About two weeks ago the International Aid Transparency Initiative (IATI) managed a significant breakthrough when a group of supporting donors finally agreed to a common standard for aid information reporting.
For those of you that are unfamiliar with IATI, it is a revolutionary attempt to increase the quality, quantity, frequency and comparability of data on international aid commitments and disbursements. A common standard means that data from multiple countries donors can be aggregated with ease. Some day, it could just take a few clicks to learn how much aid is being spent on infrastructure in all of sub-Saharan Africa or which donors operate in a very specific geographic region.
No recipient governments have officially signed up to IATI. This has little to do with official support: IATI is solely an initiative aimed at donor behaviour, and its proponents are quick to point out that there are little to no obligations for recipient governments (a handful of recipient governments have declared their support for the initiative).
Yet the adoption of a single IATI standard will have major implications for how recipient governments trade information with donors, and some of them will not be immediately beneficial:
Aid-dependent governments which are at least marginally competent make an effort to track aid spending, even that which is not entirely under their control. Tracking and analysis requires classification, and many countries have, in order to ascertain how donor aid aligns to their own objectives, create reporting mechanisms which help them do this.
For example, about five years ago Malawi created a national development plan, known as the Malawi Growth and Development Strategy (MDGS). Despite managing to be hopelessly vague, technical, and vague all at once, the MDGS was Malawi's home-grown plan, one that did not directly mimic the aspirations of donors. Such plans, which are becoming more prevalent, are crucial for recipient governments to assert their independence, and offer an opportunity for donors to honor their obligations under the Paris Declaration, which places weight on country ownership of the development discourse, as well as the responsibility of donors to align their own practices to that of the recipient's.
When worked as a civil servant in Malawi, I watched the Ministry of Finance becomes more and more adept at disaggregating development expenditure using the MDGS to generate appropriate categories. Such processes are crucial for determining whether or not a national budget is compatible with the country's objectives - they also required a huge amount of timely reporting from the donors.
So one of the ways donors can fulfill their responsibilities to promote country ownership and alignment is to report their aid commitments and disbursements using existing country systems. Yet donors still struggle to meet these meager expectations: aid data is often dropped in the recipient government's lap in a format which still needs to be interpreted and converted to match the needs of the country, an exercise which just wastes the time of already capacity-strapped governments. Still, some donors are getting better, due to pressures from more-capable recipients and their obligations under the Paris Declaration.
IATI, if implemented as a single reporting standard, could possibly undermine this marginal improvement. In theory this shouldn't be so, as an IATI standard only creates obligations for donors to report up. Yet given donors limited time and patience for adhering to country requirements (i.e. reporting their aid expenditure in MDGS terms), a large-scale adoption of a new standard is likely to take precedence over any attempts to focus on in-country reporting standards. This doesn't mean that recipient countries are going to have less information about aid on their hands (in fact, they'll likely have a lot, lot more), but that the information being provided by donors is going to possibly be less useful.
Problems are also generated by a single standard when information flows in the opposite direction: in order ot fufill their obligations on IATI, donors will not only have strong incentives to just use that standard when reporting to recipients, they will also certainly require recipient governments to use that standard when reporting aid spent on government-run projects. This is a little bit like a donor moving to a non-English speaking country, never sending the recipient government any information in the native toungue, then demanding that any phone calls made to the donor office are made in English (analogies aside, don't donors do this anyway?).
Many would consider the estimated benefits of IATI to vastly outweigh possible negative effects on country ownership, alignment and reporting habits. Owen Barder makes a compelling case for these benefits in his blog. Many of the assumptions rely on the ability of civil society to use transparency to better hold donors and governments to account. However, such accountability can only be strengthened by information that is defined by country-led debates. The MDGS, for example, has been talked about and debated in Malawian civil society for years, and so aid information which trickles into this debate is going to be inherently less useful if that data can't easily be disaggregated in a manner people are accustomed to.
Make no mistake: the push for international aid transparency is a good thing - an incredibly good thing - and a single reporting standard is probably necessary for any coherent system to be embraced. However, we need to be cautious that the crusade for top-down transparency complements, not sidelines, the slog for country ownership.
