Most emergency fund arguments are about size. Placement gets less attention and costs real money. Between January 2022 and May 2023 the Bank of England base rate rose 4.25 percentage points. The nine largest UK savings providers passed 28% of that on to easy-access accounts. On fixed-term and notice accounts they passed on 51%. Instant access cost savers roughly half the rate rise.
That's the trade in one figure. Emergency cash sits at the intersection of four things: how fast you can reach it, what it yields, who can lose it for you, and how it's taxed. Rates move, so what follows sticks to structure — protection limits, fund mechanics, tax rules — rather than the rate table of the week. Sizing is a separate question, and the case for setting the target by income volatility instead of a flat three-to-six-month rule is covered in how income volatility changes emergency fund size.
Instant access: you pay for the instant part
The FCA's 2023 cash savings market review put figures on the cost. Easy-access accounts hold 60% of balances across the nine biggest firms. Their average rate went from 0.07% in January 2022 to 1.25% in May 2023. That's a rise of 1.18 percentage points against a 4.25 point move in base rate, so 28% pass-through. Between 2004 and 2009 the same firms passed on an average of 80%.
Fixed-term and notice accounts did better. Their average rate rose from 0.3% to 2.47%, an increase of 2.17 points, or 51% pass-through. The gap between what the big nine paid on easy access and on everything else widened from 0.23 percentage points to 1.22 over the same seventeen months.
Inertia does most of the damage. In the FCA's 2024 Financial Lives survey, 76% of adults with a savings account held it with their main current account provider. Easy-access accounts were held by 45% of all UK adults. Notice accounts were held by 3%. The default home for emergency cash is whatever savings account came bundled with the current account, and that's usually the worst-paying shelf in the shop.
The upside is real, though. The money is there the same day, the balance never falls in nominal terms, and deposit protection covers it. For a fund whose entire job is to be available at nine o'clock on a bad Monday, that isn't nothing.
Notice accounts: a cliff you choose in advance
A notice account pays more because you agree to wait. Thirty, sixty, ninety-five or a hundred and twenty days, depending on the product. The FCA's data shows the notice and fixed-term bucket capturing 51% of the rate rise against 28% for easy access. That premium is compensation for a genuine constraint.
The constraint is also the problem. A failed boiler doesn't wait ninety-five days. Notice accounts work as the back half of an emergency fund rather than the front: one instant tier covering the first month of expenses, a notice tier behind it. Splitting that way recovers most of the yield while keeping same-day money for the events that actually arrive without warning.
Very few households do it. Notice accounts were held by 3% of UK adults in 2024, and by 5% of those with any savings account at all. Fixed-term accounts, by contrast, doubled from 7% of adults in 2022 to 14% in 2024. Savers responded to higher rates by locking money up completely, not by tiering it.
Money market funds are not deposits
Sterling money market funds are the institutional answer to this problem, and they've become easier for individuals to buy. They're also structurally different from a bank account in ways that matter precisely when things go wrong.
Around 90% of sterling money market fund assets sit in funds domiciled in the EU, mainly Ireland and Luxembourg. Most of that money is in low volatility net asset value funds, or LVNAVs. An LVNAV quotes a constant price per unit, but only inside a collar. If the mark-to-market value drifts more than 20 basis points from the constant price, the fund has to switch to variable pricing. The stable pound is a convention with a trapdoor, not a guarantee.
Under the inherited rules, stable NAV funds hold at least 30% of assets in weekly liquid assets and 10% in daily liquid assets. If weekly liquidity drops below 30% and daily outflows exceed 10%, managers have to consider fees, gates or suspension. Below 10% weekly liquidity, fees or suspension become mandatory. Emergency money that can be gated isn't really emergency money.
March 2020 tested all of this. Sterling money market funds saw outflows of around £25 billion, roughly 11% of total assets, and three funds lost more than 20% of assets between 11 and 20 March. No sterling fund gated. The FCA's own view was that multiple suspensions could have posed a significant threat to wider UK financial stability. Redemptions came mostly from investors selling fund units to meet derivative margin calls.
The regulator's fix, confirmed in a statement on 8 June 2026, raises the expected weekly liquid asset holding to 40% for stable NAV funds and 20% for variable NAV funds. It also breaks the automatic link between liquidity thresholds and the imposition of fees or gates, which had turned the 30% level into a tripwire investors watched and ran from. Legislation repealing the inherited Money Market Funds Regulation was expected by the end of 2026.
Two things follow for household cash. There's no deposit protection on a money market fund: you own units in a vehicle holding bank paper and government bills, and the credit risk is yours. And individual retail investors are a small minority of the shareholder base, so behaviour under stress is set by corporate treasurers and pension schemes meeting margin calls.
Short-dated government debt
The cleanest credit available in sterling is the government itself. UK Treasury bills are the short end of that market. For a household the practical routes are a money market fund holding bills, a short-dated gilt fund, or individual gilts with a year or two left to run.
The tax treatment is where this gets interesting. Gilts have been exempt from capital gains tax since 2 July 1986, under section 115 of the Taxation of Chargeable Gains Act 1992. Coupon interest is taxable, but the pull back to par isn't. A low-coupon short gilt bought below par therefore delivers most of its return as a tax-free gain. For an additional-rate taxpayer with no personal savings allowance at all, that reorders the options more than any rate comparison does.
