the same road.

The following sections outline the issues and offer recommendations to address specific problem areas, especially overlaps and differences between TOSSD and ODA. These should be taken as examples; no claim is made of completeness. However, they aim to tackle the most important points: those which, unless they are addressed and clarified, will leave the way open to differing interpretations, which in turn will lead to inconsistent reporting and thus to meaningless totals.

On the concepts of development embodied in TOSSD and ODA

“The Total Official Support for Sustainable Development (TOSSD) statistical measure includes all officially-supported resource flows to promote sustainable development in developing countries and to support development enablers and/or address global challenges at regional or global levels.” [Instructions, para. 8]

While the purpose of TOSSD transactions is “to promote sustainable development”, the “main objective” of ODA transactions is “the promotion of the economic development and welfare of developing countries”. So the first question that needs to be answered is whether there is any difference between the sets of transactions implied by these two concepts of development.

On the surface it would seem that “sustainable development” may require different criteria than “economic development and welfare” – especially in relation to environmental concerns.  For example, one could imagine economic development and welfare including activities that brought short-term benefits to humans, but were unsustainable, or damaged the environment, in the long-term. On the other hand, the sustainability of ODA actions has been a policy objective and quality requirement for decades and, recently, alignment of ODA programmes with SDGs has become an important additional goal. This suggests that what is eligible as ODA should also meet the definition of TOSSD. 

In practice, it appears that TOSSD’s definition of “sustainable development” has been drawn so broadly as to include practically everything that meets the developmental test of ODA.  TOSSD’s “Eligibility Criteria Regarding Sustainable Development” [Instructions, para. 47 to 49] are, upon analysis, extremely permissive: they appear designed to encourage reporters to include within TOSSD any even vaguely developmental activity, however tenuous its link to sustainability or the Sustainable Development Goals might be – which is another indication that ODA would fully fit under the TOSSD umbrella.

If this interpretation is correct, then the definitional differences between TOSSD and ODA concerning their concept of development vanish.  But in that case, one might ask what is to be gained by having definitions which appear different on the surface.  In any case, clarification of the relation between the two definitions is imperative to avoid confusion and ensure consistent reporting.

On the coverage overlap between TOSSD and existing DAC resource flow data

According to the overall definition of TOSSD, ODA is included fully or to a major part in TOSSD: ODA covers all official concessional resource provision to developing countries but excludes support to development enablers, support to addressing global challenges, and private flows mobilised by official interventions – all of which are included in TOSSD.

However, much of what is excluded from ODA is still included elsewhere in DAC Statistics – under non-ODA categories of resource flows to developing countries, i.e. non-concessional official (OOF), officially supported export credits, and concessional (NGO) and non-concessional (PRIV) private flows. This leaves ODA and other DAC resource flow data in a highly complex and to some extent ambiguous relationship with TOSSD. Some OOF and PRIV transactions may enter TOSSD, but TOSSD may also include some “development enablers” or “support to addressing global challenges” that do not meet the OECD/DAC definition of resource flows for development and thus appear nowhere in the OECD/DAC statistical universe. At the same time, TOSSD will exclude the vast majority of private flows reportable in the OECD/DAC system. Indeed the only private flows included in TOSSD are those mobilised through official interventions.  This leaves out of account all other private loans and investments, and all NGO and foundation flows financed by their own resources and in support of their own programmes. To be sure, this is largely consistent with the TOSSD concept, since the “O” stands for “official” – although this word must be understood in a sense that allows private flows to be included, as long as a link to official action is demonstrated.

Nevertheless, and despite widely overlapping sets, TOSSD and OECD/DAC total resource flow data clearly aim at fundamentally different coverages of developmental resource flows.  This raises pivotal questions about whether and how the existing OECD/DAC and TOSSD data systems will relate to each other, both in conceptual and practical terms. Are they to be run in parallel on a completely separate basis or to collaborate and exchange data? If the latter, in what format, on what timetable, and will reporting be differentiated between DAC and non-DAC providers?  Will TOSSD harvest data from OECD/DAC (hence harmonising definitions, classifications and procedures where necessary)? The necessity and cost-benefit of cutting across established international concepts of these resource flows, reflected in the DAC categories and widely shared across the international system, would need to be established in order to convince reporters to the OECD/DAC system to make the considerable extra effort required for TOSSD reporting – it being evident that, without the inputs of OECD/DAC reporters, TOSSD coverage rates are unlikely to come even near required levels.

