Payments - Introduction / Learning brief
Liquidity saving and gridlock resolution
Your notes
In simple terms / 01
What this means in plain language
Real-time gross settlement systems settle payments one by one, yet still economise on cash. Central queues, bilateral and multilateral offsetting, and liquidity-saving mechanisms let banks settle more payments with less intraday liquidity and break gridlock.
Settling payments one at a time, immediately, is safe but hungry for cash: a bank needs usable money on its settlement account at the exact moment each payment leaves. If every bank waited to receive before it paid, a real-time gross settlement (RTGS) system could freeze even though all its members are solvent — a standstill called gridlock. Modern systems avoid this without giving up settling each payment finally. Instead of sending payments straight out, the system holds unfunded ones in a central queue. It then looks for payments that cancel out: if two banks each owe the other, their payments can settle together, and if a larger circle of banks owe each other in a loop, a set of queued payments can settle all at once. These offsetting techniques are called liquidity-saving mechanisms. They let the system clear far more value than the cash actually on hand, and they keep a waiting circle of banks from locking up.
Complete lesson / 02
Understand the full idea, step by step
Three flatmates each owe another one money — A owes B, B owes C, C owes A — and each insists on being paid before paying. Nobody is broke, every debt is good, and yet no money moves. Payment systems can fall into exactly this trap, at the scale of hundreds of millions. This lesson is about how they get out of it.
Gridlock
Gridlock is the situation where payments sit blocked in a circle: each participant is waiting to receive before it pays, so a set of perfectly good payments cannot settle even though the banks behind them are sound. It is a coordination failure, not a credit failure — which is the clue to the cure. If the payments could all move at once, the circle would clear itself.
The queue is where the intelligence lives
In a real-time gross settlement (RTGS) system, a payment the sender cannot yet fund waits in a central queue rather than being rejected. A naive system would release the queue strictly first-in, first-out. A modern one does something cleverer: it runs offsetting algorithms that search the queue for payments which can settle simultaneously against each other. Bilateral offsetting is the simple case — two banks each have a payment queued to the other, so the pair settles together and each needs liquidity only for the difference. Multilateral offsetting is the powerful case: when several banks owe each other in a loop, the whole set settles at once, and only the net positions need funding.
| Bank Alfa: pays 80,000,000.00, receives 75,000,000.00 | net EUR 5,000,000.00 out |
|---|---|
| Nordbank: pays 70,000,000.00, receives 80,000,000.00 | net EUR 10,000,000.00 in |
| Meridian Bank: pays 75,000,000.00, receives 70,000,000.00 | net EUR 5,000,000.00 out |
| Gross value queued in the circle | EUR 225,000,000.00 |
| Liquidity actually consumed to settle it all | EUR 10,000,000.00 |
If the three payments settle in one simultaneous run, Bank Alfa's account only needs to cover its net EUR 5,000,000.00, Meridian Bank likewise, and Nordbank ends up richer. Each bank's EUR 10,000,000.00 balance is more than enough — EUR 225,000,000.00 of payments settle using a fraction of that value in cash. This is the arithmetic that makes offsetting worth building. Figures illustrative.
How the offsetting run unlocks the circle
- CLEARING
Clearing System Delta's algorithm scans the queue and finds the loop: three payments that, taken together, nearly cancel out.
- VALIDATION
It checks that each bank's settlement account can cover its net position for the set — EUR 5,000,000.00 for Bank Alfa, EUR 5,000,000.00 for Meridian Bank, nothing for Nordbank.
- SETTLEMENT
All three payments settle simultaneously across the accounts at Central Bank Omega. Not netted into one payment — three individual settlements, executed in the same instant.
- LEDGER
Each account moves only by its net: Bank Alfa down 5,000,000.00, Meridian Bank down 5,000,000.00, Nordbank up 10,000,000.00. Every payment is final.
- NOTIFICATION
Each bank sees its payment settled and its inflow arrived. The queue is clear; the circle never knew it was a circle.
COMMON CONFUSION
“Offsetting turns the RTGS into a netting system, so the payments are not really settled individually or finally.”
Each payment in the offsetting run settles individually, in central bank money, with full finality — the algorithm changes when and with what companions a payment settles, not whether the settlement is real. No participant is exposed to another while waiting for a later net settlement. That is the whole trick: the liquidity efficiency of netting without giving up the certainty of gross settlement.
You may be wondering: if only net positions need funding, how is this different from deferred net settlement?
Timing of the risk. In deferred net settlement, obligations pile up for hours and settle later — participants are exposed to each other across that gap, and the system needs defences in case one fails to pay in. In an offsetting run, there is no gap: the moment the algorithm finds a workable set, everything in it settles at once, finally. Liquidity is saved the same way; credit exposure never opens up.
