The Prefunding Trap
Cross-border payments commonly rely on prefunded correspondent balances. Public data establish the mechanism and its trade-offs, but do not publish a reproducible global stock or annual carrying-cost total.
By Anton Titov, Founder · Plexo Institute
Correspondent banking uses prefunded balances to make payment instructions executable.
Prefunding lets a correspondent payment settle when the instruction arrives. It also means liquidity is positioned in advance, often across several counterparties and currencies.
Reading Guide
Four ideas to anchor the read: the operational mechanism, the limits of public measurement, the right scope for a calculation, and the realistic limits of any unlock.
The perspective explains the prefunding mechanism, distinguishes it from a global stock estimate, describes the operational trade-offs, and sets out the inputs required for a reproducible corridor or operator calculation.
Public banking and payment datasets do not separate operational payment float from lending, deposits, repo, intragroup funding, stress buffers, or strategic reserves. They also do not provide a public mirror-removal method for nostro assets and vostro liabilities. A global number without those inputs would be false precision.
A corridor calculation must distinguish day-to-day settlement float from intraday buffers, regulatory liquidity, strategic relationship balances, customer deposits, lending, and intragroup funding. Faster settlement may reduce part of the first category; it does not automatically release the others.
Plexo does not make a continent-wide concentration or settlement-cycle claim from public aggregate data. A corridor study can assess its own correspondent access, eligible balances and settlement timing, but it must name the institutions, dates, source records and exclusions. A commercial case study does not establish a regional or global baseline.
Stablecoin or tokenized-deposit settlement can shorten funding cycles in a named corridor. A global release forecast would additionally require a representative population, a clear eligible-balance definition, duplicate removal, and stated regulatory and operational assumptions. None is implied by faster settlement alone.
The Trap
Messaging moves instructions. Prefunding makes those instructions executable.
In correspondent banking, a bank or payment provider can maintain an account at another institution in a foreign currency. A payment instruction then depends on the account arrangement, available liquidity, agreed credit terms, and the parties' operating process. The precise settlement path varies by corridor and institution.
Prefunding is one common way to make an instruction executable: liquidity is positioned before a payment arrives. That liquidity can be operational payment float, but it can also sit alongside intraday buffers, regulatory liquidity, strategic relationship balances, customer funds, lending, or intragroup funding. Public data do not separate those categories into one global prefunding number.
Mirror Accounts Make Messages Executable
A payment instruction clears only because capital is already staged on both sides.
Bank A
Debits sender locally
The customer sees a payment start, but cross-border value has not moved.
SWIFT message
Instruction, not value
The message is executable only because a balance is already waiting elsewhere.
Bank B
Credits supplier locally
The receiver can be paid because the mirror balance was prefunded.
A payment message is an instruction, not by itself the completed transfer of economic value. In a correspondent arrangement, execution depends on the account, liquidity and risk arrangements already agreed by the participants. Prefunding can support that execution, but it is not the only possible design and public data do not show a universal two-sided balance pattern.
The operational question is therefore specific: which balances are required for a named payment service, which are precautionary or regulatory, and which can be reduced without moving risk elsewhere?
How Big Is The Trap
The mechanism is observable; the global stock is not publicly reproducible.
Public sources provide complementary evidence, but none identifies a global population of eligible payment-prefunding balances without overlapping positions or non-payment funding.
| Public evidence | What it establishes | What it cannot establish |
|---|---|---|
BIS Locational Banking Statistics | Cross-border claims and liabilities at an aggregate level | The share that is operational payment prefunding rather than lending, deposits, repo, or intragroup funding |
CPMI correspondent-banking data | Payment-message activity, active correspondents, and corridors | Bilateral account balances, intraday buffers, or mirror removal |
BIS cross-border payment analysis | Prefunding is a real cross-border liquidity-management trade-off | A global prefunded stock or a counterfactual amount released by technology |
Named operator or corridor records | A reproducible calculation when the operator designates eligible balances | A global estimate without documented sampling and weighting |
The practical conclusion is not that every balance is avoidable. It is that a payment design should show which liquidity it needs, why it needs it, and whether a different settlement or netting model can reduce the operational portion without creating a new risk elsewhere.
A range is not reproducible merely because it has a lower and upper bound. The calculation must state the eligible-balance definition, observation period, source population, mirror treatment, exclusions, and counterfactual. Without those, a large number is a thesis label rather than an auditable result.
The Annual Cost
A carrying-cost estimate is valid only at the scope where its inputs are known.
