# Payout cash buffers: RTP vs ACH — when to cut the legacy 2-day hold

Daniela Ruiz · November 2, 2026

> RTP speed alone won't justify cutting the 2-day ACH hold. A formula based on settlement lag and exception rates shows when payout buffers can be safely reduced.

| Takeaway | Detail |
| --- | --- |
| Replace 48 hours with a formula. | Set the payout cash buffer per rail per beneficiary cohort as (settlement lag in hours ÷ 24) + (measured exception rate × average exception resolution days). |
| Cut the ACH-era 2-day buffer only after the exception rate. | RTP going live for a beneficiary cohort is not enough—reduction must follow measured exception rate data, not speed alone. |
| The push for real-time liquidity traces to March 2020. | Bottomline states that with the emergence of the pandemic in March 2020, employee well-being became the first concern for leaders, including CFOs. |
| A lesser buffer can lower cost of capital. | Management Study Guide states that requiring a lesser buffer and transferring funds between accounts can lower the overall cost of capital. |

This guide breaks down when to cut the legacy 2-day payout buffer by rail and beneficiary cohort.

It applies a measured exception-based rule to RTP and ACH without relying on a flat 48-hour hold.

![Payout cash buffers](https://static.mm-ais.com/article-images-ai/payout-cash-buffers-rtp-vs-ach-when-to-c-ai-d08ba5c6.jpg)

## How the buffer actually gets consumed

The payout cash buffer is consumed by two independent clocks, not by one flat 48-hour allowance. The first is the rail’s settlement lag: RTP settles continuously, so its lag is an intraday measure that can be expressed in hours, while ACH settles in batch windows, so its lag can stretch across cutoffs and into the next business day. The second is the exception clock—returns, name mismatches, and closed accounts—which runs in calendar days, not hours. Because an exception can reopen after settlement, the buffer must cover both the time funds are in transit and the time an exception takes to resolve.

Expressed as a working rule, the buffer is the sum of those two clocks: settlement lag in hours divided by 24, plus the measured exception rate multiplied by the average exception resolution days. The first term captures intraday liquidity mechanics such as prefunding positions, intraday credit, cutoff times, batch settlement windows, and continuous settlement. The second term captures the reserve needed for exceptions that do not care how fast the rail is.

That distinction matters because a faster rail changes only the settlement term. If visibility and reporting do not keep pace, treasury still cannot see, explain, or act on the exception term in real time. Bottomline’s Kevin Grant frames the treasury problem as three separate capabilities—real-time liquidity, real-time visibility, and real-time reporting—so a faster rail without faster visibility does not shrink the buffer; it only shrinks one of those three capabilities.

The exception clock is the part most likely to be underestimated. Returns, name mismatches, and closed accounts are resolved on calendar days, and their duration is driven by beneficiary data quality, exception handling, and bank response times rather than by settlement speed. A cohort with a low measured exception rate and short average resolution time needs a smaller reserve; a cohort with noisy beneficiary data does not get relief from RTP alone.

The practical check is to compare the two-clock total with the legacy 48-hour allowance for that rail and beneficiary cohort, then apply the 0.5-day gating test defined in the decision rules below. Do not cut the ACH-era buffer on settlement speed alone—cut it only when the computed buffer is below 0.5 days and same-day exception reporting is in place for that cohort; otherwise the flat buffer remains the safer floor.

## What the evidence actually shows

For payout teams weighing any change to the legacy 2-day buffer, the market evidence points less to rail speed and more to the ability to see cash in time. Bottomline's *It's prime time for real-time liquidity management* frames the treasury problem as three capabilities—real-time liquidity, real-time visibility, and real-time reporting—and attributes the post-pandemic push for real-time liquidity to CFOs needing to know their cash position under disruption. Check the article's own survey figures for any comfort-with-forecasting percentages before citing them; the point that holds from the source is that visibility, not rail speed, is the binding constraint.

The same Bottomline study attributes the post-March-2020 push toward real-time liquidity to CFO-level cash position questions in a remote operating model. With the first priority being employee well-being at the start of that disruption, the focus quickly shifted to knowing the current cash position. In effect, the demand driver was visibility under disruption, not a general race for speed. This framing is consistent with Citi's work on real-time liquidity, which treats keeping pace with a real-time world as the task for treasury.

Taken together, the survey and framing figures from Bottomline and Citi show a convergence on visibility as the binding constraint. For most of the firms surveyed, the hurdle is not "can this rail settle?" but "can we forecast our position well enough to act on what is about to settle?" That distinction is critical, because a flat 48-hour allowance can mask a visibility problem instead of solving it.

