How Many Parties Can Freeze My Money?

The household audit in June found sixteen findings and four criticals, and I noted then that I had skipped money because it deserved its own entry. Water got the same treatment and turned out to be the most embarrassing omission on the list. This one is worse, because unlike the water main, money has chokepoints that can act deliberately.

The question is not "is my bank safe." Banks are, on the whole, boringly safe with deposits. The question is: how many separate parties can decide, without warning and without my consent, that my money stops moving today? And when one does, how long does the household run?

The count

Chokepoint	What it stops	Notice required	Household runway without it
Primary bank	Rent, cards, transfers, everything	None	19 days
Card network	Day-to-day spending	None	Indefinite, degraded
Payroll provider	Income arriving at all	None	41 days
Bill payment processor	Recurring obligations	None	60+ days, with phone calls
Brokerage	Access to long-term savings	None	Not a runway question
Self-custodied reserve	Nothing. No third party.	n/a	n/a

Five parties can interrupt the flow of my money. Four of them can do it with no notice, no hearing, and no obligation to explain, under authorities that exist for good reasons and are applied by systems that make mistakes at scale.

I want to be careful here, because this is the topic where my genre goes silly. Account freezes are not common and they are not a conspiracy. Most are automated flags on patterns that genuinely correlate with fraud, and most resolve. The reason to plan for one is not that it is likely. It is that the probability is not negligible, the impact is total, the duration is measured in weeks, and the resolution is outside your control. That combination is exactly what a household should carry redundancy for, and almost nobody does.

The runway numbers

Nineteen days is the primary bank figure and the one that made me act: the days the household covers rent, food, fuel, and obligations on cash plus the secondary account, without touching investments or borrowing.

It is not enough. Published resolution times for a flagged account, when I asked institutions directly, ranged from 3 business days to "we cannot commit to a timeframe." A month is plausible. The target is 45 days and the gap gets closed with boring, unglamorous cash.

The mitigation that nearly did not work

I opened a second account at a different institution in 2024 and considered the problem solved. When I actually audited it this month, I checked something I had never thought to check: who runs the software.

Both institutions ran on the same core banking platform, from the same vendor.

That does not mean one freeze causes the other. Different institutions, different compliance departments, different decisions. But a class of failure I thought I was hedged against, a platform outage or vendor incident, would have taken both accounts down together while I sat there admiring my redundancy.

This is the outage ledger's lesson arriving in my own finances. Two services that write down the same building are one service. I moved the secondary account in June to an institution on a different core, which cost an afternoon and three forms, and I would not have known to do it if I had not asked who the vendor was. It is not a question consumers are encouraged to ask. Ask it anyway.

Self-custody, priced honestly

A portion of household reserves is self-custodied and I drill quarterly, and the drill times are already on this site. What is worth adding is the trade in its unflattering form, because this conversation is usually conducted by people selling something.

Self-custody removes counterparty risk and replaces it with operator risk. No institution can freeze it and no institution can help me when I make a mistake. The failure mode moves from "a third party stops me, temporarily, with an appeals process" to "I lose it, permanently, with nobody to call." Those are not the same risk and I do not think one is universally better.

My loosely held estimate: an account freeze affecting me in any given decade sits in the low single digits of percent. Losing self-custodied funds through my own error over the same period is lower, but only because I drill and have performed a full recovery three times. Strip out the drilling and my own error is the larger risk by a wide margin. The practice is not a detail. The practice is the entire thing.

Which is why the reserve is a portion and not the whole. Concentration in the institution is a risk. Concentration in me is also a risk. A household that has replaced five chokepoints with one chokepoint named "the guy who weighs his coffee" has not decentralized anything.

The findings

Three actions, two done:

1. Move the secondary account off the shared core platform. Done, 2026-06-22.
2. Raise cash on hand until the primary-bank runway clears 45 days. In progress, currently 31 days, on track for September.
3. Move one recurring obligation to the secondary account so it is exercised monthly rather than sitting unused. Done, 2026-07-06.

That third one is the restore-drill principle applied to banking, and it is the finding I would hand to someone with no interest in any of this: a backup account you have never paid anything from is a rumor. Institutions close dormant accounts. Cards expire unnoticed. Login credentials rot. Run something through it every month or you do not have a second bank, you have a second login page.

Confidence 0.83 on the runway figures, which is deliberately the lowest number I have published. Runway estimates depend on how the household actually behaves under stress rather than how I modeled it at a desk, and I have never tested that.