Liquidity in a 24/7 World: Managing Settlement in Real-Time

Executive Summary

The transition to 24/7/365 real-time payments (RTP) has created a profound, behind-the-scenes challenge for banks: liquidity management. In the old “next-day” payment world, banks only needed to ensure they were funded at the end of the business day. But in a real-time world, “end-of-day” is meaningless. Banks must now be able to settle large, unpredictable payments instantly, 24/7, including weekends and holidays. This shift from “net end-of-day” to “real-time gross” settlement has created a massive liquidity puzzle. Banks now risk “trapping” huge, inefficient pools of cash in settlement accounts around the clock, or worse, failing a payment at 3:00 AM on a Sunday due to a temporary shortfall. The race is on to develop new, AI-driven treasury tools to predict payment flows and manage liquidity “just-in-time.”

1. The Old World: The Comfort of “Net Settlement”

For decades, international and domestic payments (like ACH and wires) operated on a “Net Settlement” (NS) or “Designated-Time” basis.

• How it worked: Throughout the day, banks would send payment messages back and forth. The central bank or clearing house would keep a running “tab” (a net position).
• The 5:00 PM “Settle-Up”: At the end of the business day (e.g., 5:00 PM), the system would “close.” The clearing house would tell Bank A, “You sent $1.1B and received $1.0B, so you owe $100M to the settlement account.”
• The Liquidity Advantage: This was incredibly efficient. A bank only needed to manage its liquidity for this one settlement window. Its treasury team had all day to predict this final net position and borrow or lend funds as needed. Liquidity was a predictable, 9-to-5 job.

2. The New World: The Tyranny of “Real-Time Gross Settlement”

Real-Time Payment (RTP) networks (like FedNow, SEPA Instant) and modern high-value payment systems operate on a “Real-Time Gross Settlement” (RTGS) basis.

• How it works: Every single transaction is settled individually and instantly (“gross”). There is no “tab.”
• The Pre-Funding Requirement: For this to work, a bank must have 100% of the funds for every payment in its settlement account before the payment is sent. If a corporate client wants to send a $5M payment at 3:00 AM on a Sunday, the bank must have $5M in its central bank account at that exact moment.
• The 24/7 Problem: This requirement is 24/7/365. The bank’s 9-to-5 treasury team is now offline, but its liquidity obligation is not.

3. The “Trapped Liquidity” & “Fragmentation” Crises

This new model creates two massive inefficiencies for a bank’s treasury.

1. “Trapped Liquidity”: To be safe, a bank might have to park $500M in its central bank settlement account over the entire weekend, just in case its clients initiate large payments. This is a huge capital inefficiency. That $500M is “trapped”—it cannot be invested, loaned out, or earn interest. It is operationally “dead” money, held purely as a buffer.
2. “Fragmented Liquidity”: The problem is multiplied in international banking. A global bank may have 20 different settlement accounts for 20 different currencies and RTP systems. It might be “cash rich” with a €2B surplus in its European (TARGET2) account but dangerously low on funds in its US (Fed) account at 2:00 AM. Its total liquidity is fine, but it’s in the wrong pot. This “fragmentation” is the core challenge of 24/7 global treasury.

4. The Solution: Predictive, AI-Driven Liquidity Management

If the problem is 24/7, the solution must also be 24/7 and automated. Human treasury teams cannot manage this in real-time. This is where AI-driven “RegTech” and “TreasuryTech” are becoming critical.

• AI-Powered Forecasting: New platforms use machine learning to analyze a bank’s historical payment flows. They can predict, with high accuracy, the likely payment volumes for a specific client on a Sunday night versus a Tuesday morning, allowing the bank to forecast its liquidity needs hour by hour, not day by day.
• Automated “Sweeping”: These AI tools are connected via API to the bank’s various “pots” of money. When the system predicts a high-volume period for USD payments, it can automatically “sweep” or “pull” liquidity from a short-term investment account into the Fed settlement account just-in-time. When the peak passes, it “sweeps” the excess cash back out to be invested.
• Payment Orchestration: This is the “smart routing” of payments. If a non-urgent, low-value payment arrives and the system sees liquidity is tight, it can automatically route that payment via a slower, “next-day” rail, saving the precious real-time liquidity for high-value, time-sensitive transactions.

Conclusion

The move to real-time payments is far more than a “front-end” customer upgrade; it is a “back-end” revolution in how banks manage their own money. The days of end-of-day settlement are over. The new battleground is capital efficiency, and the winners will be the institutions that replace their 9-to-5 human treasury models with 24/7, AI-driven, predictive liquidity platforms.