Update: Owen Barder comments below, suggesting that I've got it all wrong (which is probable, given that he's lived and breathed IATI for the past few years):
The IATI standard explicitly includes a requirement to report the information classified according to the country’s own systems. (The details of how this will be done are yet to be worked out, but there is no stakeholder in this process – donor, developing country or civil society – who doubts that this is centrally important to whether IATI is useful.IATI is not a once size fits all top down standard in the way you imply. Some aspects of the data will be universal (eg dates of disbursements) but much of the data recorded about each aid item will depend on the kind of spending it is and on the country where the money is being spent. So the data for aid to Rwanda will be classified (among other things) according to Rwanda’s system, data for aid to Malawi will be classified (among other things) according to Malawi’s systems.
10 Comments
Matt
Eh? The IATI standard explicitly includes a requirement to report the information classified according to the country's own systems. (The details of how this will be done are yet to be worked out, but there is no stakeholder in this process - donor, developing country or civil society - who doubts that this is centrally important to whether IATI is useful.
IATI is not a once size fits all top down standard in the way you imply. Some aspects of the data will be universal (eg dates of disbursements) but much of the data recorded about each aid item will depend on the kind of spending it is and on the country where the money is being spent. So the data for aid to Rwanda will be classified (among other things) according to Rwanda's system, data for aid to Malawi will be classified (among other things) according to Malawi's systems.
The beauty of IATI is that there will be a single source of this information, containing all the richness of detail to do both inter-national comparisons and to meet intra-country needs for information.
Owen
Hi Matt,
Notwithstanding Owen's comments which are probably far more to the point than mine, from a systems perspective the advantage of a single data classification is that you only have to work out one mapping from your system internal definitions to/from the data standard. Thus whilst in the now the Malawian govt may have a bit more work to do to fit with the IATI standard (or maybe not from Owen's comments), they shouldn't have to repeat that for each separate donor. That bit of pain is usually worth the gain.
MJ
Hi MJ, nnGood point - the need to avoid having to report to several donors is one of the large expected cost-saving benefits from IATI, one of the results from the cost-benefit analysis that several of us worked on for IATI. That said, I believe that the burden of `translation' should be on donors, not recipient governments, to figure out how on-budget aid spending matches their standards. In (my) ideal world, recipient governments would only worry about their own country systems, and donors would, at their own expense, have to figure out how those systems translate into the language they want to use. nnI should point out that I'm still convinced the net benefits are positive, but just trying to get us thinking about possible negatives, which exist for *every* new initiatve under the sun.
Thanks Owen - I've update the post to include your response. I'm glad to hear that my concerns are driven by ignorance. Could you point me towards the document that details these committments? The IATI standards website is a bit difficult to approach in this regard. Other documents clearly show that there is some concern here:
"meet in full the information needs of developing country government AIMS and budgets without imposing a burden on developing countries, including complying with local definitions and classifications"
I understand the call for a link to country systems, and maybe that is enough, but I suppose it remains to be seen how well donors comply with this, and whether or not recipient governments are allowed to lead with this classification (rather than donors doing this before passing down the information on their own).
There is also still the issue that the IATI standard is comprised of other things outside the country classification, some of the work of having to conform to this will inevitably trickle down to recipient governments. We'd hope that these would be offset (by reducing reporting requirements, etc, as the estimates that our friend Asma uncovered), but they are still worth noting.
I wasn't suggesting that those involved in IATI don't care about these things, but as with most new initiatives, I'll be more convinced when I see these things in action. Otherwise, points taken.
Cheers,
Matt
Owen - I appreciate that you say 'the details of how this will be done are yet to be worked out', but any indications on how this will work?
In some countries, data is divided according to sectors and sub-sectors. These sectors and sub-sectors depend on the local conceptions of the economy and society (what Tanzania has as sectors is very different to what Malawi has, since they have very different local development plans); if we want to generate any analytics on sectors across countries, either the IATI people or the local data collectors will have to codify every activity line twice. In Burkina Faso, they already do this (they produce a report using local sector conceptions and the OECD-DAC sector conceptions), and it's a lot of work.
