Pre-funding

Pre-funding is money placed in position before it is needed. When a settlement system does not net obligations against each other and does not extend credit during the day, the only way a payment can go through is if the funds are already sitting there. So banks and payment providers park money in advance, ready to settle, and that money cannot do anything else while it waits.

It appears in two familiar places. One is the nostro account, where a bank holds a funded balance in a foreign currency at a correspondent so that it can make payments in that currency. The other is inside faster-payment schemes, where, as the BIS describes, a receiving provider advances funds to the payee before settlement between providers takes place, and pre-funded positions are one of the tools that cover the resulting credit exposure.

Why it is a cost, not just a mechanic

The Financial Stability Board lists the liquidity cost of pre-funding as one of the components that make cross-border payments expensive, alongside transaction fees, foreign-exchange charges, and fees along the payment chain. The cost is not a fee you can see on a statement. It is the opportunity cost of capital held idle, spread across every currency and every relationship, purely so that payments can settle on systems that do not connect.

The scale becomes visible when you compare a netted system with a gross one. CHIPS, a large US dollar system, settled about $2 trillion in average daily value in 2025 using roughly $96 billion of funding. It estimates that settling the same payments one by one under a real-time gross approach would have required around $442 billion. Its liquidity efficiency, on those figures, is about 26 to 1: each dollar of intraday funding supported roughly $26 of settled value. The difference between $96 billion and $442 billion is a direct measure of how much liquidity netting frees, and how much a fully pre-funded model would tie up.

The direction of travel

Pre-funding is a symptom of settlement that is neither netted, real-time, nor programmable. Netting reduces the funding needed to move a given value. Real-time settlement removes settlement risk but demands liquidity for every gross payment, which is its own version of the problem. Programmable settlement aims at the root of it: if value can settle directly and instantly when its conditions are met, there is far less need to pre-position money against timing risk in the first place. A note of caution belongs here, since some vendor figures for total capital “trapped” in pre-funded accounts circulate without a solid source. The CHIPS comparison above is the firmest published illustration of the effect.