The wider liquidity-saving toolkit
Offsetting sits at the centre of a toolkit that banks use deliberately. Priority classes let a participant mark critical payments — the funding of another settlement system, a time-fixed obligation — so the queue considers them first. Reservation facilities ring-fence a slice of liquidity for a named purpose, so an urgent afternoon payment is not starved by routine morning traffic. Timed payments let treasury release large amounts at chosen moments, smoothing demand on the account. Behind all of it, the liquidity desk watches balance, queued outflows, expected inflows, and collateral headroom in real time, and steps in when a payment lingers near a cut-off.
STRICTLY SPEAKING
Strictly speaking, every real system implements its own algorithms, priority levels, and reservation types, and the sketch here is a teaching model of the common pattern rather than any operator's rulebook. The Principles for Financial Market Infrastructures expect systemically important systems to manage liquidity risk deliberately, and liquidity-saving mechanisms are one of the standard tools for doing so.
FOR NOW, REMEMBER
- Gross settlement is safe but liquidity-hungry; unfunded payments queue rather than fail.
- Gridlock is solvent banks locked in a waiting circle — a coordination failure, not a credit failure.
- Bilateral and multilateral offsetting settle queued payments simultaneously, so each bank funds only its net position while every payment still settles finally.
- Priorities, reservations, and timed payments round out the toolkit; the liquidity desk uses them actively through the day.
TRY IT YOURSELF
At 10:00, Clearing System Delta finds Bank Alfa, Nordbank, and Meridian Bank holding queued payments to each other in a loop, and settles all three simultaneously with each account moving only by its net. Which statement about the settled payments is correct?
You have now seen liquidity from both sides — what it costs and how systems economise on it. The liquidity topic gathers the full picture: intraday credit, collateral, monitoring, and the supervisory expectations behind them.
KEEP GOINGKey takeaways / 03
Three things to remember
- 01
Real-time gross settlement is safe but liquidity-hungry, because each payment needs funds on hand at the moment it settles.
- 02
Unfunded payments wait in a central queue rather than failing, and gridlock is when banks each wait to receive before paying.
- 03
Bilateral and multilateral offsetting settle queued payments against each other, clearing more value with less intraday cash.
Practical use cases / 04
Where you would use this
A liquidity desk uses priority flags, timed payments, and reservations so that critical obligations settle while ordinary traffic competes for the rest.
A settlement system operator runs offsetting algorithms across the central queue to break gridlock and lower the liquidity each participant must hold.
A payments operations team investigates a payment that has sat queued near a cut-off and resolves it by adding liquidity, moving collateral, or re-prioritising.
Worked example / 05
Put the idea into a real situation
Illustrative example: three fictional banks are each short of cash in a real-time gross settlement system. Northreach Bank has a EUR 5,000,000.00 payment queued to Silvergate Bank; Silvergate has a EUR 4,000,000.00 payment queued to Castellan Bank; and Castellan has a EUR 4,500,000.00 payment queued to Northreach. Settled one by one, the first payment cannot move because Northreach lacks EUR 5,000,000.00, and every payment waits on another — gridlock. A liquidity-saving mechanism looks at all three together and settles them simultaneously, so only the net differences need funding: Northreach must find EUR 500,000.00, Castellan must find EUR 500,000.00, and Silvergate receives EUR 1,000,000.00. EUR 13,500,000.00 of gross payments clear while only EUR 1,000,000.00 of cash actually changes hands. The circle that was frozen a moment earlier settles in a single step.
Evidence & review / 07
Evidence & review
General model of queueing and liquidity-saving mechanisms in RTGS systems; each operator's algorithms, priorities, and reservation facilities differ
What this brief simplifies: The three-bank loop and all amounts are illustrative. Real offsetting algorithms handle many participants and partial sets, and real systems combine several mechanisms the lesson only names.
Sources for this brief3
- Official requirement
Principles for financial market infrastructures ↗ — CPMI and IOSCO (Bank for International Settlements) · Principle 7 — liquidity risk
Published by the CPSS (now CPMI) and IOSCO; contains 24 principles plus responsibilities for authorities. This site uses it only for high-level concepts such as settlement finality.
- Market practiceMarch 2003 edition
A glossary of terms used in payments and settlement systems ↗ — CPSS (now CPMI), Bank for International Settlements · Gridlock, queueing, liquidity-saving mechanisms
Terminology has evolved since this edition; newer CPMI publications refine some definitions.
- Simplified educational illustration
Payments Signal editorial teaching models — Payments Signal
Used wherever diagrams, scenarios, figures, or example values are didactic constructions rather than sourced facts; every such use carries a simplifications disclosure. All people, companies, banks, and list entries in examples are fictional.