IATI Consultations Archive

Live discussions and consultations can be found at discuss.iatistandard.org.

Version 2.01 - Iteration 2 - 5. A consistent approach to OECD DAC organisation identifiers

This is part of the formal proposal that makes up the second iteration of the Version 2.01 Upgrade process.
 
IATI organisation identifier methodology splits identifiers into two parts: a registration agency and a registration identifier. IATI maintains (and continues to build) a code list of registration agencies. The registration identifier is issued and maintained by the agency.To standardise identifiers derived from OECD DAC donor, agency and channel of delivery codes it is proposed that:
  • The OECD DAC is recognised as a registration agency with code XM-DAC.
    • (All agencies are prefixed with ISO country codes: the XM indicates a multilateral/international agency)
  • DAC donor codes should be adopted.
    • IATI at the outset substituted DAC donor codes with ISO country codes. This should be corrected.
    • The government of the United Kingdom (as a donor) is currently represented as "GB". This should now change to XM-DAC-12
  • DAC agency codes should be modified to be consistent with donor codes.
    • UK DFID is currently referenced in IATI as "GB-1". This should now change to XM-DAC-12-1
  • DAC delivery channel codes should be prefixed with the DAC agency code.
    • UNDP is currently referenced in IATI as "41114". This should now change to XM-DAC-41114
  • for discussion go to http://support.iatistandard.org/entries/41042267-Modify-DAC-derived-organisation-identifiers
  • We have mapped existing organisation identifiers that will need to change in version 2.01 against a set of proposed new identifiers. - https://docs.google.com/spreadsheets/d/11TiOakHaaTI1mcK7aMg-0T9QLdH9CbAXllUHi4i1nSU/edit#gid=0

For technical details about implementing this proposal go to: https://github.com/IATI/IATI-Extra-Documentation/issues/67

Have more questions? Submit a request

18 Comments

  • 0
    Avatar
    Ben Webb

    Are we suggesting that people who don't currently use a DAC identifier, should use DAC rather than a different registration agency?

    e.g.

    CDC Group PLC       GB-COH-038777       XM-DAC-12-2
  • 0
    Avatar
    Bill Anderson

    No, we are not.

    Many organisations have valid identifiers from more than one registration agency. There has been discussion (in corridors only) about IATI specifying an order in which agencies should be used, but this has never been agreed or proposed.

  • 0
    Avatar
    Ben Webb

    The example I gave is in the spreadsheet of proposed new identifiers. Should it be removed?

  • 0
    Avatar
    Herman van Loon

    Would this mean that from 2.01 onwards ALL iati identifiers on the activity level would also have to be changed (e.g. from GB-1-xxxxx to XM-DAC-12-1-xxxx)?

  • 0
    Avatar
    Bill Anderson

    Unfortunately it's even more complicated than that:

    For traceability chains to remain unbroken users of third party organisation and activity identifiers should ideally change them at the same time as the third party does.

    We understand this is not great, but we believe the pain is worth the gain. We will also maintain a public cross-mapping table.

  • 0
    Avatar
    Herman van Loon

    Because we already have more than 260 IATI publishers publishing in 1.x format, it will mean that the transition period most likely will be multiple years, since every publisher has it's own pace in their IT change cycle. In this timeperiod, it will be virtually impossible to implement tracebility chains.

    Tracebility though is one of the corderstones and big added values of IATI. The current practice is that linking IATI activities already is quite a challenge. Having two coexisting identification systems  at the samen time will add a horrendous complexity.

    Will the benifits will outweigh the advantages? Possibly this change could be a showstopper for the usability of IATI data in the next years, so there must be huge advantages when making this drastical change. Maybe we should only do this, when NOT making the change, will threaten the long term succes of IATI. What is the opinion of other publishers on this?

  • 0
    Avatar
    Bill Anderson

    There are 116 publishers using /transaction/provider-org/@provider-activity-id - see http://dashboard.iatistandard.org/element/iati-activity_transaction_provider-org_@provider-activity-id.html

    I don't have an exact figure, but the majority of these are using Aidstream - which will be able to do a search and replace on all their accounts at the appropriate moment.

