On this page8 sections
Use a card for anything with a fixed amount and a paper trail — rent, bills, insurance, subscriptions, fuel, anything bought online. Use cash only for the one or two categories where you personally overshoot every month, and only after you have checked that the shops you use still take it.
That is the recommendation, and it is deliberately narrow, because the case for cash is narrower than most budgeting content admits. The claim that paying in notes makes you spend less rests on a small, decades-old body of work that has never been settled. The claim that a card can cost you real money when you carry a balance rests on quarterly data from central banks. One of those is a hypothesis. The other is a bill.
Cash, debit and credit as budgeting tools
The trade-offs that actually differ, rather than the ones that sound different
- Friction per purchaseHighest
- Can you exceed the limit?No — when it's gone it's gone
- Automatic recordNone. You log it or lose it
- If it's lost or stolenGone
- Works onlineNo
- Dispute routeNone
- Debt riskZero
- Friction per purchaseLow — a tap
- Can you exceed the limit?Only via an overdraft
- Automatic recordEvery transaction, dated
- If it's lost or stolenFreeze it in the app
- Works onlineYes
- Dispute routeScheme chargeback, weaker than credit
- Debt riskOverdraft only
- Friction per purchaseLowest
- Can you exceed the limit?Up to the credit limit
- Automatic recordEvery transaction, dated
- If it's lost or stolenNot your money at risk
- Works onlineYes
- Dispute routeStrongest — statutory in some markets
- Debt riskHigh if you revolve
What the research actually shows, and what it does not
The study everyone quotes is real. Drazen Prelec and Duncan Simester of MIT ran sealed-bid auctions for genuine prizes — tickets to a sold-out Boston Celtics game and a Boston Red Sox game — and found that bidders told to pay by credit card were willing to pay more than bidders told to pay in cash. It was published as Always Leave Home Without It in Marketing Letters in 2001.
Now the caveats that most articles skip. The samples were small. The setting was a sealed-bid auction for a scarce, emotionally charged prize, which is about as far from a Tuesday supermarket shop as a purchase gets. The comparison was credit card against cash, not debit against cash. And it was run in a world before contactless, mobile wallets, or a banking app that shows your balance falling in real time. A quarter of a century later it remains the single most cited result in this area, which tells you something about how little else there is.
The mechanism people attach to it is the "pain of paying" — the idea, developed by Ofer Zellermayer in the mid-1990s and formalised by Prelec and George Loewenstein in 1998, that handing over physical money hurts in a way that a tap does not, and that the hurt acts as a brake. It is a plausible model. It is not a measurement, and later attempts to pin down its size have not converged on an answer. The honest position is that the direction of the effect is credible and the size of it is unestablished.
There is a useful reality check from the adjacent literature on budgeting tools, which asks a very similar question — does changing the interface change the spending? A randomised field experiment with 9,035 users inside a fintech app — run by a behavioural science lab and never peer reviewed — found no significant difference in spending between a control group and two budgeting conditions over 13 weeks: $675.97 in the control group against $681.08 and $673.25 (Irrational Labs / Common Cents Lab). The same study found budgeters overshot their own budgets by 1.3 to 1.4 times. A peer-reviewed UK randomised trial of money-management apps among credit union members in Derry found the people given apps were more likely to say they kept track of their income and spending, and proved more resilient to a financial shock — but the authors were explicit that this did not translate into improved financial wellbeing over the six-month window (French, McKillop & Stewart, European Journal of Finance).
Neither trial is about payment method. Both are about the same underlying bet: that a different way of seeing your money changes how much of it you spend. The effects found are small, and in one case zero. Treat any confident claim that switching to cash will cut your spending by a specific percentage as marketing.
The number on the other side is not a hypothesis
Where the arithmetic is unambiguous is the cost of using a credit card badly.
US commercial banks charged an average of 22.15% in the second quarter of 2026 on credit card accounts that were actually assessed interest, according to the Federal Reserve's G.19 release. In the UK, the Bank of England's representative quoted rate on card lending was 24.71% in July 2026 (Bank of England) — a rate that has barely moved in two years. Neither is a rate you pay if you clear the balance every month. Both are what you pay if you do not.
The arithmetic, as an illustration on a round number: revolve $1,000 for a year at 22.15% and you pay roughly $222 in interest; revolve £1,000 at the 24.71% quoted rate and you pay roughly £247. That is a worked example, not a quote — it is before compounding, before any fee, and the UK figure is a representative advertised rate rather than the rate your own card charges.
This is not a rare failure mode. About half of all US credit card accounts carry a balance from month to month, and US consumers were charged $160 billion in credit card interest in 2024, up from $105 billion two years earlier (CFPB, Consumer Credit Card Market Report). In the UK, the FCA's Financial Lives 2024 survey found 19% of adults — 10.1 million people — had revolved a credit card balance in the previous 12 months, which is 31% of the 33.4 million people who had used a card or carried a balance (FCA).
What a carried balance costs, by market
Source: Federal Reserve Board, G.19 Consumer Credit, Accounts assessed interest, 2026 Q2 (preliminary); Bank of England, quoted household interest rates, Representative credit card rate, July 2026; CFPB, The Consumer Credit Card Market: Report to Congress, 2024 calendar year
If that describes you, the payment-method question is a distraction and the payoff question is the one that matters. Work the balance down first with the credit card payoff calculator, and pick an order using snowball versus avalanche.
The rules are not the same where you live
The single biggest error in most cash-versus-card articles is quoting one country's consumer-credit law as though it were physics. It is not.
United Kingdom. Section 75 of the Consumer Credit Act 1974 makes your credit card issuer jointly liable with the retailer for certain purchases within a set value range, which is why a credit card is genuinely safer than a debit card for large or risky buys. The thresholds and exclusions are specific — check the current figures with the FCA or a consumer body before you rely on them, and note that debit cards are covered only by voluntary scheme chargeback rules, not by the Act. Ireland, despite the shared border and much shared retail, has its own consumer credit regime, not Section 75.
United States. There is no Section 75 equivalent. The Fair Credit Billing Act and Regulation Z give you billing-error rights, a cap on liability for unauthorised charges, and the right to withhold payment on a disputed purchase in defined circumstances — a different mechanism reaching a broadly similar place for fraud, and a weaker one for a retailer that simply fails to deliver. Card surcharging and cash discounts are also legal in most US states and banned on consumer cards in the UK and EU, which changes the maths on paying cash at a small business.
Canada and Australia. Purchase protection runs largely through Visa and Mastercard scheme rules and the individual card's benefits schedule rather than statute, so two cards from the same bank can offer different cover. Read the benefits guide rather than assuming.
The other thing that varies is how much rope a card gives you before the damage is serious. Household debt as a share of net disposable income differs by a factor of more than two across these markets.
Household debt as a share of net disposable income, 2024
- AUAustralia210%of net disposable income, 2024OECD, National Accounts at a Glance
- CACanada181%of net disposable income, 2024OECD, National Accounts at a Glance
- UKUnited Kingdom131%of net disposable income, 2024OECD, National Accounts at a Glance
- USUnited States99%of net disposable income, 2024OECD, National Accounts at a Glance
- IEIreland86%of net disposable income, 2024OECD, National Accounts at a Glance
Cash stuffing: the honest version
The cash envelope revival — filling labelled binder pockets with notes on payday and filming it — is where most of the current interest in this question comes from. It is the envelope system with better production values, and the core mechanic is sound: pre-committing a category to a fixed sum, then meeting a physical wall when it runs out.
Three objections get raised. Two of them are real.
"You lose the interest." This one is not real, and it is worth killing with arithmetic. Say you keep a £400 cash float across your envelopes. The Bank of England's effective rate across the existing stock of household instant-access balances was 1.65% in June 2026 (Bank of England) — so that float would have earned about £6.60 over a year sitting in a typical account. Even at the 4.30% average on newly opened fixed-term deposits, which you could not use as a spending float anyway, it is about £17. In the US the FDIC national average savings rate was 0.38% in July 2026 (FDIC), so a $400 float forgoes about $1.52 a year. Forgone interest is not the argument against cash stuffing.
"It gets stolen." This one is real, and it is the objection with teeth. There is no dispute route, no freeze button, and no insurance on a binder in a bedroom drawer beyond whatever cash limit your home policy carries — which is usually low. Keep the float to a week rather than a month if you go this route.