On TOSSD and ODA concepts of concessionality

Both TOSSD and ODA/DAC Statistics distinguish between concessional and non-concessional resource provision, however each in a different way:


For a TOSSD loan to be classed as “concessional” it needs to have a grant element of at least 35%, after applying a 5% p.a. discount rate to the repayment stream.


For a loan extended to the official sector to be eligible as ODA, it needs to meet the following concessionality criteria2:


All four of the above formulas for assessing concessionality are much less demanding than the TOSSD test of 35% concessionality using a 5% discount rate. ODA can thus include many loans to governments that under the TOSSD definition of concessionality will be considered “non-concessional” – even though, by definition, ODA should exclude non-concessional loans.

The gap between TOSSD and ODA treatment of the concessionality of loans to the private sector in developing countries is even wider. Here the ODA test only requires a 25% grant element after discounting the repayment stream at 10% p.a.  Note, however, that for whichever loans to the private sector qualify under each test, both TOSSD and the “headline” ODA measure record the actual flows on the loan, whereas for loans to the official sector, TOSSD records actual flows while the ODA “headline” measure records grant equivalents.

The wide gaps and procedural discrepancies between the TOSSD and ODA tests of concessionality risk creating confusion and a loss of transparency, since users will no longer be presented with consistent information on concessional flows. It is therefore imperative to sort out these differences or, if that fails, to consider changing the terminology (replace “concessional” by another term).

On the basis of measurement

For both TOSSD and ODA, questions now need to be answered about the very nature of the quantities they are aiming to measure. In the case of ODA, ambiguities regarding that question – of whether ODA is to measure donor budgetary effort, or inflows to beneficiary countries, or recipient benefit – have, inter alia, led to confusion, methodological dispute, and in the end to pending or unsatisfactory solutions for reporting (as explained in detail by Simon Scott in his above-mentioned blog). In order for TOSSD to avoid such an entangled situation, reporting instructions need to make clear beyond doubt what exactly TOSSD intends to capture. Definition of the point of measurement is important, but on its own it will not do the job. A clear definition of what quantity TOSSD is measuring will impact on both the choice of point of measurement and the type of financial volume to capture (e.g. financial commitments, cash-flow, or cash-flow in combination with contingent liabilities). This is essential for ensuring clarity on what a headline total figure represents.

More specifically, while TOSSD is designed as a measure based on resource flows, it includes guarantees and other contingent liabilities that stimulate flows but which do not themselves constitute a flow [Instructions, para. 107]. This seems to imply that a headline figure for TOSSD would mix flow and non-flow data, leading to incoherence in the statistical quantity being measured, and an overstatement of the total resources provided. This pitfall could be avoided either by relegating the non-flow elements to memorandum items, or by replacing the term “resource flow” by some broader term such as resource engagement or provision, and clearly stating what this term encompasses.3

On Sustainable Development Goal monitoring

A notable feature of the current TOSSD instructions is their emphasis on linking – one might almost say, chaining – TOSSD to the 2030 Agenda and its Sustainable Development Goals.  The Instructions mention the SDGs more than 30 times, they rely on them for TOSSD’s definition of sustainability (para. 47), and they even state that “As and when the 2030 Agenda is concluded and replaced by another framework, the TOSSD measure will be updated to link to that framework.”

Perhaps most remarkable is the obligation placed on statistical reporters to specify the SDG target(s) to which each individual TOSSD transaction is directed [Instructions, paras 37 and 87]. Up to ten targets can be specified for a single activity.  This effectively sets up a new system for monitoring flows in support of the SDGs, rivalling the official UN SDG indicator framework, which includes numerous indicators that use the traditional categories of ODA, OOF, and private flows including foreign direct investment (FDI).

So far, it appears that TOSSD is only being proposed to replace, or partially replace, the existing SDG indicator of total flows mobilised for development (SDG target 17.3).  But once TOSSD data, tied to individual SDG goals and targets, are regularly produced, there would appear to be considerable scope for confusion between those figures and all the other specific official SDG indicators that report on resource flows. Therefore, it will be necessary for TOSSD to also clarify its relationship to the official SDG indicator framework (see further suggestions on this below).