The catch is price risk. A gilt maturing in eighteen months moves with yields in the meantime. Hold it to maturity and you get par. Sell early into a rate spike and you don't. That's a different risk from bank default, and no compensation scheme covers it. The instinct that makes people hold cash instead of bonds, which is the wish for a number that never falls, is worth reading against the evidence on the difference between risk tolerance and risk capacity.
Premium Bonds: a lottery with a state guarantee
NS&I is the outlier, because it isn't covered by the FSCS at all. It's backed directly by HM Treasury, so protection isn't capped at £120,000. For anyone parking a large cash balance, that's a genuinely different risk profile from a bank deposit.
Premium Bonds are the best-known NS&I product, held by 26% of UK adults in 2024. As at 5 August 2026, the prize fund rate is 3.80% variable, with odds of 22,000 to 1 for each £1 bond in each monthly draw. The minimum purchase is £25 and the maximum holding is £50,000. Prizes carry no income tax and no capital gains tax.
That 3.80% is a mean, not an expectation for any one holder. NS&I allocates 10% of the prize fund to prizes of £5,000 and above, 10% to £500 and £1,000 prizes, and 80% to prizes of £25, £50 and £100. The July 2026 draw paid 6,226,179 prizes worth £433,757,200 in total, including two £1 million jackpots. Hold £1,000 of bonds and the expected number of prizes across a full year is about 0.55. Most years you'd win nothing at all. The skew means the median holder earns less than the headline rate, and small holders earn a lot less.
Set against Bank Rate, which has been 3.75% since 18 December 2025, and NS&I's own Direct Saver at 3.45%, the prize fund rate looks competitive. For a higher-rate taxpayer who has used up the £500 personal savings allowance, a tax-free 3.80% expected return is worth more still. For someone holding £2,000, variance swamps the arithmetic entirely.
Deposit protection is now the easy part
The FSCS limit rose to £120,000 per eligible person per bank, building society or credit union on 1 December 2025. The old limit was £85,000, set in 2017. The PRA had originally proposed £110,000 and raised the figure after consultation feedback and newer inflation data. Temporary high balances, such as house sale proceeds, redundancy payments and insurance settlements, are protected up to £1.4 million for six months from the date the money was first deposited.
Two traps survive the increase. Protection applies per banking licence rather than per brand, so two accounts sitting under one licence share a single £120,000 ceiling. And awareness is patchy. Only 72% of adults with a savings product knew the FSCS existed when the FCA asked in 2024.
Tax reorders everything
The personal savings allowance is £1,000 for basic-rate taxpayers, £500 for higher-rate taxpayers and nothing for additional-rate taxpayers. A separate starting rate for savings gives up to £5,000 tax-free, but only where other income is below £17,570, and it tapers pound for pound above the £12,570 personal allowance. The ISA allowance is £20,000 for 2026-27 across all ISA types combined.
So the ranking flips by taxpayer. A basic-rate taxpayer with £15,000 in an account paying 4% earns £600 of interest and pays no tax on it. An additional-rate taxpayer with the same balance keeps 55% of that interest. For the second person, a cash ISA, Premium Bonds or a low-coupon short gilt can beat a higher headline rate elsewhere. Comparing gross rates across wrappers with different tax treatment repeats the error described in what a 0.2% fee versus 1% costs over thirty years, where a small annual difference decides the outcome.
The case against optimising this at all
Critics of the whole exercise make a fair point. An emergency fund exists to be spent at the worst possible moment, and every rung up the yield ladder adds a way for the money to be unreachable exactly then. Notice periods bind. Money market funds can gate. Gilts fall in price when yields rise. Chasing sixty basis points across five accounts creates admin that people quietly abandon, and abandoned optimisation reverts to the worst rate on the shelf.
The counter-evidence is about magnitude. The FCA's own numbers put the easy-access penalty at 1.22 percentage points against other account types by May 2023. That isn't twenty basis points. On £20,000 it's £244 a year for doing nothing. Splitting a fund into an instant tier and a notice tier doesn't remove the instant tier either. It stops the whole balance sitting in the lowest-paying option.
Where the critics are plainly right is the stopping point. A single instant-access account inside the protection limit, at a competitive rate, captures most of the available value with none of the failure modes. The marginal gain from the fifth refinement is small and the marginal fragility isn't. Where the argument breaks down is in treating the default account as the safe choice. Sitting at 0.07% while base rate is 4.25 points higher is a decision too, just an unexamined one.
What would change the conclusion
Four things would.
If pass-through improved so that easy-access rates tracked base rate closely, the case for tiering weakens on its own. The FCA reported pass-through accelerating to 35% between August 2022 and May 2023, up from 14% in the seven months before that. Continued improvement shrinks the prize.
If the money market fund reforms land as drafted, with 40% weekly liquid assets for stable NAV funds and no automatic link to gates, the gating risk that argues against these funds for household cash falls. The rules weren't final when this was written, and the repeal legislation was still pending.
If Bank Rate falls a long way, absolute spreads compress and the effort stops paying for itself. At 3.75% a gap of 1.2 points is worth chasing. At 0.5% it isn't.
And if protection limits move again, the case for spreading balances across institutions changes with them. The 2025 increase from £85,000 to £120,000 already solved that problem for most households in a single step. The placement question that remains is about yield and access, not about default risk.