FedNow’s first demand on banks was a weekend treasury desk
The Federal Reserve’s 2019 FedNow plan promised round-the-clock settlement but left banks without weekend tools to move liquidity. Its hardest demand was staffing and automating treasury, accounting and fraud work at hours many banks had never covered.
Looking Back revisits past weeks one at a time, up to the most recent, and asks what it meant for the banks, health systems and public agencies we serve. The series was written in 2026, with the benefit of hindsight, and each article says which week it looks back on.
Two-minute summaryThe dek and the first sentence of each section, quoted from the article
The Federal Reserve’s 2019 FedNow plan promised round-the-clock settlement but left banks without weekend tools to move liquidity. Its hardest demand was staffing and automating treasury, accounting and fraud work at hours many banks had never covered.
- The week On August 5, 2019, the Federal Reserve Board announced that the Federal Reserve Banks would develop a new round-the-clock real-time payment and settlement service called the FedNow Service.
- Liquidity with the window shut Real-time gross settlement means each payment settles individually, at once, in central bank money.
- A ledger that closes every day The notice’s second operational change drew less comment but reached further.
- Monitoring at all hours The notice’s warning about monitoring was direct.
- Fraud and exceptions in the small hours Instant payments change the economics of fraud response.
- Who carries the cost The operational demands fall unevenly.
- With hindsight The service arrived on the Federal Reserve’s published timetable.
The week
On August 5, 2019, the Federal Reserve Board announced that the Federal Reserve Banks would develop a new round-the-clock real-time payment and settlement service called the FedNow Service. The Board’s press release said faster payment services, which move funds near-instantly day and night, on weekends and weekdays, could become widely used. It said the Board anticipated the service would be available in 2023 or 2024. The Federal Register notice added that nationwide reach, rather than initial availability, was a critical measure of success, and that reach would take longer.
The formal notice appeared in the Federal Register on August 9, 2019, as “Federal Reserve Actions To Support Interbank Settlement of Faster Payments” (84 FR 39297, Docket OP-1670). It set out the Board’s determination that the Reserve Banks should build an interbank 24x7x365 real-time gross settlement service with integrated clearing functionality. Comments on the proposed design were due by November 7, 2019.
The decision followed a 2018 notice that had asked whether the Federal Reserve should build such a service. According to the 2019 notice, more than 350 commenters addressed the question. Approximately 320, from all segments, supported a Federal Reserve service. Approximately 30, mostly large banks and private-sector operators, did not. The press release put support at over 90 percent of those taking a position.
The second half of the announcement got less attention. The Board said it intended to explore expanding the hours of the Fedwire Funds Service and the National Settlement Service, up to 24x7x365. The stated purpose was to support liquidity management in private-sector real-time services and a wide range of other payment activity. The Board said it would seek comment separately, after more analysis of operational, risk and policy questions.
The notice explained why liquidity was the problem. It said the Federal Reserve did not then offer a service that let banks move liquidity as needed, in particular on weekends and holidays, to support real-time settlement. Banks could hold extra funds in the accounts that support faster payments, but the notice called that prefunding potentially costly. Prefunding also required predicting the number and value of future customer payments.
The notice noted that a private-sector operator had asked Federal Reserve staff in 2016 to open a joint account for its proposed faster payment system. It also described the hours that applied at the time. The Fedwire Funds Service business day began at 9:00 p.m. eastern time on the preceding calendar day and ended at 6:30 p.m., Monday through Friday, excluding holidays. The National Settlement Service ran from 7:30 a.m. to 5:30 p.m. on the same days.
Two operational changes followed from a service that never closes. The notice said end-of-day balances for each master account would be calculated on every calendar day, weekends and holidays included, under what it called a seven-day accounting regime. And it warned that participating banks might need to adjust internal account monitoring to manage intraday liquidity, and to avoid a negative balance at the service’s closing time. It added that, at least initially, discount window loan originations would likely not be available on weekends and holidays.
The Board’s 2019 announcement framed FedNow largely around access: a public service that would let banks of every size offer instant payments. That story is accurate, and our article on the week of July 17, 2023 covers the core-processor decisions it forced on community banks and credit unions. This article is about something the 2019 notice said almost in passing.
A payment service that never closes needs a bank that never closes, at least in the parts of the bank that manage money and risk. The notice described a world in which payments settle in master accounts at any hour, end-of-day balances are struck every calendar day, and liquidity cannot be topped up through the usual channels on weekends. In that world, treasury, accounting, operations and fraud teams may be on duty at hours they never covered before. Our argument is that FedNow’s first and hardest demand on banks was operational staffing and automation, not technology or product design.
The demand breaks into five parts:
- Liquidity with the window shut: funding settlement accounts when the usual tools to move money are closed.
- A ledger that closes every day: seven-day accounting and what it does to reconciliation and reporting.
- Monitoring at all hours: watching the master account in real time, and deciding what people do and what machines do.
- Fraud and exceptions in the small hours: responding to irrevocable payments when the usual teams are off shift.
- Who carries the cost: how the burden differs by type of institution, and the case that it was overstated.
The first four are distinct operational functions, each with its own owners. The fifth asks who pays for them, and whether the argument holds.
Liquidity with the window shut
Real-time gross settlement means each payment settles individually, at once, in central bank money. For a bank, every outgoing instant payment draws on its balance immediately. If the balance is not there, the payment fails.
The weekend gap
In 2019, the tools banks used to move money into a Federal Reserve account ran on weekdays. The Fedwire Funds Service ran from the evening before each business day to 6:30 p.m., Monday through Friday. The National Settlement Service ran during business hours on weekdays. The notice said plainly that no Federal Reserve service then let banks move liquidity as needed on weekends and holidays.
A service that runs 24x7x365 therefore created a gap. From the Friday close of Fedwire to its reopening, a bank’s instant payment activity would draw on whatever balance it held at the close. Large weekend outflows, such as payroll or a run of consumer transfers, could exhaust it.
Prefunding and its price
The obvious answer is to hold more. The notice described exactly that: banks can increase the funds in accounts that support faster payments to cover future transactions. It also named the two costs. Prefunded money cannot be used for anything else, and the amount needed depends on predicting customer payments, which is uncertain.
Those costs fall on treasury. Someone has to forecast weekend and holiday flows, set a buffer, decide how much idle balance the bank will carry, and revisit the figure as volumes grow. That is a forecasting and decision task, repeated every week, that did not exist for a bank whose payment activity stopped on Friday evening.
The liquidity tool
The 2019 notice discussed a liquidity management tool that would allow transfers between accounts at the Reserve Banks around the clock, either through expanded hours of existing services or a new service. It reported that many commenters wanted such a tool to be automated, so that transfers could happen outside business hours without operational staff at the banks.
That comment is the article’s argument in a sentence. Those commenters were saying that a tool requiring a person to act at 2 a.m. on a Sunday solves the funding problem by creating a staffing problem. Automation was the preferred answer, but automation still needs rules: thresholds, sources of funds, approval limits and exceptions. Someone has to write, test and own those rules.
Two settlement models, two funding routines
The notice described two ways to settle instant payments in central bank money. In a private-sector service built on a joint account at a Reserve Bank, settlement happens on the operator’s own ledger. Each participant must keep enough funds in the joint account to cover its obligations around the clock. FedNow would instead settle directly in banks’ master accounts.
A bank in both models has two pools to keep funded, each on its own schedule. The notice described a liquidity tool that could move funds between a joint account and banks’ master accounts at any time, and noted that a private operator could let participants shift funds among themselves on its ledger over a weekend. Either way, someone at the bank decides how much sits in each pool and when to move it.
The notice also observed that many banks had historically wanted connections to more than one service for contingency purposes. Every extra connection is resilience, and every extra connection is another balance to watch outside business hours.
Thin networks, lumpy flows
The notice expected nationwide reach to take longer than launch. In the early years, a bank on FedNow would exchange payments with relatively few counterparties. Thin networks make flows lumpy: a single corporate client or a single large counterparty can dominate a weekend.
Lumpy flows are hard to forecast, and hard-to-forecast flows push buffers up. A bank joining early should expect its first buffers to be conservative and its forecasts to improve only as volumes grow. That argues for reviewing the buffer often in the first months, with the review owned by treasury rather than the payments product team.
