RTGS vs Net Settlement: How Design Shapes Liquidity and Risk
Gross and net settlement solve the same problem in opposite directions, and the choice decides who funds what, when credit risk appears and how much liquidity a day needs.
Key takeaways
- RTGS settles each payment individually and immediately, so finality is instant but every participant must fund payments as they happen or queue them
- Deferred net settlement nets obligations before paying them, cutting funding needs while leaving an unresolved credit exposure open until the cycle closes
- Netting is leverage on liquidity: CLS settlement members fund roughly 1% of the value they settle each day, according to CLS figures presented by the Bank of Canada
- Finality is a legal property as much as a technical one — in the EU it rests on the Settlement Finality Directive and each system's rulebook
Two designs, one shared goal
Every wholesale payment ends in the same place: funds moving between accounts held at a central bank or a commercial bank. What differs is the order of operations, and that order decides how much cash and collateral participants need to hold.
Clearing works out who owes what. Settlement moves the money that discharges the obligation. Settlement design is the choice of how those transfers are sequenced, batched and finalised — and the BIS Committee on Payments and Market Infrastructures (CPMI) draws the line plainly in its glossary: real-time gross settlement (RTGS) is the continuous settlement of payments on an individual order basis, without netting debits against credits, while deferred net settlement (DNS) settles on a net basis at the end of a predefined cycle (BIS CPMI, 2024 cross-border payments monitoring survey).
That survey counted 166 operational payment systems reported by central banks: 72 RTGS systems, 45 fast payment systems and 49 DNS systems (BIS CPMI). Gross and net are not competing products so much as different answers to the same question: do you want certainty now, or efficiency at the end?
RTGS: settle now, fund now
In an RTGS system, every instruction is settled in full, one by one, continuously through the operating day. Once settled, the transfer is final and irrevocable, and it settles in central bank money — a claim on the central bank rather than on a commercial counterparty.
Scale is the point. The Fedwire Funds Service, the US large-value RTGS system, originated 209.9 million transfers in 2024, moving an average of about $4.5 trillion a day at an average value of $5.4 million per transfer (Federal Reserve Financial Services, Fedwire Funds Service annual statistics). In the euro area, T2 settled an average of 431,067 euro payments per day worth €1,932.8 billion in 2025 (ECB, TARGET Services Annual Report 2025).
Those totals dwarf the balances behind them. That is the design working as intended: the same euro or dollar is recycled many times during a day, so liquidity is stretched by velocity rather than accumulated in full.
The binding constraint is liquidity. Because each payment must be funded as it arrives, participants need positive balances or intraday credit. Systems soften the pressure with queuing and optimisation algorithms that offset pending payments against each other before settling them gross. When the algorithms find no match, payments queue — and in a stressed market a queue can turn into gridlock, with everyone waiting for an inflow that never arrives.
Deferred net settlement: accumulate, net, then settle the residual
A net settlement system works the other way round. Obligations build up over a defined window, are netted bilaterally or multilaterally, and only the resulting net positions are settled in central bank money at the end of the cycle.
Well-known examples are euro wholesale netting through EURO1 and retail automated clearing houses such as SEPA credit transfers in the EU and the US ACH network. In US dollars, CHIPS is a hybrid: it is not a pure gross system, and it releases payment orders individually or through a netting algorithm that offsets them against each other. It shows how much value that offsetting can carry: in 2024 it settled an average of 565,591 payment orders per business day worth roughly $1.8 trillion, an average order of $3.4 million, for 41 participants (Fedwire Funds Service PFMI disclosure).
The trade-off is stark. Netting collapses gross obligations into a much smaller funding figure, but until settlement happens, participants hold unmatured credit exposure to one another. If a member fails before the cycle closes, the netting set may have to be unwound and recalculated. That risk is what motivated the first coordinated international netting standards in the 1990 Lamfalussy report, a lineage the EU's Settlement Finality Directive still cites in its recitals (BIS, Report of the Committee on Interbank Netting Schemes, EUR-Lex, Directive 98/26/EC).