     

  • 0
    Avatar
    Herman van Loon
    There are about 6 organizations with more than 200 activities, representing more than more than 90% of all the activities using this element (unicef, bmgf , worldbank , cafod , ss-org , irishaid). With this number of activities it is very unlikely they use a tool like aidstream. There will also be existing publishers who will use this field in the near future, because of requirements from donors. All those organizations will need to change their systems. Because IATI 2.01 a major change, this will be a fair amount of work which will take quite some time, assuming that every publisher will move to 2.01 (which is an open question). This means that 2 identifications systems will have co-exist for a longer period of time. Therefore during this period it will be even harder than it already is, to link activities from different publishers. The burden will be on all data consumers who would want to have insight in the money flows between organizations. So in my opinion there must be very compelling reasons to make this high impact change. What are the benefits? Which urgent (data quality) problem is being solved? Because we have only a limited number of DAC donors, why not leave the current centrally maintained identification scheme in place? What is important in my opinion is that organizations and activities can be uniquely identified. This requirement is already met in IATI 1.x. Standardization of the identification of the limited number of DAC donors is nice, but not when it will disrupt or complicate efforts to make the aid chain more transparent for the next few years. For the other organizations, there is now already a guideline for organization identifiers in place.
  • 0
    Avatar
    John Adams

    I totally agree with Hermann. I don't believe that the perceived benefit of using a standardised organisation identifier convention outweighs the disruption to the data publishers who are using organisation and activity identifiers to link activities across organisational boundaries. The time taken to propagate the change through DFID and all our downstream publishers would mean that any applications (including DevTracker) built on the traceability model would break until all publishers had fixed and republished their data.

    I also don't see the benefit in adding meaning to a code. Having XX-XXX-nn-nnnn should not convey any meaning, it is purely a convention to allow organisations to develop unique(ish) codes. As Hermann says, we already have a unique coding system for major DAC donors, so let's keep with that.

    I don't yet understand what benefit there would be to migrating to a common code convention when we already have unique identifiers for those organisations.

     

  • 0
    Avatar
    rachel.rank

    Following on from the correspondence above, it would be useful to have a list of the pros and cons of this proposal so Steering Committee members can take a more informed decision on whether the benefits outweigh the risks and costs involved.

  • 0
    Avatar
    Bill Anderson

    Hi John and Herman. I am on leave at the moment and I understand that discussions have been taking place offline with our technical team on this issue. There are many practical issues that I would like to raise in defence of this proposal, which I will do on my return, but I would like to put on record now where I think this issue sits within a bigger context.

    The data revolution is emerging as a key vehicle for post 2015 development strategies. It's objective is to improve, harness, combine and transform all relevant data into coherent, usable information for decision-making globally, nationally and sub-nationally. IATI is one of many systems and standards which needs to throw its weight behind this drive for joined up and interoperable data.

    For this to work we need to work together with other systems and standards to create a common language for data from different sources to be aligned and compared: global data standards.

    These standards may be relatively simple (formats for dates), supposedly simple (definition of supranational regions) or complex (sector definitions).

    IATI's position has always been that, wherever possible, it should adopt existing standards, and work with other standards bodies or authorities to improve global compatibility so that different systems can talk to each other. For example, our country, language and currency codes are controlled by ISO authority lists; our region codes allow for two authority lists (OECD DAC and UN).

    There is no global standard for uniquely identifying organisations. This is a major problem and is likely to become an even bigger one. OpenCorporates led the way in developing a methodology for identifying private sector entities, and IATI has adopted this for both private sector and NGOs: most of these entities are registered with one or more registration agency who issue and record (publicly or privately) unique identifiers - by keeping an authority list of the registration agencies (currently owned by IATI in the absence of any broader body willing to take it on) we can construct globally unique and meaningful identifiers for all registered institutions. This approach has a very low overhead and is agnostic to the application in which it is to be used. As such it is a likely candidate for the emerging openContracting standard and for any future work on budgeting and spending standards.

