When we started working with the UK Department for International Development (DFID) and UN Office for the Coordination of Humanitarian Affairs (OCHA) on this pilot, we needed to answer the critical question for any business: does anybody really need our service? This was the easy part — of course everybody needs a better way to track aid! We identified the specific needs of donors, UN agencies and NGOs in Sprint One, but we knew that the really hard work would come in Sprint Two, when we needed to collect the data to take the pilot forward.
At the end of Sprint Two we breathed a sigh of relief, because the data collection had not been as difficult as we expected. This was thanks to the excellent cooperation of DFID and UN colleagues, and the offices of 40 implementing organisations working with two of the Country Based Pooled Funds (CBPFs), Iraq (IRQ) and the occupied Palestinian Territories (OPT). We drew on existing OCHA databases, but also gathered more detailed transaction data directly from those implementing partners for the period 2018–19, which enabled us to paint a more comprehensive picture of their delivery chains.
These chains are highlighted in yellow in Figure 1 (below), a Sankey diagram that illustrates the flow of funds from DFID and other donors to implementing partners via the CBPFs; in this example, the yellow flows are those where we were able to get detailed data on financial transactions.
The next question was: how do we present this data and our subsequent analysis, in a useful format? Although you can see that our dataset is incomplete — the data collection was a voluntary exercise, and many implementing organisations were simply too busy to respond — when we presented our findings to DFID and OCHA colleagues, it became clear that nobody had ever collected this detailed data previously, and that it had never been presented in this way. This confirmed two of our beliefs:
- the data does exist to enable us to reconstruct entire delivery chains, but it is fragmented between multiple stakeholders in different countries, using different systems and working with different financial institutions;
- such reconstruction provides a more granular picture of delivery chains than currently possible.
Clearly there is a potential role for an innovative service like Disberse — not just in terms of improving the efficiency and transparency of aid (as we discussed in our previous blog post on transparency), but also in terms of enhancing the analytic capabilities of the aid industry. For example, we were able to pinpoint bottlenecks for liquidity in specific delivery chains which would cause delays in aid reaching the end recipients. Bringing as many of these transactions as possible onto one platform would make possible better analysis — and therefore better decisions.
The service we are piloting could help make this possible, in two ways. First, Disberse is not a tracking system but a financial institution; data would reflect actual transactions in real time, requiring no additional reporting from over-stretched staff. Second, a distributed ledger could allow all this data to continue to be managed by the responsible institutions, but more easily shared between them to improve the collective intelligence of the aid industry. This would also inform and complement existing systems such as the International Aid Transparency Initiative (IATI) and OCHA’s Grant Management System (GMS).
In our last blog post we talked about how our approach might improve transparency, and some of the issues that that raises. The other key metric we are looking at for this Pilot is efficiency, and specifically what sort of friction exists within the delivery chain, in terms of both time and money.
Figure 2 (above) shows a “typical” delivery chain, with funds flowing from the donor (in this case DFID) via the Country Based Pooled Funds (CBPFs) managed by OCHA, to contracting and implementing organisations, and eventually to a vendor. The bar below the chain shows the time it takes for funds to move through each link in the chain — and you can see that the total time is quite significant.
From start to finish, we found that the average time it takes for money to travel from donor sign-off to implementing organisation expenditure is around 97 days. Why does it take this long? We identified three sources of friction in the delivery chain:
- Financial transactions through financial institutions can take up to seven days;
- Every organisation has a certain handling time when they process a payment;
- Most organisations also have a holding time before funds move to the next link.
Each source of friction has a reasonable explanation. For example, financial institutions cannot change their internal systems just for aid flows, and so financial transactions are the same for aid as for everybody else. Handling time is often the result of the due caution that aid organisations must exercise in handling public funds; and holding time is simply the result of the fact that funds usually do not need to be spent as soon as they are received.
We therefore emphasise that these inefficiencies are not the fault of any single organisation or group of organisations. This is a systemic problem — not just the aid system but the financial system that we use to deliver this assistance — and no single organisation has ever had oversight of that system until now. One of the goals of this Pilot, however, is to show what is possible if we can build a more complete picture of the system.
We can illustrate this with another set of data points from our research: the financial risk made visible along the delivery chain through exchange rates. Figure 3 (below) visualises this, showing how using different currency pairs at different points in time can increase the exchange rate losses for a given project; while all countries face such risks, we found that implementing partners working in the Iraqi Dinar face a higher risk to their funding than others — up to + or -9.2%.
Of course stakeholders in the delivery chain are only able to see their own links in that chain. Without time-intensive retrospective research of the kind we conducted, none can see along the entire delivery chain, and therefore they cannot see these patterns; and since nobody realises the problem, nobody can identify potential solutions. If such problems go unaddressed, a significant proportion of aid funds can be lost — not because of corruption, but because of financial risks that can in fact be mitigated through collective action. A platform that identifies such problems and offers risk mitigation services could then help donors to rationalise their transactions, minimising exchange rate losses while increasing the predictability of funding for implementing organisations.
The data that enabled us to identify these issues will now become the seed data for the Simulation Exercise which we’ll perform in Sprint Three. Some of the findings above reflect the benchmarks that we’ve drawn from that data, benchmarks which describe a “business as usual” scenario. We’ll run the historical data again on the Disberse platform, and then measure our performance against those benchmarks.
Only experiments like this will enable us to answer the key question: how much can innovative financial technology, built on a distributed ledger platform, improve the efficiency and transparency of aid? Once we’ve answered those questions, and built confidence with the Pilot stakeholders, we can explore the next question: how might a major aid organisation like DFID test this technology with live transactions?