RTGS vs net settlement at a glance
| Feature | RTGS | Deferred net settlement |
|---|---|---|
| Timing | Continuous, payment by payment | Batched into defined settlement cycles |
| Finality | Immediate on settlement, in central bank money | At cycle close, after netting |
| Funding need | Full value of each payment as it settles | Net position only |
| Interbank credit exposure | Short-lived; intraday credit and queues carry their own risk | Open until the cycle closes |
| Member failure | Settled payments are final; unsettled ones queue | The netting set may need to be unwound |
| Typical venues | Fedwire, T2, CHAPS | EURO1, ACH, SEPA credit transfers (CHIPS is a hybrid) |
What netting actually saves
The clearest illustration comes from foreign exchange, where two currencies must change hands at the same moment. CLS reported an average daily settled value of $7.2 trillion in 2024, with a record $19.1 trillion settled on 20 June 2024 (CLS Group, Report and Accounts 2024).
Settling that gross would be impractical at any reasonable cost. Because CLS calculates each member's multilateral net position per currency, the funding required to settle the $19.1 trillion record was about $78 billion, and on a typical day settlement members fund only around 1% of the value of their payment instructions (CLS presentation, Bank of Canada). That is roughly a hundredfold reduction in the cash needed to move a given value.
The lesson generalises beyond FX. Netting is a leverage mechanism applied to liquidity: it lets a system move enormous value on a small fraction of the cash. The price of that leverage is a concentrated, time-boxed exposure that does not exist in the same form in RTGS.
Settlement risk: what each design removes, and what it leaves behind
Cross-currency settlement risk — paying away the currency you sold and not receiving the one you bought — is often called Herstatt risk, after the bank failure that prompted the first international netting standards. CLS addresses it with payment-versus-payment (PvP): both legs are released simultaneously, so neither side is exposed to the other's principal.
That protection does not reach every currency or every participant. CLS covers 18 currencies, and more than 37,000 third-party institutions access it indirectly through settlement members (CLS Group, Report and Accounts 2024). An institution accessing PvP indirectly inherits the risk profile of whichever member stands between it and the system — a credit, liquidity and operational dependency rather than a direct one.
In plain netting, the exposure is credit risk plus liquidity risk: a participant may owe more than expected if a counterparty defaults and the set is recalculated. RTGS removes that waiting exposure but replaces it with the possibility that a participant cannot fund a payment in time — which is why supervisors treat intraday liquidity as a risk category in its own right, alongside settlement finality and credit risk, in the CPSS-IOSCO Principles for Financial Market Infrastructures (BIS, PFMI).
Finality itself is legal, not merely technical. In the EU, the Settlement Finality Directive ensures that transfer orders entering a designated system remain enforceable even if the sender becomes insolvent, and that an order cannot be revoked after the point defined in the system's rules (European Commission, settlement finality). Outside the EU, comparable protections come from different national frameworks. The effect is similar, the legal mechanics are not, so this has to be assessed jurisdiction by jurisdiction rather than assumed.
Where the costs land
For RTGS participants, the binding cost is intraday liquidity. Funding a payment means holding balances or drawing central bank credit, and that credit is priced and collateralised. In the US, the Federal Reserve's Payment System Risk policy sets out the daylight overdraft framework for Fedwire participants (Federal Reserve, PSR policy); pricing and conditions differ by jurisdiction and by institution, so participants should check the rules that apply to them rather than assume a common standard.
Operating hours matter too. A system that closes in the evening compresses settlement into a window, and central banks periodically extend access: the Federal Reserve has announced an expansion of the Fedwire Funds Service and the National Settlement Service to 22 hours a day, six days a week, planned for 2028 or 2029 (Federal Reserve Board, October 2025, Federal Register notice). The CPMI reports that RTGS systems operate an average of around 66 hours per week, with only five of the systems surveyed running 24/7 (BIS CPMI).
