Notes on open data and transparency

Beneficial ownership, corporate disclosure, tax justice, supply chains, policy and history

Week: Jan 20, 2020 → Jan 24, 2020


1. We spent most of the week co-operating at one of our quarterly meetings so there wasn’t a huge amount of time for delivery. I often find the prospect of an enforced break from my main project to be daunting; in practice, being away is always really helpful as a creative catalyst. (I already know this but also always manage forget this.) 2. As part of a systems migration, Simon has been transferring lots of our internal tickets on beneficial ownership over to Github and closing or moving others to the new system (Notion). As a result of that, I’ve been reminded of a lot of old thinking on beneficial ownership transparency.

One thing in particular struck me as an interesting problem: in a ‘paper→electronic’ system what is the best strategy to deal with mistakes (e.g., misspelling a personal name or jurisdiction)? The question has become more interesting in the face of increasing evidence that deliberate mistakes are a significant part of the threat model for beneficial ownership registers. And the the principles for dealing with the problem also seem to generalise quite well to any information system where:

  1. There is a discretionary gatekeeping function on data inputs.
  2. The accuracy and integrity of data on the system are very important.
  3. The overall speed and efficiency of the system are important.

Therefore a paper→electronic system probably needs to take the following into account when designing its business processes:

  1. Administrators should have some leeway to correct submissions because people will make mistakes and outright rejection for any mistake may grind the system to a halt.
  2. Administrators should be able to reject a submission outright. Without this possibility, the rate of dishonest mistakes in submissions is likely to rise (and a mechanism to reduce the number of honest mistakes over time is lost).
  3. Administrators should be able to report a submission if a mistake is believed to be deliberate.
  4. The decision to correct or reject a submission should be rules-based. The rules should not be made public and should be adjusted on an ongoing basis.
  5. Where a correction is made, data and/or metadata about the correction should be recorded and preferably published. This might be: a correction flag; a record of the original submitted value alongside the corrected value; information on who corrected the submission.


1. We had one of our quarterly COGMs this week, in Manchester. The C stands for Coworking, as we have moved the formal OGM to a virtual meeting a fortnight or so after our physical time together. This now means that the OGM in COGM stands for “not an Ordinary General Meeting” but let’s not worry about that - it’s just nice to see everyone.

2. Our venue choices are one of my favourite things about our COGMs because we get to act on our cooperative values rather than just talk about them. We spent one of our COGM days at the Salford Lads Club. It’s an amazing and inspiring community space, with an air of solidity and history that makes it a pleasure to spend time in. Less coop-y, but a lot of fun, was trying to get value for money on the unlimited cakes and coffee at Ziferblat.

3. We have been moving towards more external facilitation and specialist input at our COGMs. This dramatically reduces the amount of physical and mental energy needed for a 2.5 day meeting and seems to produce better results. This time we had a half day session on our finances with Andrew Woodcock and a full day developing our strategy with Alan Heckman. Both sessions were excellent. We finished with an extended retro on the outcome of the last COGM: the internal ‘improvement program’ we have been implementing with the support of Nat Lombardio and Rich Bartlett of The Hum/Loomio - this, too, was a solid investment of time and money.

4. Bringing in external input and specialist advice helps us address two of our riskiest organisational behaviours: risk aversion and a freelancer mindset. Risk aversion makes us unwilling to invest, which has opportunity costs. Freelancer mindset creates an attitude that we can and should do everything ourselves, which impacts on workloads and may lead to us reinventing the wheel (badly). Every time we pay someone to help us, investing some of our reserves to improve our lives in future, our default behaviours are shifted in a healthier direction.

5. Getting everyone together in a room to focus on finance and strategy is an important statement on two of the co-op principles: democratic worker control and member economic participation. But periodic attention isn’t a substitute for ongoing engagement. I think this model presents a particular risk that, without proper discussion of prioritisation and areas of disagreement, we mistake a few hours of harmony on high-level principles for the absence of conflict or concern about what needs to be done.


1. After a recent breakthrough with reading in Russian (as in: it was possible), I demolished in a few weeks all of the graded readers that I had been hoarding. I just got a new pile of stuff, one level up, and have started on the first of these books. “Easy Readers” are my favourite series as they are adapted from good source material and remain relatively idiomatic, and therefore not that “easy” at all. The first of these new books (a detective story called Peter and Peter) makes me reliably sleepy after reading a few pages - so I guess that it’s the right level.

Other posts in the Weeknotes series: