When the Clock Is Closed, Instant Payments Wait
For cross-border payments, the hidden question behind speed is not only how quickly a payment instruction is processed, but when the institutions responsible for settling it are able to act. Understanding that relationship lets a reader judge whether an “instant” service is fast end to end or merely fast until it reaches a closed settlement window.
A payment can move through a fast payment system (FPS) very quickly and still wait for a real-time gross settlement (RTGS) system, a correspondent or another connected institution. The reason is simple: these systems operate on different schedules and perform different jobs. FPS are generally built for rapid retail payments and commonly operate around the clock. RTGS systems are important for settling obligations between financial institutions and for supporting the settlement of some retail payments, but their operating hours have traditionally been more limited.
The word “settle” is doing important work here. A system may accept an instruction, check and process it, and report that step back to a provider before the related obligation has reached final settlement in the next system. A sender’s screen can therefore show progress without proving that the receiving side has received funds that it can use. This is not necessarily a failure; it is a difference between the state of an instruction and the state of the money movement behind it. Cross-border speed claims become meaningful only when they say which state they measure.
The clock behind an instant instruction
Imagine a payment that begins in one time zone late at night. The sending provider may accept it immediately, and its domestic FPS may process the instruction without delay. But if the next settlement step depends on an RTGS system that is closed, the payment has reached the edge of local speed. It can be queued until the relevant system opens, or the institutions involved can keep liquidity positioned in advance so the transaction can proceed when settlement is possible. In either case, the time visible to the sender and the time of final settlement are not necessarily the same.
This is why time-zone overlap is a system property, not just a scheduling inconvenience. If two jurisdictions are several hours apart, their settlement systems may be open for only a short period at the same time. Payments then accumulate around the overlap. When the overlap is narrow, banks may need to pre-position liquidity before the payment arrives. That funding has an opportunity cost, and it may be harder to manage when the payment involves several currencies or institutions with different operating calendars. If liquidity is not available, the transaction waits for the next usable window.
Pre-positioning changes the timing problem but does not make it disappear. An institution that holds liquidity ahead of demand may help a payment move during a period when another system is open, but it is tying up resources in anticipation of a flow that may or may not occur. If the surrounding markets are closed, the institution may have fewer ways to replace an unexpected outflow. The design question is consequently two-sided: how much delay is avoided, and what additional balance, staffing and operational exposure is accepted to avoid it?
The 2025 CPMI monitoring survey gives a sense of the gap between “always on” as a concept and the actual settlement environment. The RTGS systems covered by the survey operated for an average of 73 hours per week, compared with 168 hours for continuous operation. Eleven systems reported operating 24/7. Across the weekday period when the greatest number of systems were open together—06:00 to 11:00 Coordinated Universal Time—89% of the surveyed systems were open. That is a substantial overlap, but it is also a reminder that a global payment day is not one uninterrupted window.
The global settlement window is best understood as a shared opportunity, not a guarantee. It identifies the period in which many systems are available together, but it does not mean every corridor uses the same institutions, currencies or cut-off rules. A payment involving systems outside that common period can face a smaller overlap than the headline figure suggests. The figures also describe the surveyed systems, not every payment in the world. Their value is diagnostic: they show why an institution can be open for much of a week while still failing to overlap with the particular foreign market a customer needs.
The timing matters in both directions. A payment sent outside the overlap may wait for the receiving side, while a payment sent during a crowded overlap may encounter a concentration of activity. The BIS brief links limited operating-hour overlap with delayed final settlement, higher costs from pre-positioned liquidity and risks associated with peak-hour concentration. These effects primarily arise between payment-service providers and in wholesale settlement, but they can reach end users as slower or less predictable payments.
A simple timeline makes the mechanism concrete. Suppose the sending FPS remains available through the night, but the receiving RTGS system opens in the early morning of the sender’s time zone. The sender may receive an acceptance message at 23:00. The payment then enters a waiting state rather than a completed state. If the receiving system opens at 06:00, the relevant institution must still have the funds, the payment instructions must remain usable, and the receiving provider must be ready to continue the process. The seven-hour interval is not caused by the sending rail’s processing speed. It is caused by the conditions required for the next settlement step.
A different corridor may have the opposite pattern. Its systems could overlap for several hours, but activity might be concentrated at the beginning of that period because payments that accumulated overnight are released together. That concentration can create operational pressure and increase the importance of liquidity controls. “Open” is therefore necessary but not identical to “able to complete every payment immediately.” Capacity, funding and the sequence of participating institutions still matter.
The diagram below separates domestic processing from settlement completion. It also shows why the same payment can follow two different paths depending on the state of the next system.
The sequence separates two events that are often collapsed into one: domestic processing and settlement completion. The first can be fast in both scenarios. The second depends on the operating condition of the next system. An app can therefore report that an instruction was accepted while the cross-border payment remains in a state of waiting or pending final settlement. A useful service description should make clear whether its time measure ends at acceptance, processing, settlement or funds availability.
Extending the window is a design choice, not a magic switch
The obvious response is to extend RTGS operating hours. Longer hours can enlarge the global settlement window, improve liquidity management and reduce the delay created by time-zone gaps. They can also reduce the amount of activity forced into a peak period. More overlap may support new forms of synchronisation, links between FPS and payment-versus-payment arrangements in which two currencies are settled together. In this sense, the clock is part of the foundation on which more ambitious cross-border services can be built.
The effect is not necessarily proportional to the number of hours added. An extra hour that overlaps with a major partner’s active period can be more valuable than an hour when few relevant institutions are operating. The Bank of England model cited in the BIS brief places particular value on extending the UK RTGS morning, because that creates more overlap with Asia-Pacific time zones. It estimates that a five-hour morning extension would reduce payment delays by about 6% and raise the share of payments settled within one hour from 50% to 60%. A move to 24-hour operation is modelled as reducing delays by about 60% and lifting the share settled within one hour to 75%.
Those estimates illustrate a method of reasoning, not a universal forecast. They are modelled results for a particular setting, and another corridor may have different demand, calendars, liquidity practices and system dependencies. Their important lesson is that an operator should ask which delays are caused by the clock and which extension creates the most useful overlap. A policy that adds hours without connecting them to a real cross-border need may impose cost without delivering a comparable customer benefit.
But “open longer” has real operational consequences. System operators may need additional staffing and upgraded technology. Banks must decide how to staff payment operations, manage liquidity and respond to customer demand on nights, weekends or holidays. If a system remains open while relevant markets are closed, an institution may face an outflow without an immediate way to source liquidity. The benefit of accepting a payment earlier has to be weighed against the cost of holding funds and the risk of operating when the surrounding financial system is less available.
The European T2 example shows why the operating decision can be useful even before a system reaches 24/7. Its extended hours began at 02:30 rather than at the later start used by the former TARGET2 system. The BIS brief reports that the change spread payments across a broader period and reduced the concentration that had previously occurred in the first hour. Payments from Asian banks were particularly concentrated in the early morning European hours, between 03:00 and 07:00. The extension therefore supported a real cross-border need: it created more usable overlap with institutions outside the euro area and eased pressure on the beginning of the European day.
There is also a calendar dimension. Extending a weekday schedule is different from opening on Saturdays, Sundays or holidays. A system may be technically ready to process a payment while the banks, markets or liquidity arrangements around it are not. The BIS brief notes that operators must consider customer demand, staffing, collateral or other exposures and the possibility of outflows on days when markets are closed. A continuous clock can shift responsibility to institutions that previously relied on predictable pauses, so the change needs operating rules as well as a longer timetable.
There is also a question of where the extension produces the most value. The survey reports that many RTGS systems are moving toward near or full 24/7 operation, but it also identifies a perceived lack of business case or demand as a leading barrier. Competing priorities, limited resources and high costs matter. A jurisdiction may rationally choose a targeted extension that covers the most valuable foreign-market overlap rather than pay for continuous operation that customers rarely use. The modelled UK results illustrate this kind of marginal-choice question, but they do not settle it for other systems.
Even a fully open RTGS system would not remove every cross-border dependency. The receiving institution may use a different schedule. A participant may not have direct access. A correspondent may still be needed. The payment data may need conversion, or the transaction may require a control that cannot be completed automatically. Longer hours remove one possible wait; they do not create a missing connection or grant a provider permission to use a system.
That boundary is important for evaluating infrastructure announcements. A move toward 24/7 settlement is meaningful when it increases overlap with the specific systems and institutions in a corridor. It is less meaningful as an isolated headline. The right comparison is not between an old clock and an ideal clock, but between the operating schedule of every necessary participant and the times when people and businesses actually need to send money.
The practical test is therefore a timeline. Identify when the sender’s provider accepts the instruction, when the sending rail processes it, when each settlement system is open, when liquidity is available, and when the receiving provider can complete the final step. If a claim about instant cross-border payments leaves those times unspecified, it has described a fast component rather than the whole service. If an extension is proposed, ask which specific wait it removes, which institutions benefit, and what new staffing, liquidity or risk obligations it creates.
The clock is the first hidden condition because every other arrangement must operate in time. But timing alone does not tell us who may enter the system or who must stand between the payer and the payee. The next question is the access rule: which institutions can participate directly, and what changes when a payment depends on an intermediary?