Conclusion: Rescuing TOSSD from the credibility crisis in OECD/DAC statistics

Taken as a whole, the TOSSD structure as now proposed establishes a parallel system for counting developmental resource flows, which risks undermining confidence in the existing OECD/DAC measurement structure of ODA+OOF+Private (Official Development Assistance + Other Official Flows + private charitable and market-term flows). Unless an approach is found to make the two systems truly complementary both conceptually and mathematically, and to harness the existing OECD/DAC system (i.e. definitions, reporting formats, data available, etc.) for TOSSD reporting, both universes will risk confidence erosion and reporting fatigue. This danger is particularly acute at the present time, when the credibility of ODA as a measure of aid effort has been undermined by a series of half-baked and defective rule changes.

A further challenge is posed by the highly fragmented design of TOSSD, with its many sub-categories of different nature and meaning, and many overlaps with the broader resource flow data collected by OECD/DAC since the 1960s. It will be difficult, even for those involved in reporting – let alone a general audience – to clearly understand what individual sub-totals or the TOSSD total figure are intended to signify. Without clarity on the meaning of its main figures, TOSSD will not be able to fulfil its purpose as a statistical measure: to inform political decision-making and encourage an interested public audience to participate knowledgeably in policy debates on financing for development.

Meeting these challenges will require considerable further work.  The main lines of such work have been outlined in each section above, but the following section briefly recaps them and adds some specific further suggestions.

Some final specific recommendations and food for thought

Resolving the currently ambiguous concepts for collecting and presenting data on TOSSD – and clarifying TOSSD’s relation to the existing OECD/DAC data collection on ODA and other resource flows – will require enhanced reporting instructions that offer clarifications on the following points:

  • What precisely is the universal set that constitutes TOSSD? There is a need to sharpen the overall TOSSD definition and spell out the specific ways in which its definition of sustainable development is either broader or narrower than the developmental test of ODA. The underlying measurement purpose (as explained above) needs to be spelled out clearly so as to enable reporters to understand precisely what financial volumes to include and data users to understand the meaning of these volume figures.
  • What are the key TOSSD subsets on which data are to be collected and presented?
  • What is the detailed relationship between TOSSD and ODA items? The “differences” between TOSSD and ODA addressed in a table in Annex G of the TOSSD Reporting Instructions are too general and imprecise to be reliably interpreted as to what is intended with each measurement, whereas this is an indispensable prerequisite for both consistent, comparable reporting and meaningful use of data.  For example:​​

– under the heading “Scope of flows covered”, the only information conveyed is “TOSSD: Concessional and non-concessional; ODA: concessional” – no mention is made of differences in the concept of “concessional” and that the classification in TOSSD is based on much stricter criteria than ODA concessionality.
– under the heading “Measurement”, it is stated “TOSSD: cash flow; ODA: grant equivalent” – which is imprecise as the current (2018) headline ODA measure is a mix of grant equivalent and cash flow data.

Clarifications, more accurate information, and perhaps changing the term “resource flows” – while clearly explaining the meaning of any new term – could enhance the TOSSD reporting concept considerably on these questions.

  • More broadly, information needs to be provided on how the TOSSD data set relates to the ODA data set: Is it intended as a complementary set (in the mathematical sense, based on the claim of complementarity)? Or is it intended as a specific sub-set (given the definitions of TOSSD which suggest that ODA is included in TOSSD)? Or is there a partial overlap of the two sets? All these questions concerning the relationship of ODA and TOSSD call for precise answers, since a majority of reporters will be faced with both ODA and TOSSD reporting and any ambiguity about the structure and relationship of the two is very likely to lead to confusion, and to undermine reporting morale.
  • Reporting arrangements also need to be clarified, and made as simple and straightforward as possible.  Clarification is needed on how data collected through other systems, above all the OECD/DAC collection on total resource flows to developing countries, are going to be used/made available in order to avoid duplication of efforts and inconsistencies that result from parallel reporting.

Whether or not clarifications of this kind can be provided will impact decisively on the quality and usability of the data collection in either a positive or negative sense.

Further clarification is also needed on how TOSSD figures are intended to be incorporated in the current set of SDG indicators. Although the reporting instructions explain that the TOSSD measurement is “…. expected to serve for monitoring this goal [SDG 17] and several other SDG targets”, they do not say how TOSSD figures will feed into and be coordinated with the existing set of SDG indicators. A next iteration of the Instructions should fill this gap, while recognising that the selection of global SDG indicators is under the responsibility of the United Nations.