For payout operations, this means the justification for holding the ACH-era buffer cannot rest on the existence of faster rails alone. If the majority do not claim comfort with liquidity planning, then the ability to shorten any buffer is tied directly to gaining that live view, not to the mere fact that RTP is available for a cohort. Without enough visibility to track exceptions as they arise, cutting time out of the buffer would be a guess rather than a measured decision.

In short, the evidence from these sources only establishes that visibility is the binding constraint. It explains why the conversation has moved to real-time liquidity in the first place, but it does not by itself dictate a universal cut. What it does dictate is the order of work: fix the view before you fix the hours.

## RTP buffer vs ACH buffer, compared

For this comparison alone, the side-by-side view below is the point of focus. The winner is the RTP cohort buffer on size, but only when paired with same-day exception reporting; the ACH 2-day buffer wins on predictability because batch windows make the settlement clock deterministic. That trade-off is the reason a cut cannot be justified by rail speed alone, and it is also why you must not compare a round-trip ACH cycle against a one-leg RTP payout when judging the buffer.

Worked example: for the RTP cohort, 0.1 + (0.015 × 3) = 0.145 days of payout volume; for the ACH cohort, 2 + (0.015 × 3) = 2.045 days. Express the result in days of payout volume so it can be compared directly with the legacy 2-day allowance. The 0.045-day exception reserve is identical in both cases, so the 1.9-day difference is settlement lag. That is why ACH is expensive to fund on the settlement term, not the exception term.

Cost of being wrong runs both ways. Over-buffering leaves cash idle against a payout book that could be deployed elsewhere; under-buffering risks failed payouts, overdraft use, or manual funding intervention and can force same-day funding calls. If the exception rate rises, the reserve rises with it; if settlement lag is understated, the buffer is too thin even when the exception rate looks stable.

Cut the legacy 2-day ACH buffer only after the cohort’s exception rate has been measured and held through at least one full exception-resolution cycle. A short RTP settlement lag by itself is not a sufficient trigger. Keep separate worksheets per cohort and do not blend RTP and ACH populations, because their settlement terms differ.

## Decision rules for cutting the buffer

Apply the cut decision cohort by cohort, never to the whole payout program at once. A cohort qualifies only when its computed buffer is below 0.5 days and its exception reporting is same-day; then set that cohort's buffer to the computed figure and retire the 2-day assumption for that cohort alone. The 0.5-day test keeps the legacy allowance in place wherever the new math still leaves a material intraday cushion.

| Condition | Action |
| --- | --- |
| Computed buffer under 0.5 days and same-day exception reporting | Cut that cohort's buffer to the computed figure and retire the 2-day assumption for it |
| Exception reporting lag over one business day | Do not cut; the buffer is protecting against visibility latency, not settlement lag |
| Cohort includes first-time beneficiaries or payroll | Hold the ACH-era buffer until at least one full pay cycle of exception data exists for that cohort |

If exception reporting lags by more than one business day, do not cut, even when the computed buffer is small. In that case the buffer is protecting against visibility latency rather than settlement lag, and cutting it would convert a reporting delay into a funding shortfall. Restore the cut review only after reporting returns to same-day.

If the cohort includes first-time beneficiaries or payroll, hold the ACH-era buffer until at least one full pay cycle of exception data exists for that cohort. New beneficiary populations and payroll runs have not yet produced a measured exception pattern, so the computed buffer would rest on an untested rate. Treat the first cycle as the measurement period, then re-run the cut test.

When a cohort passes the cut, document the cohort, rail, effective date, and exception rate used, and leave all other cohorts on their existing buffers. Re-test any cut cohort after a material change in beneficiary mix, payment volume, or exception reporting timeliness.

## Frequently Asked Questions

**What event triggered the push for real-time liquidity?**

The push for real-time liquidity traces to March 2020.

**Is launching RTP for a beneficiary cohort enough to cut the legacy 2-day buffer?**

RTP going live for a beneficiary cohort is not enough—reduction must follow measured exception rate data, not speed alone.

**What two independent clocks consume the payout cash buffer?**

The payout cash buffer is consumed by the rail’s settlement lag and the exception clock (returns, name mismatch).

**How does settlement lag differ between RTP and ACH?**

RTP settles continuously, so its lag is an intraday measure that can be expressed in hours, while ACH settles in batch windows, so its lag can stretch across cutoffs and into the next business day.