Stress on a weekend
A buffer sized for normal weeks leaves out the harder case, which is a weekend when something goes wrong: a rumor about the bank, a large corporate client moving funds, or an operational outage at another institution that sends payments elsewhere. The Federal Reserve’s 2024 proposal on longer Fedwire hours, discussed in hindsight, explicitly raised the risk that transferring funds to meet rapid deposit outflows on weekends and holidays could worsen liquidity problems for a bank in crisis.
A bank that offers instant payments should therefore treat a weekend outflow as a scenario in its liquidity stress testing. The questions are practical. How large an outflow could the buffer absorb, who would see it first, and what could they do before Monday? A plan that depends on weekday tools answers the last question with “very little”.
Intraday credit and its limits
The notice said the Board was considering making intraday credit, also called daylight overdrafts, available around the clock for FedNow on the same terms as other Federal Reserve services. That would cushion timing mismatches within a day. But it did not remove the obligation to end each day with a positive balance, and the notice said that expectation would apply over weekends and holidays.
The notice also said the discount window, the Federal Reserve’s overnight lending program, would likely not originate loans on weekends and holidays, at least at first. A bank short of funds on a Saturday could not count on the usual backstop. Its weekend plan had to rest on its own buffer and on whatever transfer tools existed.
What this asks of treasury
For a treasury function used to a five-day week, the practical list is long. It needs a weekend and holiday cash forecast for instant payment flows. It needs a buffer policy approved at the right level, and a rule for when and how to move funds once a transfer tool is available.
It also needs an escalation path for a balance approaching zero, with someone authorized to act. And it needs a review after each long weekend to adjust the buffer. Each item is simple in isolation. Together, it is a standing weekend duty for a function that had none.
A ledger that closes every day
The notice’s second operational change drew less comment but reached further. FedNow master accounts would have end-of-day balances calculated on every calendar day. The notice called this a seven-day accounting regime and contrasted it with the five-day regime the Federal Reserve then used.
Why the accounting changes
Under the five-day regime, end-of-day balances were recorded for each banking day, Monday through Friday. Because Federal Reserve services did not operate on weekends or holidays, there were no weekend balances to record. The notice explained that the Board considered a five-day regime for FedNow but determined, under accounting principles, that a seven-day regime was appropriate.
The notice did not impose that regime on banks’ own books. It said participating banks could use other accounting approaches internally, for example five-day accounting that records weekend transactions as occurring on Monday, and could choose to receive weekend and holiday reports on the next business day. That flexibility moves the decision to the bank. A bank that reconciles to the master account every calendar day takes on seven daily balances a week. A bank that keeps a five-day ledger needs a documented method for bridging its books to seven Federal Reserve balances.
Reconciliation and reporting
Daily reconciliation is a control. Breaks, meaning differences between the bank’s records and the Federal Reserve’s, are investigated and resolved. If balances are struck daily but reconciled only on Monday, weekend breaks wait until Monday. Instant payments are irrevocable, so a break caused by an error or a fraud is harder to recover the longer it waits.
The notice also listed the end-of-day balance as the basis for reserve and interest calculations, transaction reporting and account reconciliation. Each of those processes had to accept weekend data. A process that assumed five days of input, written years before, is the kind of thing that fails silently the first time it meets a Sunday balance.
Interest and reserve calculations
The notice tied end-of-day balances to reserve and interest calculations, as well as transaction reporting and reconciliation. Under a seven-day regime, weekend and holiday balances become inputs in their own right instead of being carried from Friday.
For a bank, that touches treasury forecasting and internal pricing of funds. If weekend balances move, interest earned on them moves too, and models that assume a flat weekend will drift. The correction is small for a bank with small flows. It is worth checking before flows grow, because the error builds up in places nobody reviews.
Calendars hidden in systems
Bank systems can carry a business-day calendar: interest accrual, statement cycles, regulatory report generation, cash forecasting and batch schedules. Each was written for a world in which money stopped moving on weekends and holidays. A seven-day ledger at the Federal Reserve does not change those calendars, but it exposes every place they meet weekend data.
Finding those places is inventory work. A bank preparing for instant payments needs a list of systems that consume master account data or payment postings, with a note of how each handles a non-business day. The ones that drop, defer or double-count weekend entries need fixing or a documented workaround before the first live weekend.
Weekday holidays
Weekends get most of the attention, but weekday holidays are harder. In 2019, Fedwire and NSS were closed on designated holidays. FedNow would run on those days like any other.
A holiday Monday turns a two-day gap into a three-day one for any weekday tool, and holiday staffing may be thinner still. A bank’s instant payment calendar should list every weekday holiday in the year, with the buffer, roster and reconciliation plan for each. Those are the days a plan built only for ordinary weekends is most likely to fail.
Evidence for auditors and examiners
Reconciliation is a control that auditors and examiners can be expected to test. A bank that moves to seven daily balances should expect questions about how weekend balances are reconciled, who reviews the exceptions and how quickly breaks are cleared. The evidence is ordinary: reconciliation reports for each calendar day, an exceptions log with dates and owners, and an aging report for open breaks.
Producing that evidence for weekends is the test of whether the operating model is real. If the reports exist only for business days, the seven-day ledger is being reconciled on a five-day rhythm, whatever the policy says.
Who does the work
Where accounting and reconciliation teams work only on weekdays, seven-day accounting does not automatically mean seven-day staffing; much reconciliation can be automated, with exceptions held for the next business day. But the decision about which breaks can wait and which cannot is a policy choice.
A sensible rule, offered as good practice rather than a requirement, separates breaks by value and by type. A small timing difference can wait. A break that suggests a duplicate payment, a missing settlement or unexplained outflows needs a person the same day. That person has to exist on the weekend roster.
Monitoring at all hours
The notice’s warning about monitoring was direct. Account balance management would become more complex where payments settle continuously in master accounts, and banks might need to adjust internal account monitoring. Banks would need to monitor transactions in real time, or hold enough funding to cover payments arising shortly before the service’s closing.
What has to be watched
Three things need watching around the clock once a bank sends and receives instant payments. The first is the master account balance against expected flows and the buffer. The second is the health of the connection to the service and of the bank’s own payment processing, since a stalled queue looks like a liquidity problem from outside. The third is the pattern of payments, for fraud and for errors.
A bank may already monitor systems around the clock for availability, for example through a technology operations center. What changed is that business monitoring, meaning balances and payment patterns, now needs the same coverage. That requires either business staff on rotation or technology staff trained to recognize business conditions and authorized to escalate them.
People and machines
The commenters who asked for automated liquidity transfers were making a staffing choice. Every function that can run on rules overnight saves a person on the roster. But each automated function needs a person who can be reached when it fails or when a condition falls outside its rules.
A workable model, again good practice rather than a requirement, has three layers. Automated rules handle the routine, such as balance thresholds triggering transfers or alerts. An on-call rotation of trained staff responds to alerts within a set time. A smaller group of decision-makers can authorize unusual actions, such as a large funding move or a decision to stop sending payments. The design question is how many alerts reach the second layer and how often the third is woken.
The closing time
The notice flagged one moment for attention: the service’s daily closing time, when end-of-day balances are struck. A bank must avoid a negative balance at that moment. Payments arriving shortly before the close need either real-time monitoring or enough funding to absorb them.
For a bank that chooses funding over monitoring, the buffer rises. For one that chooses monitoring, someone must be watching at the close every night. Either way, the closing time becomes a daily operational event, seven days a week, with an owner and a procedure.
Alerts that mean something
Round-the-clock monitoring fails in two opposite ways. Too few alerts, and a draining balance goes unseen. Too many, and the on-call person learns to ignore them. Both failures are more likely at night, when a single person is responding.
Our suggestion is to set alert thresholds from the buffer policy, not from system defaults. One threshold prompts a check, a lower one prompts action, and a third wakes a decision-maker. Each alert should say what it means and what the first step is. The thresholds should be reviewed after every incident and every long weekend, as flows change.
No more maintenance windows
A service that never closes removes the off-peak hours that technology teams use for changes. Batch systems may rely on overnight or weekend windows for upgrades, patches and tests. A bank connected to a 24x7x365 settlement service has no such window for the systems in the payment path. Changes have to be made while payments flow.
