Thoughts on Ten Steps (or at least a few of them)
While raw data is fantastic, especially for researchers and analysts, there are also a lot of people interested in aid who would like their information pre-cooked to some degree.
Hats off to Owen Barder for his tour de force of a blog post on what needs to happen in aid transparency! At AidData we cheer his arguments that data should be comparable, open, geocoded, and reusable. This is at the heart of the AidData initiative—ensuring that aid datasets integrated into the database are appropriately formatted and coded so that they can be compared with other datasets; making everything public and easily downloadable; geocoding at the activity level; and enabling raw exports rather than pre-packaged or analyzed data so that information can be (dis)aggregated for all sorts of purposes. Soon, it will be possible to export data in IATI format as well.
While all this raw data is fantastic, especially for researchers and analysts, as Owen points out there are also a lot of people interested in aid who would like their information pre-cooked to some degree: a journalist who wants to know how much donors gave to his or her country; a citizen who wants a snapshot of what donors are doing in his or her community; or an NGO that wants a list of other NGOs active in the same sector. Although AidData at the moment may look like simply a big database, behind the scenes there is a lot of work going on to make it more digestible and visual for the kinds of users who would prefer not to read through pages of codes and standard definitions in order to answer a few burning questions (stay tuned!). Along with geocoding, this will pave the way for new initiatives to crowdsource development information, which the AidData team is looking to pilot in Uganda along with Ushahidi and UNICEF.
Linking IATI-format information with country-level aid management systems will be another exciting step (aidinfo and Development Gateway have tested this concept in Malawi and Burkina Faso, working with the national governments). Some of these systems actually do already link to national budget processes: in DRC and Madagascar, for example, the governments use information in the Aid Management Platform (AMP) to prepare their annual budgets. In Senegal, budget information is imported into the local AMP so that it can be analyzed alongside aid, since as Owen says, most stakeholders don’t care whether resources come from aid or domestic resources—they just want to know what resources are available, period. And although many of these systems are not publicly accessible, they are frequently used to create official aid reports that are made public (some examples here).
Adding IATI data feeds to country systems will help governments track activities that are implemented by donors and NGOs outside government channels. I am skeptical that requiring donors to report aid information according to each country budget classification (as Owen advocates) will be feasible at the donor HQ level, since each country has its own classifications. In the spirit of country ownership, seems to me better left to the partner governments themselves to decide how aid maps to their own codes (assuming that they have sufficiently timely and detailed information, per IATI, to do so). IATI will nevertheless reduce transaction costs and improve transparency at the country level significantly.