Budget identifier

[andylolz]

Description

The budget identifier is a way of linking the activity to the recipient country government’s own budget codes.

This is only expected if the activity is in the implementation, completion or post-completion phase, and if the aid type is not budget support. This is not expected if the aid type is budget support (A01 or A02) or if the sector code is administrative costs (91010).

Proposed test

For each current activity,
i​f `activity-status/@code` i​s one of (2, 3, 4)
a​nd `default-aid-type/@code` i​s n​ot A01
a​nd `default-aid-type/@code` i​s n​ot A02
a​nd `transaction/aid-type/@code` i​s n​ot A01
a​nd `transaction/aid-type/@code` i​s n​ot A02
a​nd `sector/@code` i​s n​ot 91010
then `capital-spend` should be present

Issues identified

Since the 2016 Aid Transparency Index, the Budget Identifier codelist has been deprecated. The preferred method for linking activities to recipient country government’s budget codes is to use the most granular DAC 5-digit codes, including voluntary codes. We welcome suggestions on how to assess this.

[markbrough]

I think it's really important to include a sector code test in this indicator. Being able to link aid projects to country budget codes is important for IATI to deliver on its full potential. Following the work with WP-STAT to improve CRS purpose codes, checking whether the more granular CRS purpose codes are being used is now the best way to implement this test in line with the methodology established. The set of codes that are not mappable to country budgets are shown in the 2012 IATI Study on better reflecting aid flows in country budgets and the work undertaken by the IATI TAG Working Group on Aid and Budget Alignment.

[YohannaLoucheur]

Fully agree that testing the use of the more granular CRS codes would be important, and contribute to a better assessment of whether donors meet the data needs of partner countries.

I would suggest to add the "core funding" codes to the exclusions (B01, B02), as these can rarely be coded to a specific sector, especially not at a very disaggregated level. There are exceptions (e.g. an NGO or multilateral focused on a very very specific issue), but these are quite rare and not worth the noise created by testing all B01 and B02, in my opinion.

[andylolz]

Thanks @markbrough and @YohannaLoucheur.

We’ve been in discussion with @markbrough about this, and are still looking into the viability of testing against a subset of DAC CRS purpose codes. This partly relies on the existence of clear guidelines for publishers on how to do this, which we could then reference.

[YohannaLoucheur]

This partly relies on the existence of clear guidelines for publishers on how to do this, which we could then reference

By "how to do this", you mean using the codes? They are part of the CRS purpose codelist, they have definitions etc. There is no need for guidelines other than whatever is available to use CRS purpose codes (which is or should be on the DAC website).

[AndieVaughn]

We agree with the earlier comments to use the DAC purpose codes.

Is the first part of the test that assessing if an budget identifier is reported not there? It appears that this data is only measuring if capital spend is present.

[johnadamsDFID]

This test needs a bit more work, starting from the user need - to be able to map against country budgets. I'd support a redesign of the test, but looking for specific DAC purpose codes would exclude legitimately coded activities that didn't happen to use those specific codes.

More discussion needed.

[carlelmstam]

Might be that the test needs redesigning. But if the basic structure is kept other exceptions are needed like migration costs in donor countries and other sectors where capital spend does not make sense.

[andylolz]

Thanks for the discussion here.

Firstly, there were a number of comments about testing against a subset of DAC CRS purpose codes. @markbrough has now written some suggested guidance on how publishers should do this (i.e. don’t use six codes identified as being too broad; don’t use the fifteen aggregate codes, but instead use the more detailed ‘voluntary’ codes.) This suggested guidance is newly published as a direct result of this consultation, which we welcome, and hope it will be integrated into the IATI guidelines.

On the capital-spend test: Based on feedback elsewhere, we will modify the test to ignore default-aid-type/@code G01, rather than sector code 91010. Apart from that, we think it should be possible to declare capital-spend/@percentage (even if this means declaring it as zero).