That raises the bar for change management and resilience. It favors redundant components that can be updated one at a time, staged releases, and tested rollback. The 2019 notice argued that a second operator would promote resilience through redundancy, since a single service would be a single point of failure. The same logic applies inside a bank. A payment path with one server, one connection or one key person has no way to stay up through its own maintenance.
The Federal Reserve’s own later choice is telling. In 2025, when it decided to lengthen Fedwire hours, it decided not to open Fedwire on Saturdays. It said many participants and the Reserve Banks used Saturdays for maintenance, testing and upgrades, and that keeping that downtime could help participants reduce operational and cyber risk. FedNow, designed as a 24x7x365 service, was not built around such a weekly pause. Banks connected to it need to design for change without downtime.
Testing coverage
Coverage plans fail at the edges: holidays, staff illness, a system change that silences an alert. One approach, again our suggestion rather than a requirement, is to test the weekend model the way a bank would test disaster recovery. That means a scheduled exercise in which an alert fires at an inconvenient hour, a measure of time to response, and a review of whether the person who answered had the authority and information to act.
Fraud and exceptions in the small hours
Instant payments change the economics of fraud response. Once a payment settles, it cannot be pulled back through the payment system itself. The 2019 notice recognized this, describing the overall safety of faster payments as depending in part on how well fraud can be detected and prevented, given their irrevocable, real-time nature.
Why the clock matters more
With batch payments, there can be time between initiation and settlement in which a suspicious payment can be held. With instant payments, the decision to send is the decision to settle. Fraud controls must run before the payment leaves, in the flow, in seconds.
That shifts fraud work from review to rules. Rules decide automatically whether to send, hold or reject. People then deal with what the rules hold and with what customers report after the fact. If those people are off shift, held payments wait and customer reports sit unread. In a world of instant payments, a fraudster who knows the bank’s weekend roster has an advantage.
The exceptions queue
Not every exception is fraud. Customers send to the wrong account, enter the wrong amount or dispute a payment. The notice reported that more than 90 commenters addressed the importance of fraud prevention services. The 2020 service details, covered in hindsight, described tools for payment inquiries, reconcilements and certain exceptions in the first release.
Tools help, but a person has to use them. A customer who sends money to the wrong account on a Saturday expects the bank to try to recover it on Saturday. Whether a bank can do that depends on whether it has a person on duty with access to the tools and authority to contact the receiving bank.
Holding a payment, and telling the customer
Real-time fraud rules make real-time decisions about customers. A rule that holds a large first-time payment on a Saturday evening protects the customer from a scam, or blocks a legitimate purchase. Either way, the customer wants to know what is happening, and the bank needs someone who can review the hold and release or reject it.
A hold that waits until Monday turns an instant payment into a slow one and may cost the customer more than the risk it prevents. The bank therefore has to decide which holds it can review out of hours, set thresholds accordingly and tell customers in clear terms how holds work.
Customer service as a control
Weekend customer service becomes part of the fraud control. A customer who reports a scam within minutes gives the bank its best chance to stop related payments, freeze an account or warn the receiving institution. A customer who reaches a recorded message until Monday gives it none.
This connects the front line to the back office in a way batch payments did not require. Weekend service agents need a direct path to the fraud on-call. The fraud on-call needs authority to act on the account. Both need procedures written for instant payments, not adapted from card disputes or wire recalls.
Directories and misdirected payments
The 2019 notice reported that more than 100 commenters discussed a directory, and that many called it an important driver of adoption. A directory would let payments be routed using a receiver’s phone number or email address instead of bank routing and account details. The notice explained the appeal: account details can be hard for a sender to obtain.
Convenience changes the error and fraud profile. A payment addressed by phone number depends on the directory being right and on the sender typing the right number. It also opens room for impostors who persuade a customer to send to their number. Whatever routing method a bank adopts, its exception and fraud staff need to understand how a payment was addressed in order to trace it, and they need that understanding on weekends too.
Training the weekend roster
The people on a weekend roster may not be the people who designed the processes. An analyst covering fraud on a Sunday may usually work on card disputes. A reconciliation clerk may cover a function they see once a month. Instant payments need procedures that a capable person can follow without the designer in the room.