To establish mathematical complementarity with ODA and thus avoid double-counting, items collected in ODA would need to be either removed from, or clearly flagged within, the TOSSD collection. For instance, in-donor-country costs for refugees is one item where TOSSD and ODA reporting overlap.  ODA reporting is limited to the first twelve months of refugees’ stay, but TOSSD reporting imposes no such time limit. This mingling of cost elements in TOSSD reporting undermines clarity both in reporting and interpretation of the figures produced.[1] And this appears to be just one of the examples where ODA and TOSSD overlap and where the TOSSD figure would include both ODA-eligible and non-ODA-eligible parts; other items that apparently overlap include peace and security activities and technical cooperation.

In addition, treatment of the overlap with other, non-ODA, resource flows collected by OECD/DAC since the 1960s (non-export credit Other Official Flows, officially-supported export credits, private grants from NGOs and foundations, private flows at market terms, and the recently developed identification of private flows mobilised through official intervention) should be addressed clearly in further iterations of the TOSSD Reporting Instructions.

A final specific suggestion relates to the pending decisions on how PSI (private sector instruments for development) should be reflected in ODA. Here, the stalemate situation that led to the current, unsatisfactory provisional reporting instructions could be resolved by envisaging a solution that only counts the capitalisation of PSI-extending agencies in ODA, but the outflows from the various instruments in TOSSD. A clear-cut separation along this line could remedy the present meaningless jumble of incommensurable data within ODA, with some donors reporting their net contributions to PSI-extending agencies and others the net outflows of those agencies.  This particular disorder in PSI reporting could be rectified if TOSSD reported the flows to developing countries while ODA reported the net contributions to these flows made by provider countries. TOSSD could thus even play a significant role in contributing to a recovery of ODA reporting from its current ailments.

[1] TOSSD Reporting Instructions, edition 5 June 2019, Preamble, para. 6

[2] LDCs, oLICS, LMICS and UMICs are groupings of developing countries based mainly on per capita income. Grant elements are the percentages of loan amounts that remain after subtracting their discounted repayment streams. For further explanations see here.

[3]Such an approach could draw inspiration from an element in the recent discussions on modernising the ODA measure, where proposals to include some eligible portion of extending a guarantee were turned down on the ground that ODA was a (net) flow measure and thus could not account for guarantees and contingent liabilities at the point of their extension, but only at the point when they were called and a government had to pay for a loss incurred, resulting in a flow. This recognition of a valid statistical principle (not to mix chalk and cheese) in the ODA debate seems to have been overlooked so far in the TOSSD concept. In fairness, however, one must also acknowledge that other recent steps in ODA modernisation have also flouted its former principle of not merging items of very different nature into a total, since for the first time in ODA reporting (on 2018 data) grant equivalent and flow data are mingled to produce the headline ODA figure that measures performance against the 0.7 % ODA target. Generally, recent developments in the measurement of resource provision to developing countries have moved away more and more from basic statistical principles, changing the nature of data produced from “statistics” to “politically desirable results”.  However, the fact that both ODA and TOSSD have been going down this road does not make it any more defensible.

[4]Page 38 of the Instructions states: “Use the taxonomy of modalities of co-operation to distinguish between support during the first 12 months and support beyond that period”, yet the taxonomy of modalities provided in paragraph 95 (pp. 23-4), lists only one category of “Support to refugees/protected persons”.

Note on the author:

Hedwig Riegler was Austria’s Statistical Correspondent for OECD/DAC from 1996 to 2011 and in this role responsible for collecting, compiling and finalising Austria’s national development finance statistics for submission to the DAC. In this function she was a member of the OECD/DAC’s statistical body, the Working Party on Development Finance Statistics, for which she served as Vice Chair from 2005 to 2009 and as Chair from 2009 to 2013. She has been doing consultancy work for Austria’s authorities involved in OECD/DAC reporting and UNFCCC and EU climate finance reporting and participates in debates on the quantitative aspects of development cooperation and climate finance. She contributed a number of articles to the annual publication of Austria’s main think tank on development policies, the Austrian Foundation on Development Research (OEFSE).

In der Rubrik „Kommentar der Anderen“ bietet die AG Globale Verantwortung ExpertInnen die Möglichkeit, aktuelle und relevante entwicklungspolitische Themen zu kommentieren sowie ihre Meinung zu präsentieren. Das Ziel ist, Debatten über Entwicklungspolitik zu ermöglichen, den demokratischen Diskurs zu fördern und die Bedeutung der Umsetzung der Agenda 2030 hervorzuheben. Die inhaltliche Verantwortung für den Text liegt ausschließlich bei den Autoren. Die AG Globale Verantwortung teilt nicht notwendigerweise die vorgetragenen Ansichten.