A useful carrying-cost calculation needs an eligible average balance, a currency and period, a reference rate, and a stated counterfactual. Public global datasets do not supply those inputs for payment prefunding as a distinct category.
The calculation starts with a consenting operator or documented corridor.
The rate, currency, tenor, and risk basis must be named.
The mechanism is economically meaningful even without an invented global total: idle nostro and vostro balances can support payment execution while limiting how an operator allocates liquidity elsewhere. The material question is the measured cost in a specified corridor, not an unsupported worldwide multiplication.
A rate is not observed opportunity cost until the balance, currency, duration, risk policy, and alternative use are defined. A named operator or corridor can state those assumptions. A worldwide flat rate cannot turn an undefined stock into a reliable annual cost.
Why The Trap Persists
The trap is a governance gap, not simply a technical limitation.
Plexo's analysis is that the constraint is not only technical. Payment execution also depends on rules for credit, compliance, liquidity, liability, exception handling and dispute resolution. Bilateral arrangements can make those responsibilities clear to the parties, while creating repeated liquidity-management work. Whether a multilateral or tokenised design reduces that work is an empirical question for a named operating model.
The Unlock Fails Where No One Owns The Layer
Each bank can optimize its own nostro, but only shared governance can remove duplicated balances.
Pair A-B
Float staged twice
Each corridor keeps its own safety balance.
Pair A-C
Another mirror
The same institution repeats the buffer elsewhere.
Pair B-C
More trapped stock
The problem is shared, but the accounts are bilateral.
Local optimizer
Can trim one bank's forecast
Better treasury tools reduce one balance sheet, but they do not create a common rulebook.
no cross-bank ownerShared governance layer
Only a neutral netting owner can compress duplicated balances.
A shared clearing design has to allocate governance as well as technology: participant access, credit, finality, compliance, recovery, liability and dispute handling. Existing multilateral arrangements address defined payment or FX use cases; their existence does not prove that one model can replace every correspondent relationship.
A shared ledger can change how records and settlement instructions are coordinated. It does not remove the need for accountable institutions to manage compliance, risk, recovery and disputes.
Participants may retain a familiar correspondent arrangement because it meets their current requirements for access, credit, compliance, resilience and customer service. That does not prove the arrangement is costless, or that every alternative reduces the same balance. A comparison must identify the corridor, the eligible balance, service-level obligations and the risks retained by the alternative.
Who Holds, Who Captures
The economic effect depends on the institution, corridor and balance category.
Public aggregate data do not identify who holds the operational component of prefunding globally or who captures its economic effect. The following roles are a diligence frame, not a claim about a universal market structure.
| Actor | Position | Economic effect |
|---|---|---|
Respondent banks and PSPs | Nostro holders at correspondent banks, often across several currencies and counterparties | Liquidity is positioned in advance and cannot be allocated elsewhere while it remains eligible payment float. |
Correspondent banks | Hosts accounts and provides services under bilateral arrangements | Terms, remuneration and liquidity treatment require account- or corridor-level evidence. |
Payment operators | May maintain working liquidity under their own service and risk rules | A named operator can disclose the eligible balance and funding cycle; it cannot stand in for a global estimate. |
Treasury and operations teams | Manage forecasting, reconciliation and liquidity allocation | The operating burden and its cost must be measured in the relevant institution or corridor. |
De-risking can concentrate correspondent access among fewer counterparties, but a defensible concentration claim needs a named institution, corridor, observation period, and source. Plexo treats this as a corridor-research question, not a global percentage claim.
A commercial case study can illustrate a named operating change, but Plexo does not use a vendor case as evidence for a global prefunding stock or a general release rate. The correct research unit is a documented corridor with its own eligible balances, dates, counterfactual and exclusions.
What Would Unlock It
Three unlock archetypes operate at different scale and timelines.
Three archetypes of unlock exist, each with different scale and timeline. Their effect should be measured at a named corridor or operator level rather than assumed globally.
Different Pools Need Different Unlocks
Stablecoins release operational float first; buffers and strategic reserves need policy and governance change.
Trapped capital pool
Operational prefunding
Daily corridor balances can turn faster when settlement compresses.
Intraday buffers
Risk, cutoff, and Basel assumptions must move before this pool releases.
Strategic reserves
Counterparty access and de-risking politics keep this capital defensive.
Mechanism 1
Fast settlement raises turnover
Less money waits between payment cycles.
Mechanism 2
Netting compresses gross flows
Only net positions need final settlement.
