The customer wants a completed payment
A business paying a supplier and a household sending money abroad care about how much arrives, when it can be used and what happens if something goes wrong. Correspondent banking supplies relationships that make those services possible where a bank does not directly hold the required local accounts or infrastructure.
The CPMI’s 2016 report examines those relationships and measures to reduce their costs and information problems. It is useful historical context for payment access, not a current count of available routes or a ranking of providers. The important comparison is the complete service, from sender to usable funds at the destination. [1]
The account relationship behind the payment
Correspondent banking lets one bank provide payment and related services to another, often across borders. A nostro account is the respondent bank’s account at another bank; from the account-holding bank’s perspective, the same relationship is a vostro. The labels describe perspective, not two different pools of money. [1][2]
The CPMI’s July 13, 2016 final report remains useful for the mechanism and information problems. Its historical statistics should not be presented as current network coverage. Current transaction volumes or provider performance require separate dated evidence.
Messaging is not settlement
An instruction can arrive before usable funds, and a confirmation can refer to one leg rather than final availability to the customer. Map the sender, correspondents, foreign-exchange provider and beneficiary bank. Identify which institution holds the money at each stage and whose obligation is being discharged.
Analysis: the key operational measures are completion time, exception rate, deductions and usable funds at destination. A faster message format can improve visibility without eliminating funding constraints, screening holds, local cutoffs or incorrect beneficiary details. Technology claims should specify which part of this chain improved.
Why banks prefund
Prefunding supports payments before incoming receipts arrive but ties up and creates exposure to the account-holding bank. A balance that appears excessive on an average day may be needed for peak outgoing flows or a holiday mismatch. Conversely, a large standing buffer can conceal weak forecasting and poor intraday visibility.
Risk-based due diligence for foreign financial-institution correspondent accounts is a separate legal layer. FinCEN materials describe obligations under 31 CFR 1010.610, including enhanced requirements in specified circumstances. A liquidity-efficient route is not automatically an acceptable compliance route. [3]
Compare the full cost of getting usable funds to the recipient
Hypothetical payment comparison using a common currency equivalent: Route A charges $25 plus an assumed foreign-exchange cost of 0.60% on a $10,000 payment, for a visible combined cost of $85. Route B charges $50 plus 0.20%, for $70. The route with the lower advertised transfer fee is more expensive under these assumptions. Actual comparisons also need a common exchange-rate reference, a matching quote time and any intermediary or recipient deductions.
Now assume the recipient must finance the equivalent of $100,000 for two additional days because funds arrive after its supplier deadline. At an assumed 8% simple annual rate on a 365-day basis, that timing difference costs about $43.84 before facility fees. The numbers are illustrative; they show why exchange pricing, deductions and availability time belong in the same assessment.
Analysis: a reliable slower route can be preferable to a nominally fast route with frequent exceptions if the customer can plan around it. A time-critical payment may justify a more expensive service. Report performance by corridor, payment size and relevant business calendar so averages do not hide customers whose funds are delayed.
Worked funding example
Hypothetical: a bank keeps $10 million across several correspondent accounts to support expected flows. At an assumed 4% annual opportunity cost, the gross carrying cost is $400,000 a year before any interest earned on those balances. Reducing average idle balances by $3 million would save $120,000 under the same assumption.
But suppose the reduction causes the bank to miss a settlement cutoff twice a month and incur emergency funding and customer-remediation expense. Compare the full cost and stress performance, not only average balances. A reserve at one correspondent may not be transferable in time to cure a shortfall at another.
Access and resilience can pull the network in different directions
Analysis: combining volume through fewer correspondents can improve pricing and reduce fragmented balances. It can also make the service depend more heavily on a small number of institutions. Maintaining additional routes costs money and operational effort, but a functioning alternative may preserve customer access during an outage or a relationship exit.
For a smaller institution, the tradeoff affects which international services it can offer at a sustainable price. For the end user, it affects the choice of destinations and the predictability of payment completion. Due diligence, data quality and transparent responsibilities help keep a route usable; they should be evaluated alongside service reach rather than treated as a separate product afterthought.
A route-level operating dashboard
Suggested analysis:
Scroll horizontally to see all columns.
| Measure | Why it matters | Interpretation limit |
|---|---|---|
| Peak funding need | Captures intraday timing | Daily net flows can hide large peaks |
| Idle balance by currency | Identifies opportunity cost | Some balances are required contingency buffers |
| Exception aging | Shows stranded payments and customer impact | Separate compliance review from technical repair |
| Route concentration | Tests substitutability | Two brands may use the same upstream bank |
| All-in delivered amount | Captures fees and FX effects | Compare matching amounts, corridors and times |
Measure completion, cost and recoverability together
Compare the incumbent and proposed route using delivered amounts, end-to-end availability, exception handling and realistic peak funding needs. Include rejected payments, returns and the work required to resolve them. A faster instruction acknowledgment alone is not evidence that the recipient has been paid.
Analysis: the case for a new route strengthens when it improves customer outcomes and total economics without introducing an unmanageable dependency. It weakens when lower quoted fees are offset by exchange costs, fragmented or difficult recovery of stranded funds. Changing correspondent relationships, currency conditions or customer volume can alter the result even if the messaging technology stays the same.
Sources
- CPMI, Correspondent banking final report; July 13, 2016SourceBack to text: ↑1↑2
- BIS, On the global retreat of correspondent banks; 2020 researchSourceBack to text: ↑
- FinCEN, correspondent-account due diligence discussion; FIN-2017-A002Official sourceBack to text: ↑