    There is currently no agreement (and, to be honest, little interest) for extending this methodology to public entities. How, for example, can we define the Rwandan Ministry of Health in a machine readable format that could be understood (or at least cross-mapped) by different systems? This is a serious problem that needs a solution and IATI has, and should continue to, engage with others in efforts to solve it.

    It looks unlikely that a global institution is about to pop up from out of the blue and create a whole new system of identifiers. Our best bet is to continue with the existing methodology and look to federated authority lists that may cover a single country, region, or multilateral framework. It is for this reason that we are proposing to recognise the OECD DAC code lists for donors, donor agencies and channels of delivery as a single authority list providing reliable and meaningful identification of a particular subset of agencies.

    When we drafted the IATI standard we had not thought through the importance of authority lists in maintaining the integrity of codes, identifiers and definitions. We also saw the world very much through the lens of traditional donors and DAC reporters. So we created an 'IATI' list of organisation identifiers derived (but different) from the DAC with no provenance and no controls. One "non-rule" for DAC reporters and a proper rule for everyone else. This was wrong as it didn't reflect the multi stakeholder nature of IATI and it did not create a foundation on which a sustainable standard can be built. What up until now has been a minor flaw will, in ten years time, become a nightmare.

  • 0
    Avatar
    Herman van Loon

    Hi Bill. Thanks for your eleborate answer during your leave. I still do not understand why it is nessesary to change the existing identifiers of all DAC reporters and why it will cause a nightmare in ten years time. Can you please elaborate on that topic a little bit more?

    The IATI list of DAC reporting organisations seems to be manageble (it has currently 420 entries).This list is for a major part an adoptation of the existing DAC identifiers for bilateral donors  and multilateral organisations. So the current list is for a very large part already DAC compliant. IATI can follow the DAC using the currently existing scheme, without changing all the identifiers which are already in use. Maybe only a few inconsistent codes need to be resolved if you want keep in sync with the DAC.

    The proposed change will not solve the problem of the large volume of public entities which are not DAC members.  So if I am not mistaken the broader issue of the identification of public entities in general is not solved by this change. Also the problem remains of organisations on the IATI list for which there is no counterpart on the current DAC list. How will those organisations in the new identification scheme be represented? Is the DAC willing to fulfill the role as defining authority for those organisations? Or will there be an exception for those organisations?

  • 0
    Avatar
    Bill Anderson
    Hi Herman and John

    A. Regarding the adoption of a standard methodology

    1. The proposal is that we adopt a standard methodology that applies to the identification of all organisations.
    2. The proposal is that the methodology should not be IATI specific but be applicable to a range of existing or emerging data standards.
    3. The proposal is that all organisation identifiers consist of two parts: a code for a recognised registration agency; and an identifier issued by that agency.
    4. The first part MUST validate against a code list containing all recognised registration agencies. This list is currently owned by IATI (but it would be better if it were owned by a broader body of institutions).
    5. This methodology is similar to IATI's approach to other elements and attributes where more than one authority list exists. (Region, sector, etc) A "vocabulary" (i.e. Registration Agency) defines which authority list should be used.
    6. While sectors and regions store the vocabulary as a separate attribute, the proposal combines the two as a single value because the organisation identifier - and more importantly the activity identifier (which contains the organisation identifier) - is used to search and cross-map between activities and publishers, particularly for traceability purposes. The activity identifier MUST be globally unique and it is much easier to use and validate this within a single value rather than across three separate values (registration agency, registration number, activity number)
    7. There remains the challenge of identifying registration agencies beyond those already identified. The methodology does, however, allow for ongoing additions to the list of recognised registration agencies. (Sweden have been considering establishing SE-GOV as an agency to allow them to identify departments and agencies not on the DAC list.)
    8. There is no reason why IATI itself cannot maintain its own list, PROVIDED THAT the methodology is followed. (In fact I would argue for UNOPS to become XM-IATI-41AAA as while '41AAA' was constructed to look like a DAC challenge code, it isn't)
    9. My reason for saying that we will have a nightmare in 10 years time if we do not adopt a standard methodology now, is that without a set methodology it will be impossible to track the origin of identifiers and control their uniqueness. And it will be impossible for different standards to accurately share identifiers that have no formal governance.