Policy gate
A global release rate needs a defined population
Public data do not isolate one global pool or support a universal release forecast.
| Unlock path | Examples in the source | Timeline and scale |
|---|---|---|
Account and tokenised-money designs | A named operator or payment-system arrangement | Potential effect depends on settlement, credit, access and liquidity rules; no global release rate is implied. |
Regulated stablecoin settlement | A regulated issuer and licensed operating model | May change settlement timing for an eligible flow while retaining reserve, compliance and working-capital requirements. |
Multilateral or PvP arrangements | A defined corridor and participant rulebook | Can alter coordination and settlement risk, but the liquidity effect must be measured against the prior arrangement. |
A meaningful release scenario must state the population, eligible balance, counterfactual, adoption assumptions, and remaining liquidity requirement. It cannot be inferred from technology adoption alone.
Operational settlement float, regulatory buffers, and strategic relationship balances have different release conditions. A credible forecast needs their observed mix in a defined population and should report a corridor-level range with its assumptions rather than a global percentage.
- Operational prefunding: day-to-day settlement float that may respond to netting and faster settlement.
- Intraday and regulatory liquidity: balances held for risk policy or regulatory requirements.
- Strategic relationship balances: liquidity maintained for counterparty access and service resilience.
A new rail may reduce part of the first category. The others require separately evidenced operational or policy changes.
Evidence Boundary
Public sources establish a mechanism and trade-offs, not a global historical stock series.
Correspondent banking, payment-system reform and newer settlement designs have each evolved over time. This perspective does not infer from that history that a single global pool grew, shrank or concentrated in a particular way: public sources do not isolate eligible payment prefunding as a consistent global time series.
A change in correspondent-relationship counts, payment volume or adoption of a new rail does not by itself reveal the eligible prefunding balance. A historical claim would need account-level or operator-designated balances, a stable eligibility rule, mirror removal and a consistent population across dates. Plexo does not publish that series today.
Settlement providers, tokenized-deposit pilots, and policy bodies increasingly discuss liquidity fragmentation. Their materials support a mechanism and a design question, not a public global stock calculation. The next useful evidence is a transparently measured corridor, not a larger headline.
Counter-Arguments & Limitations
The mechanism is real; the global number must remain unclaimed until it is reproducible.
Two objections define the boundary of this perspective. The first is that public aggregate banking data cannot identify a global payment-prefunding stock. The second is that a new settlement design can relocate, rather than eliminate, part of the liquidity requirement.
When Bank A holds a nostro at Bank B, the same economic relationship can appear as an asset and a liability. Adding both sides without a mirror-removal method fabricates a global total. The right response is not to choose a preferred number: it is to calculate only records with a documented counterparty identity and duplicate treatment.
The offset case: regulated stablecoin issuers hold reserves, and licensed operators maintain working capital for settlement timing and risk. A new design can therefore relocate part of the liquidity requirement. The net effect is an empirical question for a named corridor: compare eligible prefunding before and after the change, while recording issuer reserves and operator working capital separately.
About the Author
About This Perspective
Scope, disclosure, and method.
Plexo operates a regulated multi-stablecoin clearing network for cross-border B2B settlement and has an interest in architectures that reduce unnecessary operational prefunding. This perspective therefore separates a source-backed mechanism from an uncalculated global market claim. Any Plexo corridor calculation should disclose its scope, exclusions, duplicate treatment, observation period, and counterfactual.
Public sources support the mechanism and trade-offs of prefunding but do not identify a reproducible global stock. A Plexo calculation is permitted only at operator or corridor level: define the eligible balance, exclude customer deposits, lending, regulatory reserves and intragroup funding, remove nostro/vostro mirrors by counterparty, state the observation period, and name the reference rate and counterfactual before calculating a carrying-cost proxy. A global estimate requires a defined population, representative sample, weights, and duplicate-removal method. This Perspective is not investment, tax, or regulatory advice.
Continue Reading
Cross-Border 2030 - the four-layer replacement architecture and which corridors go where by 2030.
The Unlock - how freed nostro capital re-enters the productive economy.
The $2.5 Trillion Gap - the unmet trade-finance demand driven by related capital lockup.
References
Bank for International Settlements, Cross-border payments.
BIS Committee on Payments and Market Infrastructures, Correspondent banking data.
World Bank, Cross-Border Fast Payments.
BIS Committee on Payments and Market Infrastructures, Access to payment systems.
References
4 references- Cross-border payments — Bank for International Settlements
- Correspondent banking data — BIS Committee on Payments and Market Infrastructures
- Cross-Border Fast Payments — World Bank
- Access to payment systems — BIS Committee on Payments and Market Infrastructures