Money in a savings account paying less than inflation loses purchasing power every year, and it does so with more certainty than any investment loses money. At 2 per cent interest against 3 per cent inflation the real return is about −1 per cent a year — over a decade that is roughly a tenth of the value gone, guaranteed.
That is not an argument against holding cash. It is an argument for holding the right amount, because the certainty cuts both ways: cash is the only asset that is definitely there when you need it.
How much should sit in cash?
Three to six months of essential spending is the standard answer, and the range exists because the right figure depends on how replaceable your income is.
| Situation | Months |
|---|---|
| Two stable incomes, in demand | 3 |
| Single income, stable | 4–6 |
| Variable or commission income | 6–9 |
| Self-employed, lumpy income | 9–12 |
Essential spending rather than total spending is the figure to use — housing, food, utilities, transport, minimum debt payments and insurance. It is usually far lower than what you actually spend, which makes the target smaller and more achievable than people expect.
Why hold it at all if it loses value?
Because the alternative to cash is not a better return, it is selling something at the worst possible moment. An emergency fund exists so that a boiler failure or a redundancy does not become a forced sale of investments in a bad market, or a balance on a card at 22.9 per cent.
Measured against that, a 1 per cent real loss on three months of spending is cheap insurance. The credit card article covers what the uninsured version costs.
What about irregular but predictable costs?
They are a separate problem with a separate answer: a sinking fund. Insurance renewals, car servicing, Christmas, a boiler service and replacement of anything with a known life are not emergencies — they are known costs with unknown timing.
Dividing the annual total by twelve and moving it monthly turns a series of shocks into a flat line. The arithmetic is trivial and the behavioural effect is not: a 600 annual insurance bill is a crisis in one month and 50 a month is a budget line.
Keeping them separate from the emergency fund is what stops the emergency fund from being quietly consumed by things that were never emergencies.
Does the account matter?
More than the difference between accounts suggests, because the base rate is doing most of the work. What matters is not chasing the top of the table but avoiding the bottom of it — the gap between a legacy account paying almost nothing and a competitive one is frequently several percentage points.
Access is the constraint that decides the rest. An emergency fund in a notice account is not an emergency fund; the sinking fund can be, because its timing is known.
Tax treatment changes the comparison in some jurisdictions, and the after-tax rate is the one to compare against inflation.
Where should the fund actually live?
Somewhere boring, separate, and reachable within a day or two. Separate matters more than the rate: money in the current account is spent, and money one transfer away is not.
Splitting it is a reasonable refinement — a smaller instant-access tier for the things that happen this week, and the rest somewhere paying slightly more with a short delay. What defeats the purpose is any arrangement where accessing it costs a penalty, because the penalty will land at exactly the worst moment.
When does cash stop being the right answer?
Past the emergency fund and the sinking funds, for money you will not need for years. Over long horizons the certainty of a small real loss compounds into a large one, and the compounding article shows what the same money does at a positive real rate.
The dividing line is the time horizon rather than the amount. Money needed within a couple of years belongs in cash whatever the rate; money not needed for a decade is being actively eroded there.
Does the amount change when rates are high?
The target does not — it is denominated in months of spending rather than in a return. What changes is how much the fund costs you to hold: at 5 per cent interest against 3 per cent inflation, cash earns a positive real return and the trade-off largely disappears.
That is the period in which holding a slightly larger buffer is close to free, and the period in which people most often move the money into something else. The decision should follow the horizon rather than the current rate, because rates change faster than the reason the fund exists.
Questions people ask
How do I calculate the real return? Approximately, subtract inflation from the interest rate. Precisely, divide: (1 + rate) ÷ (1 + inflation) − 1, which matters once the numbers get large.
Is inflation the same for me? Not exactly — the CPI article covers why a national index and a household experience diverge, mostly through housing.
Should the emergency fund grow with income? With essential spending, which usually grows more slowly. Recalculating annually is enough.
What about keeping some in cash at home? A small amount is genuinely useful for a power or systems outage. Beyond that it earns nothing, is uninsured and is at risk in ways a bank balance is not.
Does an offset account change the answer? It removes the tax question entirely: offsetting earns the mortgage rate tax-free by reducing the interest charged rather than by generating interest, which is why it often beats a savings account paying the same rate.
Hold enough to avoid a forced sale and no more than that. The inflation calculator shows what a real loss compounds into, the savings calculator covers the growth side, and the budget calculator is where the monthly essential figure comes from.