    B. Regarding the adoption of XM-DAC and the changing of 'NL-1' to 'XM-DAC-7-1'

    1. IATI already has an undertaking with the OECD DAC that in the interests of consistency within the Busan common standard all codes used within the CRS will be accurately reflected in IATI and that IATI will not modify these lists in any way. 
    2. IATI practice up until now has been to replace the DAC donor code for governments with the ISO country code. (Eg. DAC donor code '7' was replaced with 'NL') This is inconsistent with our common standard commitment. (NB that DAC donor codes refer to institutions, not countries)
    3. These codes, although they don't look like DAC codes, are in fact controlled by the Working Party on Statistics. If Netherlands wants to add a new agency code (eg 'NL-6' under the current system) IATI's existing practice is for the '6' to be approved by DAC before being used in IATI
    4. As stated in 9 above it would be possible for existing codes to stay the same, provided they became IATI, rather than DAC codes, AND added an XM-IATI prefix.

    C. Regarding the practicalities of such a change

    1. We accept that this is disruptive, but the lack of a consistent methodology will damage us in the long run, as argued above.
    2. If we do not fix the current flawed methodology now, it will be exponentially more disrupting when we get to the next integer upgrade.
    3. Only the Netherlands and the UK are currently involved in traceability. 
    4. I would of thought that it would not be too difficult for both donors to maintain a transition period in which both old and new provider-org ids would be recognised
    5. The IATI technical team will be available to assist in the transition.
    Could you indicate which of these points are not acceptable to you?
    Best
    Bill
  • 0
    Avatar
    Herman van Loon

    Thanks Bill for your eleborate answer. The main objections are against point A.8 and C.1&2:

    Why would there be any harm in leaving the organisation identifiers allready being issued in place? Just use the new methodology for new organisations not already on the IATI list. So this list will not grow anymore. The only problem is that in the long run a fixed part of the list does not fully comply with the new standard. This will in my opinion not lead to any practical problems.

    Adopting this migration approach would avoid having two identification schemes at the same time during the transition period. The length of the transition period would be depending on the time the IATI publishers using these identifiers need to upgrade to 2.01. Because of the huge change 2.01 represents, this will probably be a lengty period. During the transition time, it will be more difficult then it already is for IATI publishers to  'link up the chain'.

    I hope that linking up the chain  will in the near future will not be limited to UK and the Netherlands, because enabling chain transperancy is in my opnion one of the most important reasons why we have IATI. Any obstacle linking data across organisation boundaries should if possible be avoided. Therefore I am reluctant to agree with this proposal. I am still not convinced the benefits outhweigh the costs.

  • 0
    Avatar
    Bill Anderson
    Thanks Herman Objecting to A.8 surely also means that you don't accept A. 1-4 : that there should be standard, global methodology for all identifiers. Is this correct? Bill
  • 0
    Avatar
    Herman van Loon

    I have no strong opinion on this. A system of self-registration, as John suggested, could also work. We do that right now for IATI publishers already. Self-registration would have the big advantage of having no need of governing bodies. Self-registration could just assign a meaningless unique identification number to an organisation. I just wonder if the fact that for a number of organisations there is no clear governing body (e.g. government organisations, local NGO's in some development countries), does make a self-registration system necessary anyhow.  With self-registration you could even connect the meaningless identifier also to other 'official' identifiers. It would be up to the registering organisation.

    What we are ctrying to create is a world wide system for organisation identifiers under control of governing bodies. Is this feasible? Is this needed for IATI to be succesfull? But if the IATI community thinks so, I have no objections to the proposed scheme, as long as it is limited to new cases.

