Peak net debit
A bank's largest net debit at the end of any minute in the day, counting only payments in Sampo, and zero if the bank never goes into net debit. A payment that is never settled still counts, at its own time.
Evidence
The evidence for each claim
The home page rounds Sampo's results. This page gives each figure in full, how it is measured, and what the model covers.
Each claim and its source
| Claim on the site | Figure in full | What it measures | Source |
|---|---|---|---|
| 49% less liquidity | 49.3% lower modeled in-scope peak net debit | Cohort monthly peak net debit (the sum of each bank's monthly peak), with and without Sampo, averaged over the complete months February and March of the synthetic history. On an average-day basis the reduction is 59.0%. | FNA five-bank synthetic model referenced to Fedwire |
| $55bn across five large banks, simulated | $55.42bn | The same cohort monthly peak net debit, in dollars, summed across the five banks. | FNA five-bank synthetic model |
| 2 hours earlier on average | 2.1 hours on average, when banks send most of their payment intents before the opening | Average advance of each payment's settlement against settlement without Sampo, across the cohort, in the scenario where banks know 70% of payment value at the opening. | FNA five-bank synthetic model; Simulation and Bank views |
| $554m a year | $554.2m | Annual value: for each bank, its saving × 100% recognition × a cost of liquidity of 1% a year, summed over the five banks. A sensitivity on bank-supplied inputs, not booked profit. | FNA five-bank synthetic model; calculator |
| 63% less value left to the last hour of the day | 43.5% of value without Sampo; 16.0% with Sampo | Share of the day's payment value settled in the last hour of the day, with 30% of payment intents known at the opening. | FNA five-bank synthetic model |
| Submitted early | 90% of large-value payments (90.4%; 87.7% of value) | Share of payments submitted before their own time. | FNA five-bank synthetic model |
| 22% less liquidity on CHIPS | 22%, on about $1.7 trillion of daily payments | Liquidity needed with reordering payments alone, against their recorded order | FNA analysis of CHIPS data for The Clearing House |
| 20% on CHAPS | About 20%, about £4bn a day, with no loss of throughput | Reduction in CHAPS settlement banks' intraday liquidity needs after the liquidity-saving mechanism | Bank of England; Davey and Gray (2014) |
| $4.6 trillion a day on Fedwire | About $4.6 trillion; more than half settles in the final third of the day | Daily Fedwire Funds value and its intraday distribution | Federal Reserve |
The model
The Sampo results come from a synthetic simulation. It contains no production bank data and no Fedwire transaction data. The 125 days are synthetic weekday labels, not an observation period.
The synthetic book is referenced to published Fedwire structure: a core of five banks with a fixed network of payment relationships, payment sizes fitted to published shares of payments above $100m, $1bn and $2bn, and an intraday profile with one peak at the opening and another in the late afternoon.
For each day, the generator draws new amounts, the times the banks send the payment intents, and each payment's own time, from a documented seed sequence.
Five further synthetic days from the same generator, with payment intents only and no Sampo results, are on the AI Sandbox page: synthetic sample days.
| Banks | 5 |
|---|---|
| History | 125 synthetic days |
| Payments | 62,500; 500 each day |
| Payment values | About $500m to $5.1bn; median about $972m, mean about $1.08bn |
| Value per day | About $539bn; 11.74% of 2025 Fedwire daily value |
| Payment intents sent at the opening | 29.85% of payments |
| Balance sampling | Fedwire convention: net all settlements in a minute, sample at minute end |
How the saving is measured
Without Sampo
Each bank submits every payment at its own time. FNA measures each bank's peak net debit.
With Sampo
Sampo advises banks to submit some payments early, never later than their own time. Its no-worse-off guard aims to reduce the risk that a bank is worse off.
The comparison
The saving is the peak without Sampo less the peak with Sampo. A bank is worse off on a day when its realised peak with Sampo is above its peak without Sampo. The guard is not a guarantee.
The method
Parts of Sampo's rule are patent pending. This page describes the data and the measurement, not the rule.
What the model leaves out
The metric is modeled in-scope peak net debit, not a bank's total intraday liquidity requirement.
The model omits opening account balances, payments below $500m and other out-of-scope flows, other rails, collateral and credit facilities, real RTGS queue, rejection and finality behaviour, cancellations, and the time between submission and settlement.
It demonstrates the mechanism and sizes a pilot. It does not estimate a Fedwire tail, actual daylight credit, bank profit, regulatory capital or a required liquidity buffer.
What a lower peak is worth
The value of a lower peak depends on two numbers that each bank sets: the share of the reduction its approved methodology recognises, and its cost of liquidity, an annual rate.
Annual value = reduction in peak × share recognised × cost of liquidity
FNA computes it for each bank, at that bank's own share and rate, and sums it over the banks.
A lower modeled peak does not by itself release a regulatory buffer, HQLA or capital. Try it with your numbers.
How a bank validates this
An evidence pilot measures the opportunity on the bank's own payments. The result is that bank's figure, not a network average.
Replay
FNA replays the agreed in-scope historical or shadow flows.
Compare
FNA compares the bank's peak without Sampo with its peak with Sampo, under an agreed measurement convention.
Test
FNA and the bank test adherence and exception scenarios, including the days coordination helps least.
Translate
The bank applies its own liquidity, stress testing, funds transfer pricing and regulatory methodology to any reduction.
Decide
The bank decides whether the measured opportunity justifies advisory production.
Join the pilot
FNA is selecting design-partner banks for a USD shadow-data pilot. Become a design partner