Emergency Buffer Sizing

A plan or a bet

A plan that works only when everything goes right is not a plan. It's a bet — and life is a bad counterparty. It doesn't negotiate, it doesn't warn, and it collects in full.

This chapter is about the deliberately unoptimized corner of a well-run financial life: the cash buffer. It sits still. It earns modestly. By every spreadsheet instinct it's underperforming — and it is, mathematically, the cheapest insurance you will ever own. Understanding why requires taking seriously an idea the writer Morgan Housel calls room for error:¹ the deliberate choice to be less than fully optimized, so that a shock bends your finances instead of breaking them.

Why surprise is certain

The case for a buffer is routinely misargued as being about emergencies — as if the question were whether your transmission will fail this year. That's the wrong probability. Any specific emergency is unlikely on any given day; surprise itself is certain over time. Some appliance, some medical bill, some car, some roof, some job wobble — across years, the probability that nothing surprises you rounds to zero. The buffer isn't a bet that something specific goes wrong. It's an acknowledgment that something will, on a schedule you don't get to see.

And the cost of being unbuffered isn't the surprise's sticker price. Without slack, a $500 surprise doesn't cost $500 — it lands on a credit card at 24%, which kills your grace period (Credit Card Interest Mechanics's switch), which means every subsequent purchase accrues interest from day one, which starts the fee-and-interest cascade of Overdraft, NSF, and Late Fees, which can unwind months of payoff progress from Intelligent Avalanche–Avalanche vs. Snowball vs. Intelligent Avalanche. The $400-surprise statistic from the Federal Reserve — 37% of U.S. adults could not cover a $400 emergency with cash or its equivalent, a share unchanged since 2022² — is really a statistic about how much of the country pays the cascade price for every surprise, not the sticker price. The same $400, absorbed onto a card at 24% and paid at minimums, roughly doubles in eventual cost. The buffer's job, stated precisely, is to make bad luck boring: an event that moves a savings balance instead of an event that restructures your month.

How big — folklore versus your actual volatility

The folk rule — "3 to 6 months of expenses" — is fine as a destination and paralyzing as a starting requirement; for a household that can't cover $400, six months of expenses is a number so distant it functions as a reason not to begin. Two corrections make sizing practical.

First: the first $1,000 does the most work per dollar. Shock sizes aren't uniformly distributed — the common ones (repairs, copays, replacements) cluster in the hundreds. A starter buffer of $500–$1,000 intercepts the most frequent shocks and breaks the cascade cycle at its most common entry point. The marginal value of buffer dollars declines from there — which means the beginning, not the end, is the urgent part.

Second: past the starter level, the honest sizing input is your volatility, not a folk multiple. Stable salary, low fixed costs, two earners: the low end suffices. Variable income, dependents, a single income, older cars and an older roof: more. The question isn't "what do the rules say?" but "how low does my cash actually swing, and what's the largest surprise my life plausibly produces?" — which is a knowable number, not a vibe (and Plenee's cash-flow projections show your realistic floor across a typical month, turning buffer sizing from folklore into arithmetic on your own volatility).

And where: somewhere boring and reachable — a high-yield savings account earning real interest. Not checking (idle — Idle Cash's invisible fee). Not investments (which may be down exactly when you need them — the double-loss of selling depressed assets to fund an emergency). Boring, yielding, one day away.

The optimization trap

The buffer's enemies are not spendthrifts — they're optimizers. Every dollar in the buffer "should" be earning more somewhere, says the spreadsheet, and the spreadsheet is right until the day it's catastrophically wrong: the fully-optimized household, everything deployed at maximum yield, meets one surprise and finances it at 24% — converting years of optimization premium into one cascade. Optimization without slack is fragility wearing a costume. Room for error looks inefficient every single day except the one that matters; the buffer's return isn't its interest rate — it's the 24% cascade it prevents, times the certainty that surprise eventually arrives.