I'm very much in favour of more information, better organised, but I think the two most important aspects of this are for the data to be locally useful (and so following whatever the local requirements for disaggregation and classification are) and not burdensome, because it's very difficult as it is to collect data from all donors and analyse it. In my experience centrally collected data differs significantly from locally collected data in terms of sector disaggregation (rather than overall volumes).
The details of what has been agreed so far are at http://iatistandard.org
It is clear that every project in - to continue your example - Malawi will have to be coded according to international classifications (eg DAC CRS codes) AND local Malawi codes (eg for economic sector, administrative unit etc).
We don't yet know how much double coding will be required (may guess is that each country will have a local set of classifications against which aid projects will be classified by donors as well as the international classifications), though I'm told that in many sectors (though not all) the codes are pretty much universal. To the extent that projects have to be separately coded for local data definitions, donors (not recipient countries) will do it as part of their routine data recording, usually as part of their internal management information system. Indeed, as you say, they are already having to do classify their aid according to these local definition so that they can report to in-country data collection processes. Since donors are already having to classify and report this (or, more pertinently, they should be even if some of them are not, yet) IATI itself won't add to the burden on donors. On the contrary, it should reduce the burden on them by systematising data collection within donor organisations and reducing unnecessary duplicate reporting, on the "code once use often" principle. It will also ensure that the information resulting from these existing efforts is more complete, coherent and consistent with other classifications, and that it is widely and easily available.
There are some technical papers on the options which are in the Technical Advisory Group part of the IATI website. I think the most relevant one is:rnhttp://www.aidtransparency.net/wp-content/uploads/2010/08/Session-4-Budget-Alignment.doc
As you can see, this is work in progress. But there is no insurmountable technical problem nor any political disagreement.
The key point is that information needs of developing countries about aid must be built into the IATI standard, and everybody involved in IATI agrees with that, and has agreed to make that part of the standard. And don't let the idea of an international standard mislead you into thinking that this means that the definitions will be the same for every country: they won't. This part of the standard will respect local definitions, practice and information needs as defined by the developing countries themselves, not defined by some international group.
Owen
I have heard so many times the mantra that we should use the data classification of the partner, however, from most partners, I never saw any classification. If all OCDE countries can classify all their government activities in the same way, would this be so difficult to extend the system to all countries?
The advantages of de facto standards, (like MS Word) are comparable to negotiated standards. The advantages of standards for communication are indeed so important that we should be investigating the moral consequences of not moving to a standard, whatever standard, instead of discussing the fine-print.
Owen -
Thanks for the further clarification and links - I'm glad these things are already being taken seriously. The iati standard website isn't the most user-friendly site in the world, although I suspect that will change with time.
Sam -
I agree that a `common language' is useful for cross-country efficiency, but it isn't neccessary useful for within-country efficiency. While I was at the Malawi budget office, the IMF tried to impose a new accounting standard (GFS), which was comparable with other countries that the IMF had pushed this on.
GFS was a good standard, but not neccessarily better than our home-grown set up, plus it had no direct link to the sort of things we liked to talk about (how the budget relates to the MDGS). It was clear that the people who were benefiting the most were at the IMF - they could now compare data across countries - rather than the Malawians, who would have been forced to change their entire budget system for the umpteenth time. Obviously this is a different scenario than the IATI standard, but it's just an example that not everyone country need the same in-country system.
Owen - thanks for the clear explanation.
Sam - The key point to bear in mind is that the economy and development sphere is not simply an objective entity. It's form depends very much on how people and institutions engage with it, and how they think about it. Conceptions about the broad range of things that contribute to human development differ wildly across developed and developing nations; now you can make a good case that it would fit some purposes to eliminate this variation, but if you believe that partner countries should set the agenda for their own development process, this quickly becomes untenable.
The ancillary point is, of course, that the conceptions of the society are also not merely ways of seeing, but also ways of shaping societies. Things that are deemed to belong in the public sphere and in certain ways are inevitably affected differently based on how they are seen and how they are brought into the public sphere.