Our suggestion is short, scenario-based runbooks: a large unexpected outflow, a stalled payment queue, a customer reporting a scam, a reconciliation break above a threshold. Each runbook names the first actions, the authority limits and the escalation contact. Rotating staff through exercises built on these runbooks keeps the weekend model from depending on one or two experts.
Who carries the cost
The operational demands fall unevenly. They depend on the bank’s size, whether it already runs around the clock, and how it chooses to participate.
Large banks
A large bank with card, trading or international operations may already staff some functions around the clock. For it, FedNow extends existing coverage to new processes. The cost is real, but the model exists.
Its harder problem is scale. A large bank’s weekend instant payment flows may be large enough that buffer sizing becomes a balance sheet question. The 2025 Federal Reserve notice on expanded Fedwire hours, covered in hindsight, records that large banks largely supported longer hours, which fits an institution for which weekend operations are already normal.
Community banks and credit unions
For a community bank or credit union with weekday treasury and accounting staff, each demand in this article is new. The 2019 notice said thousands of banks might face significant uncertainty about the overall benefits of offering faster payment services and about whether RTGS-based settlement suited smaller institutions.
Such institutions have options that reduce the staffing load. The 2020 service details said a bank could use a service provider to send and receive payments on its behalf, and could settle in a correspondent bank’s master account. The first release would also allow banks to join as receive-only participants. Each option moves some operational work to a partner. None removes the need for someone at the institution to own the decisions about liquidity, exceptions and fraud.
What a partner can and cannot carry
A service provider or correspondent can run connections, monitor technical health and, by contract, apply some rules. It is less suited to deciding whether to hold a member’s payment, call a customer about suspected fraud or approve a large funding move. Contracts should say which decisions the partner makes, which it escalates, to whom and how fast, at every hour the service runs.
A receive-only start defers the outgoing-payment risks, including most fraud exposure and most of the liquidity drain. It does not defer the ledger. Incoming payments settle at all hours, so seven-day balances and reconciliation apply to a receive-only participant as well.
The workforce side of round-the-clock
Staffing a function around the clock starts as a headcount question but goes further. It changes job design, pay, on-call rules, fatigue and retention. Some people in treasury and accounting chose those roles for their weekday hours. Asking them to join a weekend rotation is a change in their terms, and the bank has to decide how to compensate it and how to keep it fair.
The alternatives each carry trade-offs. Hiring dedicated weekend staff is expensive for a small institution and leaves them isolated from the weekday team. Rotating existing staff spreads the load but needs cross-training. Outsourcing to a provider shifts cost into a contract but leaves decisions that cannot be delegated. A bank should choose deliberately, because the default is that the same few people answer every alert.
Measures worth watching
A bank that has built a weekend model needs evidence it is working. Useful measures include the share of the weekend buffer used at its lowest point, the number of after-hours alerts and the time to respond to them, and the number of reconciliation breaks older than one day. Another is the number of customer fraud reports received outside business hours and the time to first action on each.
These are our suggestions, not supervisory requirements. Their value is that they show trends. A buffer used more each month, or alerts arriving more often, is a signal that flows are outgrowing the model before a weekend proves it.
The case that the burden was overstated
The strongest counter-argument is that FedNow payments were expected to be small. The 2019 notice itself said transaction values were expected to be relatively small compared with other master account activity, such as Fedwire transfers. Small values mean small liquidity swings, a small buffer and limited weekend risk. On this view, a bank can automate the little that needs doing and keep its weekday staff.
The counter-argument holds for liquidity at low volumes. It is weaker for the other three functions. Seven-day accounting applies whatever the value. Fraud and error handling scale with the number of payments, not their size, and each one matters to the customer who made it. And volumes are a choice made by customers, not by the bank, once the service is offered. A bank that sizes its operations for today’s small flows needs a trigger for revisiting the model as flows grow.
Deciding to wait
Joining was voluntary. A bank that cannot yet staff or automate the functions described here can decide to wait, or to start as a receive-only participant, and say so. That is a legitimate risk decision, and it is better made openly than discovered after launch. A receive-only start still brings the seven-day ledger with it, so it narrows the staffing gap without closing it.
