Why Prefunding Persists

None are villains. Each responds rationally to inherited incentives. G-SIBs earn 100-300 bps NIM on $3-5T in nostro/vostro deposits. Central banks prefer visible flows over efficient ones. Compliance vendors earn per-relationship rents. Incumbent PSPs capture spread on the cost stack.

Most payments analysis treats prefunding as a technical problem awaiting a technical solution. Better rails will fix it. This framing explains why prefunding has not been fixed yet, because better rails did not exist. It does not explain why prefunding persists even as better rails appear.

Political economy explains the second half. Once you see who benefits, the pace of reform starts to make sense.

This is Mancur Olson's Logic of Collective Action applied to payments infrastructure. The beneficiaries of reform, meaning everyone who sends a cross-border payment, are large, diffuse, and unorganized. The beneficiaries of the status quo, meaning a few G-SIBs, a few compliance vendors, and a few incumbent PSPs, are concentrated, organized, and well-resourced.

In every political arena where these two groups face each other, the concentrated group wins for longer than the economics would suggest. Payments reform is a particularly stark example because the economics have been against the status quo since roughly 2015.

G-SIB segment reporting aggregates correspondent banking with other wholesale banking lines. The float component is not broken out. This is not accidental. It is defensible accounting but it is also convenient opacity.

Reform advocates cannot cite the exact number. Critics of reform can claim it is small. The information asymmetry favors the incumbent side of the debate.

A central bank in a reactive archetype, meaning most emerging markets, inherited crypto rules written for speculation. Extending those rules to payments would require legislative action, capital budget for new supervision, and political willingness to admit that the existing framework is misclassifying flows.

None of these are zero-cost. The cheapest option for a reactive central bank is to maintain the existing framework and let the informal flows grow unsupervised. This is why reactive-to-permissive transitions lag so far behind when the economics would otherwise favor them.

Each reform bloc has a preferred endpoint. Fintech operators want their licensed rails to win. Central banks want tokenized deposits or CBDCs to win. Corporate treasuries are agnostic but risk-averse. Consumer advocates want the cheapest option but reject unlicensed rails.

These preferences do not compose into a single coalition. The reform side is a loose federation. The status quo side is a tight oligopoly. The asymmetry of organization matters.

A balance-of-payments crisis removes the luxury of slow reform. Nigeria's 2023-2024 FX stress forced the CBN to reverse its 2021 crypto ban. Argentina's 2024-2025 peso dynamics are pushing similar changes.

FX crisis is a particularly effective trigger because it aligns political incentives across factions. The incumbent banking system is visibly failing. The technocratic argument for allowing stablecoin rails becomes politically survivable.

By 2028, 15-25 jurisdictions will have all three forces aligned. This is not a global reform. It is corridor-by-corridor substitution. The $10 trillion prefunding trap does not collapse. It shrinks corridor by corridor as the status quo coalition loses locally.

The aggregate effect over 2026-2030 is a 20-30% reduction in globally prefunded capital. That is the number to watch. The total is still enormous. The trend is what matters.

The argument: This piece treats G-SIB float income as a primary defender of the status quo. In practice, G-SIB treasury operations hedge nostro currency exposure and the "NIM on nostro deposits" framing oversimplifies modern bank funding. Some G-SIB correspondent units are net loss centers when fully allocated for compliance and capital costs. The political-economy story may overweight a revenue stream that is smaller than it appears.

Partially valid. The exact quantum is unknowable from public disclosures, and that opacity is itself part of the argument. But three offsetting points: (1) FSB correspondent banking data shows G-SIBs have selectively grown nostro books in profitable corridors while exiting unprofitable ones, suggesting the float economics are positive enough to defend; (2) tokenized deposit pilots from G-SIBs have universally chosen architectures that preserve bank balance sheet intermediation rather than disintermediate it; (3) even if float NIM is small, the embedded fee revenue and customer relationship value is large. The political defense holds even if the specific float number is debated.