**What formula should replace the flat 48-hour hold?**

The payout cash buffer per rail per beneficiary cohort should be set as (settlement lag in hours ÷ 24) + (measured exception rate × average exception resolution days).

**What broader concern drove this liquidity focus for CFOs at the time?**

With the emergence of the pandemic in March 2020, employee well-being became the first concern for leaders, including CFOs.

## Quick answers

| What formula should replace the flat 48-hour hold for the payout cash buffer? | Set the payout cash buffer per rail per beneficiary cohort as (settlement lag in hours ÷ 24) + (measured exception rate × average exception resolution days). |
| --- | --- |
| When should the ACH-era 2-day buffer be cut? | Cut the ACH-era 2-day buffer only after the exception rate. |
| Is RTP going live for a beneficiary cohort sufficient to reduce the buffer? | RTP going live for a beneficiary cohort is not enough—reduction must follow measured exception rate data, not speed alone. |
| How is the payout cash buffer consumed? | The payout cash buffer is consumed by two independent clocks, not by one flat 48-hour allowance. |
| What can a lesser buffer do? | A lesser buffer can lower cost of capital. |

Also worth reading: **RTP vs Same-Day ACH: Fees, Routing, and the 24/7 Clock**: [RTP vs Same-Day ACH: Fees,](https://mosa.money/blog/rtp-vs-same-day-ach-fees-routing-and-the-247-clock.php) · **RTP vs ACH 2026 Payouts: Early-Pay Discount Math**: [RTP vs ACH 2026 Payouts:](https://mosa.money/blog/rtp-vs-ach-2026-payouts-early-pay-discount-math.php) · **2026 Payout Costs: Bank Transfer (ACH) vs Real-Time (RTP) vs Stablecoin — Switch >$10k**: [2026 Payout Costs: Bank Transfer](https://mosa.money/blog/2026-payout-costs-bank-transfer-ach-vs-real-time-rtp-vs-stablecoin-switch-10k.php)

### Related reading

- [Run Payout Reconciliation Locally: Mistral's Le Chonk 2026 - 0 Cash Data to Vendors](https://mosa.money/blog/run-payout-reconciliation-locally-mistrals-le-chonk-2026-0-cash-data-to-vendors.php)
- [Daily cash position: Automated Clearing House, Real-Time Payments, card by 4 p.m. or hold](https://mosa.money/blog/daily-cash-position-automated-clearing-house-real-time-payments-card-by-4-pm-or-hold.php)
- [Instant Payouts vs Standard ACH: When the Fee Justifies Speed — 2026 Payout Rules](https://mosa.money/blog/instant-payouts-vs-standard-ach-when-the-fee-justifies-speed-2026-payout-rules.php)
- [Tolan's $30M Raise: A Verify-Before-You-Commit Diligence Guide](https://mosa.money/blog/tolans-30m-raise-a-verify-before-you-commit-diligence-guide.php)
- [2026 Payout Costs: Bank Transfer (ACH) vs Real-Time (RTP) vs Stablecoin — Switch >$10k](https://mosa.money/blog/2026-payout-costs-bank-transfer-ach-vs-real-time-rtp-vs-stablecoin-switch-10k.php)
- [How 2026 campaign committees reconcile card, ACH, and crypto donations to FEC filings](https://mosa.money/blog/how-2026-campaign-committees-reconcile-card-ach-and-crypto-donations-to-fec-filings.php)

### Latest

- [How much does agent ID delay cost: 2026 step-up checks vs $80,000 fraud on...](https://mosa.money/blog/how-much-does-agent-id-delay-cost-2026-step-up-checks-vs-80000-fraud-on-automated-clearing-house-ach-real-time-payments-rtp-and-wire-rails.php)
- [Daily cash position: Automated Clearing House, Real-Time Payments, card by 4...](https://mosa.money/blog/daily-cash-position-automated-clearing-house-real-time-payments-card-by-4-pm-or-hold.php)
- [Instant Payouts vs Standard ACH: When the Fee Justifies Speed — 2026 Payout...](https://mosa.money/blog/instant-payouts-vs-standard-ach-when-the-fee-justifies-speed-2026-payout-rules.php)

Canonical: https://mosa.money/blog/payout-cash-buffers-rtp-vs-ach-when-to-cut-the-legacy-2-day-hold.php
Markdown: https://mosa.money/blog/payout-cash-buffers-rtp-vs-ach-when-to-cut-the-legacy-2-day-hold.php/index.md