Illicit financial flows (IFFs) are often discussed in terms of the money that leaves a country but tackling them requires much more than identifying where the money goes. It requires understanding how systems work: where the bottlenecks are, what incentives shape behaviour, which institutions have the information and authority to act, and where seemingly sensible solutions break down in practice.
Over the past eight months, six African country-teams have been working through these questions using the Problem Driven Iterative Adaptation (PDIA) approach. Rather than starting with a predetermined solution, the teams have been working to understand a specific problem in their own context, test their assumptions, and identify practical entry points for change.
Six countries, two different types of problems
Half the cohort is wrestling with some version of the same challenge: money moving through crypto and other virtual assets faster, and less visibly, than the systems meant to track it.
Rwanda - tracing financial flows linked to virtual assets: The Rwandan team is looking at crypto-related fraud and scams, where the speed and cross-border nature of transactions can make it difficult to trace where funds originate and where they ultimately end up.
South Africa - understanding crypto-related outflows: South Africa's team is examining crypto-related outflows, with estimates suggesting that around R648 billion may have left the country through crypto transactions over a six-year period.
Zambia - strengthening the response to crypto-linked financial crime: The Zambian team is focusing on crypto-linked money laundering and terrorism financing. Its work has highlighted gaps in tools, data and coordination between the institutions responsible for identifying and responding to these flows.
The other three teams are chasing illicit flows through more traditional channels: trade, natural-resource exports and cross-border crime where the money is easier to picture but no easier to trace.
Kenya - following the money through trade: The Kenyan team is looking at trade mis-invoicing and the challenges of identifying and addressing discrepancies in the value of goods being traded. Around USD 1.5 billion may have been lost through trade mis-invoicing in a single year which is equivalent to around one and a half times the country's annual health budget.
Namibia - following illicit flows through the fishing sector: Namibia's team is examining illicit financial flows in the fishing sector. Fishing exports are worth close to N$14 billion a year, yet most fishing companies report little or no taxable profit. The team is exploring what this tells us about the movement of value through the sector and the institutional challenges involved in capturing it.
Liberia - tackling drug trafficking: The Liberian team is focusing on the detection, investigation and prosecution of drug trafficking. In 2026, drug seizures were equivalent to around 7% of GDP, highlighting the scale of the challenge. The team has also identified capacity and coordination constraints between different institutions.
The teams are working on very different IFF challenges, but some common lessons are already emerging. Despite the differences between these problems and their nature, the teams are finding that the challenge is rarely simply a lack of policy or regulation. More often, the difficulty lies in how rules, institutions, incentives, information and capabilities interact in practice.
Three lessons from eight months of problem-solving
1. The real test comes when solutions meet reality: capabilities, incentives and people
At this stage, many of the conversations are largely taking place within government institutions. But the actors who will ultimately determine whether many of these solutions work, that is, traders, banks, crypto platforms and fishing companies, are largely still outside the room. This matters because the gap between a new law, regulation or procedure and what happens in practice often comes down to two questions:
Can the organisations and people involved do what is being asked of them?
Do they have a reason to do it?
South Africa's work on crypto flows illustrates the challenge. As the FSCA Deputy Commissioner pointed out at the workshop, there can be a significant gap between what the rules require and what banks are practically able to do. In deciding how to regulate this new area, FSCA concluded that a consultative approach was essential: bringing banks and platforms into the process rather than legislating at arm's length. That instinct is worth holding onto as teams move from designing rules to applying them. Zambia's team is testing something similar: its consultations with virtual-asset service providers (VASPs) are helping to bring those directly affected by potential reforms into the conversation, both to build buy-in and to understand the practical constraints around information-sharing between banks and VASPs.
Namibia's fishing-sector work offers another way of thinking about the same issue. Tax laws and rules already exist, but they haven't on their own been enough to get companies to comply and disclose true ownership. Enforcement and audits will matter, but the team could also explore how to make compliance more binding in a different way. For example, whether access to fishing quotas could be linked to disclosure of beneficial ownership. That would shift the approach from simply asking companies to comply, to giving them a concrete reason to.
The teams are therefore increasingly looking beyond the design of solutions to questions of functionality: Does it work? For whom? Under what conditions? What gets in the way? And what would need to change to make it work better? This also points to the importance of bringing actors beyond government into the process where relevant. Businesses, financial institutions, communities and other actors may experience the problem differently, and may hold knowledge that is essential to understanding why a solution does or does not work.
2. Frontier solutions: Explore the policy design space by building on what already exists
One of the most important lessons has been the value of starting with what is already there, the systems, the capabilities, the relationships and resources.
Kenya's work on trade mis-invoicing is a good example. The team has deliberately looked at how to build on existing agreements and inter-agency structures rather than proposing something entirely new.
Zambia's team applied the same logic: rather than building a new reporting channel for crypto-related suspicious transactions, the team is working with the Financial Intelligence Centre to add crypto-specific fields to the existing reporting platform. And rather than creating a standalone crypto-training programme, the team wants to incorporate a crypto module into the financial investigations curriculum that is already being taught.
Together, these examples point to something like a frontier-solutions approach: rather than designing an entirely new system and hoping it holds up against real-world incentives, teams start from the systems, relationships and capabilities that already exist, which have, in a sense, already been tested against those incentives and make small, deliberate adjustments at the edges of what's there. Each adjustment is a chance to learn something concrete about where the real constraints lie, before committing to a bigger, more expensive redesign. It's a more modest way of innovating, but also a more honest one: it treats the existing system not as an obstacle to be replaced, but as a source of evidence about what will and won't work.
3. Define success and keep revisiting it
The teams have been encouraged to think about what success looks like, but this is not something that can necessarily be settled once, at the beginning of the process.
As teams test their understanding of a problem, they often discover that the real bottlenecks lie somewhere different from where they initially thought. Their understanding sharpens, assumptions get challenged, and new constraints or opportunities emerge. As a result, the definition of meaningful progress may also need to change.
This is an important part of working iteratively. Success is not simply a fixed destination against which teams measure whether they are on track. It is also something teams need to keep testing and refining as they learn more about the problem and the system around it.
For instance, a team might set an early target around identifying the owners behind crypto wallets. That could be a genuine first measure of success until the work reveals that many of those wallets are deliberately structured to hide the real owner behind layers of intermediaries. At that point, the original measure of success may no longer capture what matters. The aim is not to get the definition of success exactly right from the outset. It is to develop a clearer, more useful understanding of success as the team gets closer to the real problem.
The distinction between a solution that exists and one that works is important. A reform may technically be implemented while failing to produce the change it was intended to achieve. Teams therefore need to keep asking not only "Have we implemented the solution?", but "Is it functioning, and is it helping to address the problem?"
Conclusion
Eight months into the process, the teams have moved considerably closer to the problems they are trying to solve. In many cases, this has meant discovering that the problem is more complex and sometimes quite different from how it initially appeared. That is not a sign that the teams have gone off track. It is part of the process.
PDIA is built around precisely this kind of learning: taking action, seeing what happens, testing assumptions and adapting based on what is learned.
The next phase will therefore be less about producing a perfect solution and more about testing emerging ideas against reality. What happens when a proposed change is introduced? Who responds? What works? What does not? Where are the real bottlenecks? And what needs to change in the solution or in the surrounding system for it to function?