A country’s anti-money laundering (AML) spending may say more about who lives next door than about the risks inside its own borders, according to new research from Napier AI.
The Napier AI / AML Index 2026–2027, produced with GlobalData, ranks 42 markets. Its central finding is that nations bordering heavily sanctioned states spend more on AML compliance than their domestic risk would justify. Napier AI chief data scientist Dr. Janet Bastiman calls this pattern the “border line of financial crime”. Neighbouring systems absorb risk that crosses frontiers and digital payment corridors, through predicate crimes, illicit cash flows and pressure on correspondent banking relationships.
The wider picture shows some progress. Napier AI estimates that $5.09tn, or 4.30% of global GDP, was laundered in 2025. That is 3.84% lower than the previous year and the first decline in the Index’s three editions. The cost of fighting it, however, is far from evenly spread.
Napier AI chief data scientist Dr. Janet Bastiman said, “When we began looking at the outliers (the countries spending disproportionately. relative to their economic size or apparent domestic risk) geography kept emerging as the explanatory variable. It is not that these nations have worse institutions or weaker political will. They are simply closer to the source of the problem.”
The Index measures compliance cost as a ratio of estimated money laundering losses, with an ideal band of 1% to 3%. Poland, which borders both Russia and Belarus, sits furthest outside it at 40.40%, the highest ratio recorded, although that figure has fallen for three consecutive years. France (22.60%), the Netherlands (20.35%) and Germany (18.12%) also overspend significantly, even though their geography and voting records resemble those of lower-spending peers. Napier AI therefore treats proximity and political alignment as separate factors.
Alignment produces a counterintuitive result. Napier AI measures it through UN General Assembly voting against sanctioned states, and it found a weak positive correlation between closer alignment and overspend. The report attributes this to sanctions lag and institutional inertia: compliance budgets are built around earlier sanctions regimes and are slow to unwind. The US shows the reverse pattern. Despite the sharpest political disagreement with the states it sanctions, its 9.10% ratio is lower than the model predicts, though still above the ideal band.
For compliance teams, the implication is that exposure now depends on neighbours and allies as well as on domestic controls.
Dr. Bastiman said, “If you’re a bank operating in a country that is two borders from a sanctioned state, you may be carrying risk that your current framework isn’t calibrated to identify,” Dr. Bastiman says. “The conventional model treats each jurisdiction as a self-contained unit. The data suggests that it is no longer sufficient, with second-order effects hopping borders and disseminating financial crime risk further than historically anticipated.”
Napier AI urges leaders to examine three areas: payment corridor exposure, whether their models treat geopolitical adjacency as a risk variable, and how quickly screening can adapt to a changing sanctions map. It estimates potential AI-driven compliance savings across the 42 markets at $143.08bn. In its interviews, 68% of experts said their AML engines are ready for AI, but most have yet to realise those savings.
Because spend lags politics, the border line will keep moving. A single new sanctions package could push a market into the high-spend zone without any change in its own policy.
Dr. Bastiman said, “Geopolitical risk has always had financial crime consequences. What we can now show is that those consequences are measurable, mappable, and to some extent, predictable. The borderline shifts. The question is whether institutions and regulators are looking at the same map, and if they can reflect that shift into financial crime defenses as quickly as the risk fluctuates.”
For more, read the full report here.
Copyright © 2026 FinTech Global
Copyright © 2026 RegTech Analyst





