Seiche’s board reads 46 out of 100, STRAIN, led by a dated reserve path into Treasury settlement. But SOFR remains below IORB, the Standing Repo Facility is unused, and market stress is low: this is a capacity warning, not a confirmed funding squeeze.
The material change on Seiche’s dollar-funding board is in the calendar, not the observed price of overnight cash. The dated reserve-path component is saturated, contributing 11.0 points to a composite reading of 45.6, classified as STRAIN. Its immediate forcing event is a $282B Treasury auction settlement on 2026-08-25.
That warning has not yet been confirmed by the market plumbing it is meant to anticipate. SOFR was 3.66% as of 2026-08-17, 1.7bp below IORB at 3.65% as of 2026-08-19. The question is therefore narrower than whether funding is already tight: can dated drains make dollar-funding capacity less elastic before overnight rates show broad scarcity?
The chain begins with reserves and dated balance-sheet demands. Reserves fell $49.3B in the week to $2,944B as of 2026-08-12. The Treasury General Account rebuilt $56.6B and foreign reverse-repo balances absorbed $39.7B, while a residual release of $36.5B offset part of those pressures. Six ledger legs sum to the reserve change to the dollar; $18.9B, or 0.28% of assets, remains in the unnamed residual.
Seiche’s mechanism is that an auction settlement can test the distribution and intermediation of reserves, rather than merely their aggregate level. The reserve-demand model places current reserves $759B below a fitted kink of $3,703.1B, with an in-sample fit of 0.62. That kink is a modelled structural marker, not an observed market boundary. The negative SOFR–IORB spread is direct evidence that the marker has not yet translated into broad overnight scarcity.
The relevant historical desk guide is the December 2025 year-end squeeze. Reporting incentives reduced willingness to intermediate while balance-sheet demand was seasonally high, and Standing Repo Facility use made the constraint observable. This is background, not a forecast: the comparison is construction-point-in-time, not a publication-vintage backtest.
The near-term test is the $282B settlement on 2026-08-25. Seiche’s reserve-path scenario puts reserves at $2,958.8B on that date, with a worst-case path of $2,897.6B. The scenario is a Seiche derivation rather than a forecast of a market price. Its significance is timing: the settlement arrives while the reserve estimate is below the fitted kink.
The calendar extends the test. On 2026-08-31, a $24.0B settlement coincides with a worst-case reserve path of $2,877.3B. Further worst-case paths are $2,848.3B on 2026-09-15, $2,820.9B on 2026-09-16, and $2,784.0B on 2026-09-30. Seiche’s published month-end forecast is 2.2bp, within a band from -1.8bp to 6.8bp, with severity at 1 on its calm-to-extreme scale.
Official-sector flows add context but do not independently establish a squeeze. Over 13 weeks through 2026-08-12, foreign officials’ Fed custody Treasury holdings fell $90.2B while foreign reverse-repo parking rose $34.8B. The combined official footprint still declined $55.5B to $2,954B. In Seiche’s accounting, 62% of the custody decline left the official sector rather than rotating within it.
The board’s main signal is divergence. Plumbing stands at the 65th percentile and market stress at the 29th percentile: plumbing leads price by +35 percentile points. The pooled five-business-day event read is 6.3%, while the published book remains neutral.
The counter-case can defeat the thesis because it is already visible in current prices and facility use. SOFR is 1.7bp below IORB, an abundance signal rather than a scarcity print. The effective federal funds rate was 3.63% as of 2026-08-17, also below IORB. These are not signs of an active scramble for marginal overnight dollars.
The backstop is also untested. Standing Repo Facility accepted take-up was $0.00B as of 2026-08-18. Market stress was at the 29th percentile of its own history, with VIX at 15.19 and high-yield option-adjusted spreads at 2.7% as of 2026-08-17.
The strongest alternative explanation is not simply that a backstop would contain stress. It is that no material strain is developing: settlements may clear cleanly, reserve distribution may remain functional, and the negative SOFR–IORB spread may be more informative than the modelled reserve path.
The observable test is the period around the 2026-08-25 settlement. The capacity thesis gains support only if the dated drain is followed by deterioration in SOFR relative to IORB, Standing Repo Facility use, or broader market-stress confirmation alongside the plumbing warning.
It is weakened if the settlement passes with SOFR still below IORB, zero or immaterial Standing Repo Facility take-up, and no confirmation from broader market pricing. The 2026-08-31 month-end window provides a second test. Calendar pressure alone does not confirm a funding event; the proposed mechanism must become observable in pricing or backstop use.
Follow the sequence rather than a single level: Treasury settlement, reserve redistribution, willingness to intermediate, secured overnight rates relative to IORB, and backstop use. A break at one link need not propagate. Confirmation requires movement across the chain.
For system dollar-funding capacity, follow Seiche’s daily dispatches, the current dispatch, and its data overview. For institution and lender balance-sheet risk, see LiquiLens. For market liquidity, crowding, and exit capacity, see Undertow and its exit material.
This analysis uses Seiche’s methodology, series index, data overview, and current dispatch. The observed balance-sheet, custody, foreign reverse-repo, rate, facility, and Treasury-calendar figures are represented in Seiche’s point-in-time synthesis.
The composite, dated reserve paths, fitted kink, percentiles, and event read are Seiche derivations, not observed market prices. Confidence measures agreement across independent signals, not the probability of a market outcome. Source publication time is not yet collected uniformly, so event, desk-knowledge, and publication clocks remain separate.
Research boundary: Seiche covers system dollar-funding capacity. LiquiLens covers institution and lender balance-sheet risk. Undertow covers market liquidity, crowding, and exit capacity. Related Palimpsest readings are context-only and never enter the composite or a model feature.
Research and market data, not investment advice.