For DNS participants, the costs look different: lower funding needs and typically lower per-item fees, offset by end-of-cycle exposure, less transparency during the cycle, and the operational cliff-edge of settlement cut-off times, where a missed deadline pushes an entire cycle to the next day.
Tiering cuts across both models. Smaller banks and non-banks access large-value systems indirectly through a correspondent, saving membership and technology costs but taking on the credit, liquidity and operational risk of that intermediary.
Why the line keeps blurring
Pure categories are becoming rarer. RTGS systems now embed liquidity-saving mechanisms that look a lot like intra-cycle netting, because gross settlement without them would require far more intraday credit than participants want to hold.
Fast payment systems settle retail payments individually with immediate finality, frequently in central bank money, and the EU's instant payments rules push a growing share of retail traffic onto that model — covered in more detail in our piece on the EU Instant Payments Regulation. Securities settlement follows the same pattern, with delivery-versus-payment in central bank money, shorter cycles and a live debate about running hours — see T+1 settlement in the US and Europe.
For retail-facing readers, the practical version of this plumbing sits in payment rails such as ACH, SEPA, FedNow and RTP, where the same gross-versus-net trade-off decides whether a payment arrives in seconds or overnight.
Risks and costs
- Liquidity risk. RTGS requires funding at the moment of settlement; queuing and central bank credit ease the strain but not the cost of holding collateral.
- Credit risk. DNS defers settlement, leaving participants exposed to one another until the cycle closes, with unwind risk if a member fails first.
- Operational risk. Both designs depend on critical infrastructure with narrow operating windows. An RTGS outage stalls payments one by one; a DNS outage stalls the whole cycle.
- Concentration and tiering risk. Indirect participants depend on a correspondent's robustness — and on still being an acceptable client during a stress event.
- Cross-currency residual risk. PvP removes principal risk in covered currencies, but not for currencies outside the arrangement or for institutions accessing it indirectly.
- Legal and jurisdictional risk. Finality protections, insolvency treatment and netting enforceability differ by jurisdiction and warrant local legal review.
- Fee structures. Transaction pricing, membership costs and any intraday credit charges vary widely across systems, and should be checked against the specific scheme's current rulebook rather than assumed.
Bottom line
RTGS and netting are not better or worse — they are differently priced. Gross settlement buys immediate, irreversible certainty and charges for it in liquidity and collateral. Netting buys capital efficiency and charges for it in deferred exposure and unwind risk. Most modern infrastructure now blends the two, running netting logic inside gross settlement to move closer to both. Knowing which mechanism a system relies on, at what moment, and under whose rulebook, is the first step to reading its liquidity and risk profile correctly.
Settlement mechanics described here are operational and educational, not investment or trading advice. Systems, fees and legal frameworks change; check current rulebooks and the rules of the jurisdiction you operate in.
Follow the shortlist
Tracking how market structure changes affect investors? Our global ETF and stock research watchlist is tracked daily. Premium members get the full list, the reasoning behind every entry and updates when entries change.
See premium accessInformation and education only, not investment advice.
Related
T+1 Corporate Actions: Dividends, Rights and Spin-Offs
How next-day settlement moved the ex-date onto the record date in the US, what that means for dividends, rights issues and spin-offs, and the deadlines investors must reset.
MarketsWhat Is a CCP? Clearing, Risk Waterfalls and Retail Trades
A central counterparty stands behind most modern trades. Here is how CCP clearing works, what the default waterfall does in a crisis, and what it means for retail investors and traders.
MarketsCorporate Actions Explained: Dividends, Mergers and Spin-Offs
How dividends, splits, rights issues, mergers, spin-offs and tender offers work — ex-dates, record dates, broker cut-offs, position and tax effects, and the cost of doing nothing.