Why are we still telling people to keep “3-6 months in a HYSA” as if operational risk doesn’t exist? In every real-world stress I’ve seen, the sequence of events matters more than the average yield on cash. Job loss, expense spike, plus a random bank fraud hold or brokerage outage is what breaks households. The old advice ignores payment-rail fragility, redemption gates, collateral calls, and basic settlement mechanics.
I’m proposing we shift from “emergency fund” to “contingent liquidity stack” engineered for time-to-cash and correlated stress. Rough sketch:
- 0-24 hours: small balances spread across two unrelated banks plus one credit union; one government MMF with same-day liquidity at a brokerage that supports instant debit card/ACH; high-limit credit card float; a minimal physical cash buffer. Goal: pay rent, food, medical without waiting on ACH holds.
- 24-72 hours: broker-held T‑bills or a gov MMF that can be used as collateral for a pledged asset line or margin debit if needed; pre-established, unused personal LOC (not HELOC) from a different institution than primary bank; ability to push funds via RTP/FedNow from at least one account.
- 3-10 days: weekly T‑bill ladder maturities; backup account at a too-big-to-fail bank if you normally bank with a fintech or regional; documented plan for COBRA or marketplace premiums and deductible shock.
- Intentionally excluded or deemphasized: TreasuryDirect I‑Bonds (illiquid timing), employer 401k loans/hardship withdrawals (slow), HELOCs (freeze risk in downturns), single-institution HYSAs with long Reg CC holds on large inbound transfers, prime MMFs with potential gates.
Questions for the hive mind:
1) Has anyone compiled actual hold-time data by bank for large ACH pulls during account openings or “suspicious activity” reviews? I’ve seen 5-9 business days on brand-name HYSAs when it matters most.
2) Which brokerages reliably allow immediate liquidity against T‑bills/gov MMFs without selling them (i.e., margin or pledged asset lines) and what are the haircuts under stress? Any datapoints from March 2020 or March 2023?
3) Government MMFs didn’t gate in 2020, but prime funds did. Is there any scenario where a 7‑day gate could hit a government MMF? If not, why are we still steering people to bank HYSAs with opaque hold policies instead of brokered gov MMFs plus a debit/ACH rail?
4) Credit availability correlation: during March 2020 and the regional bank mini-crisis, did anyone have lines cut or HELOCs frozen while their brokerage margin remained available, or vice versa? What actually held up?
5) Payment rails: who has successfully used RTP or FedNow in an emergency? Which banks actually let consumers push meaningful amounts instantly, and are there daily caps that make the feature marketing fluff?
6) Diversification of “cores”: does splitting cash across institutions that run on different banking cores/processors (e.g., FIS vs Fiserv vs Jack Henry vs in-house) reduce correlated outage risk, or is that a myth?
7) What’s a sane target for this stack? For example, cover 60 days of baseline expenses even if: portfolio is down 30%, HELOC is frozen, one bank account is locked for 7 business days, and you need to front-load a deductible. How would you size each layer to that spec?
8) Tax angle: any gotchas using margin or pledged lines to bridge a few days before T‑bill maturity? Phantom interest, wash sales, or state tax quirks when you later sell down?
9) Behavioral risk: credit card float is cheap liquidity, until it isn’t. Anyone use autopay plus a “liquidity escrow” to prevent running up balances during unemployment without noticing?
I’m not arguing yield doesn’t matter; I’m arguing sequence and operational reliability matter more. If we’re serious about resilience, shouldn’t the default advice be a time-bucketed liquidity plan across multiple rails and institutions, not just “dump it in a HYSA and hope the transfer clears”? Would love to see real data points, failure stories, and concrete setups that actually survived a frozen account, a blackout window, or a broker outage.