"Half the shops won't take it." Also real, and increasingly the binding constraint. Card-only cafés, ticketless car parks, self-service checkouts that take notes but not change, and any online purchase at all. A cash system that forces you to break the system twice a week is a system you will abandon by week three, which is the usual failure mode described in common budgeting mistakes.
The middle path most people end up on
The cash-or-card framing is a binary that the market has moved past. What actually works for most households is a digital envelope: the pre-commitment of cash, with the record-keeping of a card.
Three versions, in increasing order of effort:
- A second account. Open a fee-free current or checking account, move the month's discretionary money into it on payday, and carry only that card. When it hits zero, the category is over. This reproduces the hard limit almost exactly and keeps every transaction logged.
- A per-category limit in your budget app. Less friction than a separate account, but you get an alert before the wall rather than at it. Best for people who respond to a warning; useless for people who dismiss notifications.
- A virtual or per-merchant card. Increasingly available from mainstream banks and challengers. Useful where one specific merchant is the problem.
A note on prepaid cards, which get recommended constantly for this job: many carry load fees, monthly fees, ATM fees, or inactivity fees that will quietly outweigh the £6.60 of interest you were worried about. In most markets a second fee-free account does the same job for nothing. Check the fee schedule before you check the marketing.
A decision path
Which setup you should actually run
- 1. Are you carrying a credit card balance?If yes, stop here. Move all spending to debit or cash, keep the card only for the balance you are clearing, and treat payoff as the project. The interest dwarfs any behavioural gain from payment method.
- 2. Do you overshoot a specific named category?Not 'I spend too much' — a named category, two months running. If no named category, you have a tracking problem, not a payment-method problem. Card for everything and review weekly.
- 3. Do the merchants in that category take cash?Coffee, corner shop, pub, market: usually yes. Delivery apps, online, subscriptions, most city car parks: no. If no, cash is not available to you here and a second account is the answer.
- 4. Run the narrow versionCash for that one category, weekly withdrawal not monthly, card for everything else. Review after two months and either extend it or drop it.
Two things that decision path deliberately refuses to do: recommend cash to someone who has not identified a specific problem category, and recommend it to anyone carrying card debt. If you cannot name the category, start with the category list and two months of plain tracking first.
Tracking cash without hating it
Cash's one genuine defect as a budgeting tool is that it leaves no record, so a cash-heavy month is a hole in your data exactly where you most needed the data. Three habits close it:
- Log the withdrawal, not the purchases. Treat the £120 you take out on Saturday as one transaction against the category. You lose the detail, you keep the total, and the total is what the budget needs.
- Log at the point of withdrawal. Not the evening, not the weekend. The recall gap is where cash budgets go wrong.
- Reconcile the envelope once a week. Count what is left, compare it against what the app thinks is left, and adjust. A week's drift is fixable; a month's is fiction.
If you are building the system from scratch rather than patching one, the complete budgeting guide covers the structure this fits into, and how to stick to a budget covers the part that fails first. If cash is your response to running out of money before payday rather than to one runaway category, the underlying fix is in stop living paycheck to paycheck.
When each method genuinely fails
Cash fails when your problem categories are online, when you travel, when you are trying to build a credit history (cash builds none, and neither does a debit card), when you need a dispute route on a large purchase, and when carrying it is unsafe. It also fails quietly when the tracking discipline slips, which is the most common way it ends.
Cards fail when you revolve, when the friction is genuinely too low for a specific habit, and when the statement is the only feedback you get — a month is far too long a loop for a category you overshoot in week two.
There is no universal answer, and anyone selling you one has not looked at the evidence. What there is: a default that suits most people most of the time, and a narrow, deliberate exception where you have shown yourself the exception is needed.
Put this into practice
iBudget lets you log cash withdrawals and card transactions against the same categories, so a cash envelope stops being a blind spot in your numbers.
Start budgeting free — free plan, no card required, no bank logins.
About iBudget
iBudget helps couples and families take control of their finances with simple, collaborative budgeting tools. Track spending, set goals, and build wealth together.
Start Your Budget
