Intercompany settlement: why it is slow and what fixes it

Intercompany settlement is the movement of money that discharges obligations between entities of the same corporate group. It is the payment flow a multinational controls end to end on paper, both counterparties, both sets of books, both bank mandates, and it is routinely one of the slowest, most manual flows the group runs. Paying a stranger and paying your own subsidiary use the same rails, and the rails do not know the difference.

Why paying yourself crosses someone else’s plumbing

Picture a German subsidiary settling an intercompany invoice with an affiliate in Singapore. The group owns both entities. The payment still leaves a German bank account, enters the correspondent chain, passes through whatever intermediaries connect those two banking systems, clears sanctions screening at each hop, converts currency somewhere in the middle, and lands in Singapore against a cut-off clock set in another time zone. Each intermediary moves balances across nostro and vostro accounts it maintains for the purpose, and each takes time and fees for doing so.

Nothing in that chain is aware that the sender and receiver share a parent. The group owns both ends of the payment and none of the middle. And the middle has been thinning: the BIS has documented a decline of about 22% in active correspondent relationships between 2011 and 2019, which concentrates flows through fewer intermediaries and does nothing to make them faster. The official improvement program is behind schedule too; in October 2025 the FSB concluded the G20’s end-2027 cross-border targets are unlikely to be met at the global level.

If the group routes internal flows through a treasury center currency, the cost compounds: a peso obligation settled through a dollar treasury center can convert twice, paying spread both times, on a flow that never left the group.

What it actually costs

Trapped working capital. Cash in transit between subsidiaries belongs to the group and is usable by no one. Because internal settlement is slow and its arrival time uncertain, entities hold local buffers sized for the worst case, pre-funding positions that faster settlement would make unnecessary. Multiply a few days of float across every internal corridor and the group is financing its own plumbing.

Close friction. Month-end reconciliation chases money that has left one entity’s ledger and not arrived on the other’s. Intercompany accounts that should net to zero across the group do not, until someone finds the payments still in flight and the fees deducted en route.

FX leakage. Internal obligations in different currencies convert through external markets at external spreads, sometimes twice, and the timing of conversion is dictated by payment mechanics rather than treasury’s view of the rate.

The ladder of fixes

Groups attack the problem in three escalating steps.

1. Netting. A multilateral netting program collapses hundreds of intercompany invoices into one net payment per entity per cycle. It shrinks the number and size of settlements dramatically, and we cover it in full in the netting guide. What it does not change: the residual net payments still cross the external chain, and anything urgent bypasses the calendar.

2. The in-house bank. Subsidiaries hold accounts on the group’s own books, and internal obligations settle as book entries, instantly. This is the strongest classical answer, and the most demanding: it needs real treasury infrastructure, intercompany loan documentation, and jurisdiction-by-jurisdiction regulatory and tax analysis. Entities in exchange-controlled countries often cannot join, so the in-house bank covers the easy geography and the hard geography stays on the old rails.

3. A shared settlement layer. The structural version of the same idea: give the whole group one settlement fabric, so that an internal obligation settles directly between entities, in seconds, with finality both sides can see, whatever corridor it crosses. This is where programmable rails enter, and it matters that they do not have to mean digital assets. A settlement layer can move value over existing bank rails, tokenized deposits, or stablecoins as corridor and policy dictate.

Each rung keeps the ones below it. A group with a netting center and an in-house bank still benefits when the settlement leg underneath both gets faster and verifiable.

Where Frame fits

Frame is the settlement layer for global finance: one integration to orchestrate payments at scale across fiat rails, stablecoins, and tokenized deposits, with compliance enforced on every transaction and settlement in seconds instead of days.

For enterprise treasury, Frame works as an overlay on existing bank rails, and no stablecoin is required. Intercompany flows are the natural first use: the counterparties trust each other completely, the pain is measured in the group’s own trapped cash, and nothing about the fix requires changing banks. Frame routes each internal settlement across whichever rail fits the corridor and the policy that governs it, and every settled transaction produces verifiable evidence that its conditions were met, which is what both the netting center and the auditors wanted from the old process and rarely got. Settlement in seconds turns the working-capital buffer question from “how much float do we carry” into “why are we carrying float at all.”

The companion page, settlement for enterprise treasury, covers what treasurers should evaluate in any settlement infrastructure, this one included.