If you’ve read our previous posts about this pilot, you’ll know that we’ve been asking what people need, discussing what transparency means, and digging into the data behind a couple of the Country Based Pooled Funds (CBPFs) managed by UNOCHA. All of these activities were building towards one thing: the Simulation Exercise.
The What, How and Why of the Simulation are pretty simple. With support from UNOCHA colleagues in New York and Geneva, and in collaboration with colleagues working on the CBPFs for the Occupied Palestinian Territories and Iraq, we gathered historical data on financial transactions, focused on the period 2018–2019.
We seeded that data into a version of the Disberse platform and ran all the transactions again, recreating the funding delivery chain during that period. We then compared the historical performance of the CBPFs against the Simulation in order to assess how our platform could improve that performance.
In practice, of course, these things are never simple; our plan to hold co-design workshops with DFID staff to develop use cases was ever-so-slightly disrupted by a global pandemic, for instance. Yet we successfully completed the Simulation, and we can now share some of the results.
Moving funds more quickly
Before we start, you should read these definitions — they’re key to our analytical model, so they’re important if we want to understand what we’re looking at. It is also useful to remember that CBPF delivery chains are divided into Tiers, and when funds flow between different organisations they are usually flowing between different Tiers towards final delivery.
- Transaction time: How long it takes for funds to move from one organisation to another, usually through a third-party financial service provider.
- Handling time: How long it takes for an organisation to process funds from receipt to transfer to another organisation.
- Holding time: How long an organisation holds funds on account before transferring them to another organisation or spending them.
- Delivery time: The sum of Transaction, Handling and Holding, which shows how long it takes for funds to move through the delivery chain.
The Simulation showed that Transaction times for CBPF funds would be greatly decreased if all stakeholders in a delivery chain used the Disberse service. The service decreases the number of transactions with other financial institutions needed for delivery; so the longer the delivery chain, the greater the improvement in speed. (Figure A gives a summary.)
Although this is a small gain for each individual transaction, when aggregated across all the organisations funded under an entire CBPF it represents a huge potential time saving. In the two CBPFs that we surveyed, the aggregate number of days lost in transaction time was 400 days, but using the Disberse service reduced that to zero days lost.
Handling and Holding times were not reduced in the Simulation because they are internal to aid organisations — and so not something that Disberse could disintermediate! However the Simulation model confirmed that organisations using the Disberse service could increase the transparency of these times, and that we would be able to give those clients more data to improve analysis of their internal processes and to identify potential efficiency gains.
Decreasing financial costs and risks
Existing financial systems make it almost impossible to provide a true accounting of transaction costs, which meant we had more limited data to work with regarding cost savings. Despite this, the Simulation suggested that there was potential for savings of 70–85%, again through disintermediation; keeping funds on-platform removed the need for exchanging them between multiple currencies as they moved through different countries.
In addition the proposed cost structure of the Disberse service is quite different to other financial institutions — a one time service fee which would be paid by donors to reduce overall costs along the whole delivery chain — and our calculations suggested that this would also reduce costs by eliminating some banking fees.
We also identified potential for further savings and lower risks if all organisations sharing a delivery chain used the service. Aggregate risk in a delivery chain is increased by a single weak link, but because that risk is spread out along the delivery chain, no single organisation can identify them, let alone address them. This is a classic collective action problem which exposes the entire delivery chain.
Donors should be concerned with mitigating this aggregated risk in order to preserve the value of their funds. The Simulation showed that the Disberse could eliminate most (although not all) of this risk if all delivery chain stakeholders used the service; data collected from the entire delivery chain would generate improved analysis that could not be performed by any individual organisation.
Identifying Use Cases and Analytics
From the specific use cases described by interview respondents in earlier sprints — Delivery Chain Mapping, Audit and Control and Liquidity Forecasts — we identified several strategic opportunities that we could provide to clients, including: optimising the timing of currency exchanges, negotiating better exchange rates based on volume, and developing alternative funding strategies. Having real-time data also highlighted a number of analytics and mapping opportunities that hopefully can be explored further.
For example Figure B shows the number of transactions per partner tier in any given month. This shows the velocity of funds within the delivery chains; understanding when funds are moving, and which tiers are most active during which months can help the Pooled Fund stakeholders to plan more effectively — not just to manage cash flows, but also to identify seasonal patterns. No single organisation would be able to generate this data, but we were able to incorporate this visualisation into the Simulation version of the service.
Another example of these results can be found if we return to Handling and Holding times. Although Disberse cannot disintermediate internal processes, Figure C shows that we can extract data on organisational performance that allows us to provide metrics to enable organisations to compare their performance with other organisations in the same Tier.
These organisational efficiencies can be communicated so that both internal changes and in-platform functionalities can support improvements. Disberse could facilitate streamlining by building added-value services to support those procedures — for example, we already have a dual signature function on platform which could speed up payment approval.
Learning Lessons and Moving Forward
Sprint 3 successfully showed how financial transactions can be executed and tracked on the Disberse blockchain-based platform, and the benchmarking analysis gives us initial insight into the potential benefits, including lower transaction costs, shorter delivery times, and real-time tracking through fully transparent delivery chains.
The biggest challenge we faced was pulling together data fragmented between multiple organisations using different systems in different countries; and which is complicated by the accounting challenges inherent in the structure of pooled funds and individual projects — for example, reconciling expenditures within multi-donor and multi-year projects.
Our IATI colleagues had warned us that it is currently difficult to perform the sort of analysis that could improve performance and accountability in the sector. IATI has done tremendous work in improving aid transparency, but we all agree that improving aid transparency is not something that any single organisation can do by itself. The Simulation demonstrated that the Disberse platform provides a complementary approach that would further improve our collective capability to track aid.
Both this improved tracking and the increased efficiency gained through using the platform are relatively limited when confined to a small number of delivery chains, they scale quickly to deliver network effects that increase these benefits to all delivery chain stakeholders. These benefits are not solely reliant on Disberse’s blockchain architecture, but due to the services built on top of the blockchain.
However we identified that the network effects generated by wider adoption of the platform would create more possibilities to leverage the potential of a distributed ledger to facilitate the distributed governance of aid — something that is currently missing but which is critical for localisation efforts. We hope to explore this — amongst other use cases — in the next Sprint.