I’m hoping people can join in - and even call me out as wrong - on this thread. I’ve purposely written it this way…!
Traceability: what we think happens:
- Donor A publishes a Disbursement of £100, on the 1st Feb 2017.
- This transaction has a reference: 12345
- They use receiver-org to clearly identify that Organisation B is destination of the funds
- They also use receiver-activity-id to point to a specific IATI activity, published by Organisation B, that receives the fund
and
- Organisation B also publishes IATI data, which includes an Incoming Fund for £100 on 1st Feb 2017
- This transaction has a reference: 12345
- They indicate that the provider-org is Donor A, and also include the provider-activity-id to signal that the source of the funds is an activity published by Donor A
Together, this looks like great traceability. We can connect two organisations together, alongside their activities, via a couple of transaction.
But! The transactions are not actually linked together. Each publisher can provide their internal reference for the transaction, but it’s not possible to cite other transaction references, in the same way we can for organisations and activities.
This might seem like nitpicking. After all, the above two transactions look like they link together. All available parts of the standard are in place, so good enough, right?
At a human level - eyeballing the data - we can probably make the assumption that this D is connected to the IF. But (and this is a reason for our standard) how can a machine read and connect them?
This becomes tricky when - for example - the values of the transactions don’t match, for various reasons. I have been discussing with one organisation how they receive a single IF from a donor, and then make several D to partners, from that. It’s possible to see this in their accounting system, but when expressed via IATI, it’s difficult to then match things up - particularly when there are multiple transactions flowing in and out.
More fundamentally, we might want to ask: is this important for the IATI community? Are we concerned about tracing and linking specific transactions together? Or - can we accept that traceability in IATI will not be so precise?
And - do post an alternative take on this, if you have one.
Hi Steven Flower ,
Some comments:
1 - your assumption is that donor A knows the activity identifier of the recipient organization at the moment of disbursement. That is often not the case, since that activity might not yet exist in B’s administration.
2 - When organization B refers to the correct provider-org-activity-id of donor A, there can be no misunderstanding which activity is being funded where the funding is coming from.
3 - when this happens at the same date and the amounts are the same that is more than good enough for traceability purposes.
4 - when dates or amount differ slightly, you can still get a reasonable picture of what happens.
5 - if they differ greatly, this is a good reason to further look into the data quality of each publisher.
So I fully agree with Bill Anderson : we shouldn’t try to make IATI financial auditor proof banking system. That would divert our focus from far more interesting subjects like results reporting, etc.
Point taken.
(Plus some padding 'cos Discuss says my reply has to be at least 20 characters - that’s a bit silly)
I think there is likely consensus in this thread that IATI is not trying to provide traceability between transactions.
The question then is how we make sure we don’t give that impression to the outside world when talking traceability.
In the work we have been doing with DFID, the framing was in terms of “Delivery chains” and I wonder if this would be a better framing for any future discussions?
Thanks Bill Anderson Herman van Loon and Aria Grabowski for your contributions
To clarify, I am not advocating for IATI to change/switch/adopt to an “banking system”. I just wanted to highlight how it can very nearly be perceived as that. Having spoken to numerous donors, publishers, data users and interested bystanders - not often fortunate enough to spend hours with the standard and those around it - it has become clear that this perception is widespread.
And - this isn’t a surprise imho. We use the terms:
For the person in a finance team asked to implement IATI, they can quite easily believe it is about specific transactions, that relate to each other. As we talk more about traceability, and use these terms, then others may look for the specifics I outline above, rather than the gist or “general picture”.
bill_anderson:Understood. Can other people find that clarity too?
I don’t think this is about my nitpicking, or what’s the right or wrong implementation of the standard. If we are serious about widespread adoption, then we’re going to have to spell things like this out for data producers and users means. Repeatedly.