  • 0
    Avatar
    rachel.rank

    I've been following the discussions above and am still struggling to understand the overall benefits of this proposal. I agree with Herman and John that it does not seem to solve the problem of the large number of public entities which are not DAC members. It also sounds as if it will be heavily disruptive, and if DFID and NL are the only two publishers currently using traceability and there are enough things to work out already, we don't want to make their (and others') lives harder.

    I would like to propose that we continue to use the existing IATI codelist and structure until a more robust methodology has been established (including issues such as how to deal with Ministries in partner countries, donor agencies not on DAC lists like EC-FPI, etc.). Once we have that new methodology, we will have the pain of transition to a new methodology, but it doesn't seem sensible to suffer that pain twice. Can we not stick with the present approach::

    1) Is the code on the IATI codelists?

    2) If not, is it structured in line with the IATI methodology?

    It would be useful to establish a working group to look into this, including participants from other initiatives. In particular, it would be good to explore John’s suggestion of a system of self-registration.

    Interested to hear what others think about the idea of setting up a working group.

  • 0
    Avatar
    Bill Anderson

    I would like to see whether we now can reach a consensus to move forward - specifically for the 2.01 upgrade.

    (Discussion on this thread has moved on to the technical mailing list here - https://groups.google.com/forum/#!topic/iati-technical/pjXkBP5Mef4 - so I am going to post this in both places.)

    The issue is not directly related to the changing of DAC 'codes', but specifically to how we validate organisation and activity identifiers.

    • In Strengthening the Core of the Standard we have  proposed that all organisation identifiers MUST be prefixed with a code from the OrganisationRegistrationAgency code list.
      • I think we now have consensus on this.
      • There is an outstanding proposal that the prefix (vocabulary/uri) is stored in a separate attribute. This proposal still accepts the necessary association between an identifier and a registration agency.
      • There is still a lot more work to be done on:
        • public entity identifiers
        • a self-registration registry
        • a common approach by IATI, OpenContracting, HXL and others to solving gaps and problems
      • but this outstanding work does not negate the establishment of the rule.
    • The second proposal made in Strengthening the Core is that the activity-identifier MUST be prefixed with the reporting organisation's identifier
      • Opposition to this is at the heart of Herman van Loon's dogged contributions to this discussion and we now fully understand and accept his argument.
      • As an activity-based standard, persistent activity identifiers are necessary. There are reasons why organisation identifiers may need to change as a result of mergers, acquisitions or reorganisations. But to track activities over time, and the traceable flow of resources between activities - changing activity identifiers is unworkable. 
      • We need to agree a new approach to guidance on activity-identifier. A possible solution is to maintain the current methodology (reporting-org/@ref prefix) when first reporting an activity PROVIDED that it is never changed in subsequent updates, even if the reporting-org/@ref changes. This has the added benefit of showing the origin of an activity where organisational changes have taken place.

     

    So, the actual changes required to the 2.01 proposal are (in Strengthening the Core of the Standard) the removal of two lines (struck through), and the addition of one (underlined) - as below

    Reporting organisation

    • The reporting-org element is MANDATORY.
    • ALL the following rules must apply to the organisation-identifier in reporting-org/@ref
      • It is mandatory
      • The agency prefix MUST be a valid code in the IATI code list
      • The identifier MUST be the prefix to the iati-identifier
      • The identifier MUST be the same as that recorded by the publisher on the IATI Registry
      • The identifier MUST only contain alphanumeric characters and hyphen, underscore, colon or period

    IATI activity identifier

    • The iati-identifier is MANDATORY
    • It MUST be globally unique among all activities published by a single reporting organisation
    • Once an activity has been reported to IATI its identifier MUST NOT be changed in subsequent updates,
    • It MUST be prefixed with the organisation-identifier found in reporting-org/@ref
    • The identifier MUST only contain alphanumeric characters and hyphen, underscore, colon or period
Please sign in to leave a comment.