Why your USD payment routes through 3 correspondent banks

When you check the GPI trail of an emerging-market USD wire, you often see three or even four intermediary banks. This is not inefficiency — it is the consequence of how correspondent banking works.

Your bank holds a USD nostro (foreign-currency) account with at least one US bank. That US bank is the gateway to Fedwire / CHIPS. If your bank has only one USD correspondent, every USD wire passes through it.

The beneficiary bank also has its own USD correspondent in the US. If that correspondent is different from your bank's correspondent, the payment requires a movement from your-bank-correspondent to beneficiary-bank-correspondent — typically over CHIPS.

So the basic minimum is: your bank → your bank's USD correspondent → beneficiary bank's USD correspondent → beneficiary bank. That is four banks.

A fifth hop appears when your bank does not have a direct correspondent relationship with a tier-one US bank and instead uses a regional intermediary. Small banks in some African and Asian markets route via a Middle Eastern correspondent that has the US relationship.

Each hop adds time (sanctions screening at each bank), cost (each correspondent deducts a small fee unless OUR was specified), and risk of error (data quality degrades at each translation).

The trend post-2008 is shrinkage of correspondent networks — large US banks have ended relationships with emerging-market banks deemed high AML risk, forcing those banks into longer chains. This is one of the structural drags on cross-border payment efficiency.

Key takeaways

• Sender bank + sender USD correspondent + beneficiary USD correspondent + beneficiary bank = 4 banks.
• A 5th hop appears when small banks route via a regional intermediary.
• Each hop: more screening time, more potential charges, more error risk.
• De-risking shrinks correspondent networks, lengthening chains.
• GPI exposes every hop with timestamps and charges.

Frequently asked questions