Current Discussion: Quantifying percentage of funds allocated from an activity to COVID-19

[IATI-techteam (IATI Technical Team)]

One question about COVID-19 IATI data, is how can we tell exactly how much of a given activity is being budgeted, committed and spent on COVID19. This is particularly relevant where activities may be a mix, or contain some redirected funds.

Two options have been proposed:

• Add in COVID-19 to transaction narratives. This allows publishers to detail which commitments, incoming funds and disbursements are specific COVID-19 funding. Most organisations can already do this. However, this is cumbersome to find for data users, not all transactions may be COVID-19 specific, and timeliness of data is an issue.

• Or the sector element could be used. A binary ‘COVID-19’ and ‘not COVID-19’ category could be used. Either from adding a new codelist or organisations using vocabulary 98 and 99. This can be published alongside other sector elements and provide data users a way of interpreting budget allocations and other financial details. However, this option must also include the narrative element, how would this be used? If reported at transaction level, transactions will need to be separated into COVID-19 and not COVID-19.

Please see @rory_scott 's post and @JieruZ 's post on using the Sector element for previous discussions.

[stevieflow (Steven Flower)]

Thank you for detailing this specific topic

• Description text to transactions: agree that it could be easier to implement (for some) but less helpful for machine-analysis

• Transactions/sector - I think this highlights an important gotcha, but not sure if it is explicit…

The standard says two things in relation to sectors at the transaction level:

1. If transactions use sectors, then there must be no sectors at the activity level (I shared observations on this a two and a half years ago)
2. If one transaction has a sector, then every single other transaction must have a sector (regardless of type of transaction)

I interpret the suggestion from the @IATI-techteam that there be a covid and non-covid sector code, so that users could quickly flag the relevant transactions. As per [2] above - any publisher would need to categorise ALL their transactions, to be in line with the rule stated in the standard.

Others may observe that this may in turn devalue the purpose of the sector classification at the transaction level, as it is just being used as a on/off flag, rather than provide richer detail.

If a publisher already uses 98 or 99 at the activity level, they would also - in line with [1] - have to remove that data.

[stevieflow (Steven Flower)]

Also: We should not discount the idea of extensions to explore this topic. The X of XML is for Extensible (@David_Megginson!) !

Whilst there might be a technical overhead for both publication and data use, working with an extension can give us the room to trial, discuss and develop - whilst not getting into the complexity and frustration of change processes around the standard.

For example:

<iati-activity xmlns:covid="http://covid.iatistandard.org/"

...

<transaction>

....

<covid:flag>true</covid:flag>

[AVaughn (Andie Vaughn)]

Hi all -

USAID recently realized (oops) that placement of the humanitarian scope at the transaction level was causing a critical validation error. However, there is no way for us to put the scope element at the activity level without losing significant granularity in our information. We already put many details (such as country, sector) at the transaction level to improve the granularity of activity information. Additionally, including the scope at the activity level would over-count USAID’s efforts to specific responses - such as Covid-19 - dramatically.

• In terms of visualizations:

  • D-Portal: There is no way to identify USAID Covid-19 transactions (because the way we are reporting is not valid). For example, USAID/Nepal’s Suaahara II project has a value of about $60M with $1.1M of that directed toward Covid-19 response. There is no way on d-Portal to see this activity or which transaction is Covid related as we currently report now with the scope at the transaction level.

  • New Covid-19 Prototype: This was updated to only pull the transactions which include the Covid-19 GLIDE codes so as not to significantly exaggerate the USAID funding for Covid-19 - but we are hoping to find a long-term solution that is consistent with the rest of the IATI community.

USAID is currently looking at including a narrative in the transaction description, but is also looking at other options (such as an extension) and would like to discuss best practices with other publishers.

Any ideas would be much appreciated.

[JieruZ (Jieru Zheng)]

As sectors can be used for activity or transaction level (not both), if this were the way forward to quantify funds allocated to COVID-19, we should allow the possibilities to choose either to publish on COVID-19.

• If publisher chooses to use sector code at activity level (vocabulary 98/99), then percentage attribute should be used to indicate % funds allocated to COVID-19 for that activity;
• If publisher chooses to use sector code at transaction level, then no percentage attribute (all funds of that transaction go to COVID-19)

Ideally a sector code for COVID-19 use only would be the optimal solution. In the interim time, sector vocabulary 98/99 can be used for this purpose.

[IATI-techteam (IATI Technical Team)]

@AVaughn many thanks for posting what USAID is currently doing. For identifying transactions, in the IATI publishing guidance, we currently advice that organisations include “COVID-19” in the transaction description.

The other options currently discussed are:

1. Use of sector codes- well noted by @stevieflow and @JieruZ that sector can either be used at activity or transaction level. As currently there is no sector code for COVID-19, vocabulary 99 with a binary code of “Covid-19” and “non-Covid-19” will have to be used.
   In relation to this we are also in touch with the OECD to understand how COVID-19 will be referenced in the CRS. There is a meeting in mid-May on tracking Covid-19 related expenses in the CRS and we will update the community on this.
2. Add an extension to the transaction element. We would advise that we first consider the options within the standard before looking at extensions.
3. It will also be interesting to hear how organisation are currently using the transaction @ref and whether in their internal system transactions are already references to COVID-19.

[stevieflow (Steven Flower)]

Thanks

As currently there is no sector code for COVID-19, vocabulary 99 with a binary code of “Covid-19” and “non-Covid-19” will have to be used.

Here be dragons. This is in effect suggesting dedicating 99 to a tagging for covid.

What about those organisations that already use the 99 vocabulary for something else / internal?

Secondly, I do not believe that the sector field in IATI is intended for any form of flagging/tagging. This was the whole basis of the discussion around “non-statistical” use of the sector field (see here). This led to the addition of the tag field.

Thirdly (and perhaps more complexity!) , we have to consider - and decipher - where the standard sits in terms of use of sector codes between activity, transactions and multiple vocabularies. We already know that the standard states that if ONE transaction has a sector a particular vocabulary, then ALL other transactions must too. This seems to be regardless of the range or number of vocabularies. I think this could benefit from attention (of a Working Group!)

how organisation are currently using the transaction @ref

We have a discussion on this here. I know that USAID, DFID and others use transaction/@ref to output a form of UID for a transaction (even if just relative to the parent activity). I would not favour using this as a proxy flag for the covid use-case - we would starting to be moving far away from what is intended for that field (even if it is quite unclear in the first place!)