The Prefunding Trap

For an EM bank holding $100M nostro at a 3% deposit rate while its own cost of capital is 15%, the opportunity cost is $12M per year on a single corridor. Liquidity reserves against correspondent exposure and FX risk on USD-denominated balances add further cost. None of this appears as a fee line item, which is why fee comparisons systematically understate the real cost of correspondent banking.

When a bank in Nigeria wants to accept customer instructions for USD payments, it must hold USD balances at its US correspondent. These balances sit idle until customer orders deplete them, after which the bank must replenish.

The capital cost has three components:

1. Opportunity cost: the balance earns correspondent deposit rates, often low, rather than the bank's cost of capital. For an EM bank with 15% cost of capital, holding $100M at a correspondent earning 3% represents $12M per year in opportunity cost.

2. Liquidity reserve: regulators require banks to hold additional liquidity against correspondent exposure, adding to the capital consumed.

3. FX risk: if the bank's home currency depreciates while it holds USD nostro balances, the book value changes. Hedging costs add to the total.

At the industry level, these costs total billions annually. Most of the cost is invisible because it does not appear as a line item on cross-border transaction fees. It is embedded in the capital structure.

Continue reading: Beating Gridlock covers the capital economics in detail.

The technical solution is available. Adoption has been slow because the economic distribution is asymmetric:

• Costs accrue to smaller and emerging-market banks: they hold disproportionately large prefunding relative to their balance sheets, consuming scarce capital.

• Benefits accrue to larger correspondents when they hold the nostro: they earn deposit spread on the prefunded balances and gain relationship leverage.

When smaller banks solve their prefunding problem through multilateral netting or stablecoin settlement, larger correspondents lose deposit float and relationship leverage. The larger banks have less incentive to facilitate the transition. This asymmetry is the core of the trap. It is not about technology. It is about incentives.

The groups with the strongest incentive to unlock prefunding, including smaller banks, emerging-market banks, and SMEs, are the groups with the least power to drive industry-wide change. They can adopt alternatives individually by moving to stablecoin settlement or joining multilateral netting platforms, but they cannot unilaterally change the correspondent banking structure.

The groups with power to change the structure, including large correspondents and G7 banks, have the weakest incentive. They benefit from the current structure and have limited reason to accelerate change.

This is why change happens slowly through external pressure: Basel IV capital charges that make prefunding more expensive for correspondents, regulatory pressure for multilateral platform adoption, and stablecoin operators that peel off corridor-by-corridor. None of these are fast.

Three specific Basel IV changes raise capital costs on correspondent banking:

1. Output floor of 72.5%: constrains internal model advantages, raising capital against trade finance and correspondent banking exposures.

2. New credit conversion factor on off-balance sheet commitments: the source cites a 10% CCF on trade commitments that was historically lower.

3. Standardized approach revisions: raise RWA calculations for many correspondent banking counterparties.

The cumulative effect is that correspondent banking becomes more capital-intensive for the larger banks that provide it. McKinsey estimates this triggers emerging-market corridor repricing and possibly exit from lower-volume corridors.

This is the first time in decades that the spread-earning side of the prefunding trap has material reason to reduce exposure. The consequence: alternative rails such as stablecoin settlement and multilateral netting gain addressable market that was previously defended by incumbent economics.

1. Identify who bears the cost and who earns the spread. Until the balance-holding side has incentive to change, the system is stable despite the inefficiency.

2. External pressure accelerates faster than internal reform. Basel IV, not voluntary industry action, is unlocking progress. This is the pattern across most payment reforms, including Dodd-Frank on interchange and PSD2 on open banking.

3. Alternative rails matter. Stablecoin settlement and multilateral netting do not wait for incumbents to reform. They take corridor share incrementally, shifting the economic balance over time.

The prefunding trap will resolve. The timeline depends on the rate of Basel IV implementation, stablecoin network maturity, and multilateral platform scaling. It is a multi-year process, not a sudden shift.

Continue reading: The Routing Error covers the structural inefficiency.

The argument: prefunded nostros are not idle inefficiency. They are operational liquidity that guarantees customer payments clear regardless of intraday settlement timing. Eliminate prefunding and the system loses settlement certainty: payments fail when correspondent intraday lines are exhausted, customer instructions queue rather than execute, and operational risk increases. The trap framing ignores the function the float performs.

Valid critique of pure prefunding elimination, weak as a defense of the current scale. The piece does not argue for zero float. It argues that $10T is far above the operational minimum required for settlement certainty. Multilateral netting and atomic PvP settlement preserve certainty at materially lower float requirements: CLS clears $5T daily with ~$4B of actual movement, a 96% efficiency benchmark in the source framing. The same architectural primitives applied to stablecoin clearing would compress the $10T figure by 50-80% without compromising settlement certainty. The function is real; the scale is not.