Waiting has its own cost, since customers may move to institutions that offer instant payments. The board should weigh that cost against the operational risk of launching without the weekend model in place. Both sides of that decision should be written down with the assumptions behind them.
Planning the roster before the product
The practical conclusion is about sequencing. It is natural to plan FedNow as a product launch: connection, core processor, customer interface, pricing. The 2019 notice suggests the roster should come first. Before a bank sends its first instant payment, it should know who watches the balance on a holiday weekend, who reconciles Sunday’s ledger and who answers a scam report on a Saturday night.
Those answers set the cost of participation more than the technology does. They also determine whether the bank can offer the service safely at all, which is a decision for the board as much as for the payments team.
With hindsight
The service arrived on the Federal Reserve’s published timetable. In August 2020, the Board announced service details (85 FR 48522). The FedNow Service would have a 24-hour business day every day of the week, including weekends and holidays, with end-of-day balances reported for each of those days and access to intraday credit. Because the service would run around the clock, it would incorporate a liquidity management tool as a core feature. That tool would let participants transfer funds between one another for FedNow-related liquidity needs, and support transfers to joint accounts used by private-sector instant payment services.
The 2020 notice listed optional features for the first release: fraud prevention tools, receive-only participation, request for payment, and tools for payment inquiries, reconcilements and certain exceptions. It confirmed that a bank could use a service provider or settle in a correspondent’s account. Commenters had asked the Board to raise the $25,000 transaction limit proposed in 2019, and the Board said the Reserve Banks would set a limit consistent with market practices before launch.
In December 2022, the Board adopted changes to its Policy on Payment System Risk to support the service (87 FR 75254). The notice said intraday credit would be available on a 24x7x365 basis to FedNow participants on the same terms as other Federal Reserve services. It said the FedNow close would align with the Fedwire close on all calendar days, and that an end-of-day balance would be calculated for each calendar day.
On July 20, 2023, the Federal Reserve announced that the FedNow Service was live. The release named 35 early-adopting banks and credit unions, plus the Treasury’s Bureau of the Fiscal Service, as ready to use it, along with 16 service providers. It described a commitment to work with the more than 9,000 banks and credit unions in the country to make the service widely available over time.
The value ceiling kept rising. In September 2025, Federal Reserve Financial Services announced that it would raise the FedNow network transaction limit from $1 million to $10 million, effective November 2025, and called it the second limit increase that year. It said financial institutions could still set lower limits based on their own risk parameters, and that the service then had more than 1,400 participating organizations. Larger permitted payments raise the stakes of the weekend liquidity and fraud questions this article describes.
The liquidity side of the 2019 plan has taken longer and, so far, come out smaller. In May 2024, the Board proposed extending Fedwire Funds and National Settlement Service hours to 22 hours a day, seven days a week, every day of the year (89 FR 39613). In November 2025, it announced a narrower decision (90 FR 51356). Fedwire Funds would move to 22 hours a day, six days a week, Sunday through Friday, including weekday holidays, with NSS following. The Board expected the Reserve Banks to implement this in 2028 or 2029.
The 2025 notice explains the retreat in terms this article would recognize. It says a large majority of small and midsize banks opposed the 2024 proposal, citing costs and challenges related to staffing, operational changes and competitive pressure from large banks. It does not extend Fedwire to Saturdays, because many participants and the Reserve Banks use them for maintenance, testing and upgrades. It says participation in expanded hours is optional. It says the Board would consider going further, up to 22x7x365, no sooner than two years after 22x6 begins. It also says the Federal Reserve is exploring whether to expand discount window operating days.
What held up is the operational reading. The Federal Reserve’s own documents show the liquidity gap the 2019 notice identified addressed inside FedNow, through its liquidity tool and round-the-clock intraday credit. The Fedwire and NSS expansion announced in November 2025 is set for 2028 or 2029 and stops short of the 24x7x365 the 2019 notice raised. Those documents also record staffing among the objections smaller banks raised. FedNow did not ask banks to become technology companies. It asked them to keep the parts of the bank that manage money awake.
Spotted an error? Tell us and we will check it and correct the article where needed.