Choosing between Fedwire, FedNow, ACH and CHIPS

The choice between the four systems comes down to four questions that an analyst should ask for every payment. First, how large is the payment. Very large wholesale payments belong on Fedwire or CHIPS, routine retail and commercial payments belong on ACH or FedNow. Second, how urgent is the payment. If it must settle in seconds at any hour, FedNow is the answer; if it must settle today during business hours with finality, Fedwire is the answer; if it can wait a day or two, ACH is the cheapest answer. Third, what is the cost tolerance. ACH is the cheapest by a wide margin, FedNow is low cost, Fedwire and CHIPS are more expensive because of the liquidity and operational demands. Fourth, what is the volume. A thousand payments belong in a file, not in a thousand individual instructions. Value bands make the choice concrete. Below 1,000 dollars, ACH or FedNow, because the Fedwire fee alone would be a large fraction of the payment. Between 1,000 and 10 million dollars, FedNow if it must arrive in seconds, Same Day ACH if it can wait until the end of the day, and Fedwire if it must be final in central bank money today. Above 10 million dollars, Fedwire or CHIPS, because FedNow and the RTP network both cap at 10 million dollars per transaction (FedNow from November 2025, RTP from 9 February 2025), and ACH is not built for single payments of that size. These bands are rules of thumb, not hard limits, and the analyst should confirm the actual limits with each bank, because a bank may set a lower FedNow limit for its own customers and a higher Same Day ACH limit for a corporate client. Urgency is about the settlement, not the instruction. A customer who instructs a payment at 10am and is told it will arrive today may mean three different things: credited within seconds, settled in central bank money today, or credited by the end of the day. FedNow gives the first, Fedwire gives the second, Same Day ACH gives the third. The analyst must translate the customer's words into the settlement promise, and write that promise down, because a customer who expected seconds and got end of day will complain, and a customer who expected finality and got a batch credit that can be returned will complain louder. The screen should state when the money will arrive and when it is final, in plain words, before the customer confirms. Cost is the sum of three things, and analysts often count only one. The first is the per payment fee the bank charges, which is highest for Fedwire and CHIPS, low for FedNow, and lowest for ACH. The second is the liquidity cost, which is the balance the bank must hold to fund the payment: the full gross amount for Fedwire, the net amount for CHIPS, a share of the master account balance for FedNow, and nothing intraday for ACH because it settles in batches. The third is the operational cost, which is the work of handling exceptions: ACH returns are routine and cheap to process in bulk, while a Fedwire exception is rare and expensive because the values are large and the time pressure is high. A payment that looks cheap on fees can be expensive on liquidity and exceptions, and the analyst should present all three costs to the business. Operating hours close the decision. FedNow runs 24 hours a day, 7 days a week, including weekends and holidays, so it is the only choice for a payment that must move at midnight on a Sunday. Fedwire and CHIPS run on business days during extended hours, so a payment instructed on a Saturday waits until Monday. ACH runs on business days with file cutoffs, so a payroll file submitted on Friday settles on Tuesday if Monday is a holiday. The analyst must build the calendar into the routing logic: a payment instructed outside business hours should be offered FedNow if it fits, or queued for the next business day with the settlement date shown, never silently delayed. The landscape is changing, and a good analyst keeps up with that change rather than treating the four systems as fixed. FedNow adoption is growing as more banks join, and the RTP network operated by The Clearing House offers a parallel instant option that many large banks already use. Same Day ACH windows have expanded and the per-transaction limit rose to 1 million dollars in March 2022 (Nacha), which makes ACH faster than it used to be. Fedwire has moved to ISO 20022, which makes its messages richer and easier to reconcile with international payments. The direction of travel is clear. Instant is becoming the default expectation for more use cases, and batch is becoming the choice for high volume, low urgency, low cost payments. The analyst who understands both the current state and the direction of travel can design a payment flow that works today and can evolve tomorrow without a rebuild. A final practical point is that most US banks offer all four systems, and the customer often does not know or care which one is used. The customer cares about when the money arrives, what it costs, and whether it is certain. The analyst's job is to translate the customer's need into the right system, to make the choice visible where it matters, and to hide the complexity where it does not. A payment screen that asks the customer to choose between Fedwire and CHIPS is a design failure; a payment screen that asks when the money must arrive and how much it is, and then picks the right rail, is a design success. That translation from business need to system choice is the core skill this chapter teaches. The failure mode of a wrong choice is worth spelling out, because it is how analysts learn. A payroll sent through Fedwire costs thousands of dollars in fees and needs millions in intraday liquidity for no benefit. A 2 million dollar trade settlement sent through ACH will not settle today, will not be final, and can be returned, which leaves the trade exposed. A 50,000 dollar supplier payment sent through CHIPS may sit in the queue for hours waiting for an offset, and the supplier will call to ask where the money is. Each of these failures comes from choosing the rail on habit rather than on the four questions. Writing the four questions into the requirements document, with the answer for each payment type, is what stops the next failure.