The fair way to split a trip abroad: each expense at its own day’s rate
Here’s the short answer, because the common trip-splitting advice gets it wrong: on a multi-day trip in a foreign currency, convert each shared expense at the exchange rate on the day it happened - not one blended rate at the end. A dinner you fronted on Monday and a rental car you fronted on Saturday were paid in two different currencies as far as your bank is concerned, because the rate moved in between. Collapsing them into a single rate silently reassigns who owes what.
This isn’t a fussy edge case. It’s the exact problem accounting standards and tax authorities already solved - and the rule they landed on is the opposite of the popular “pick one base currency and convert at the end” advice you’ll find on trip-splitter sites.
The one-rate trap: Whoever fronted expenses on the days the rate was worst absorbs the drift. It’s invisible, it’s systematic, and over a week-long trip it lands on the same person every time - the one who volunteered to pay.
Why one blended rate quietly overcharges someone
A group trip isn’t one bill. It’s a ledger of expenses spread across days: dinners, taxis, a museum, the hotel, a rental car. Each one is paid on a specific date, and on each date the exchange rate between your home currency and the local one is slightly different. Currencies float. They move every single trading day.
When the trip ends and one person tallies everything, the tempting shortcut is to grab today’s rate - or an average - and convert the whole ledger at once. That shortcut assumes every expense happened at the same rate. It didn’t. The person who paid the 30,000-yen dinner on the day the yen was strong against the dollar paid more dollars than the person who paid an identical 30,000-yen dinner three days later when it had weakened. A single blended rate erases that difference and hands it to whoever was on the wrong side of it.
What actually happened:
Monday dinner = 30,000 yen at Monday’s rate
Saturday car = 30,000 yen at Saturday’s rate
Two different dollar costs - same yen amount.
What the one-rate split pretends:
Both = 60,000 yen at one rate.
The drift between Monday and Saturday? Absorbed by the fronter.
The size of the error scales with two things: how much the rate moved during the trip, and how lopsided the fronting was. A group where one person paid everything on the volatile days is exactly the group where the one-rate method is least fair.
Accounting already settled this: use the rate on the transaction date
You don’t have to take a bill-splitting app’s word for the right method. International accounting has a formal standard for exactly this question - how to record a transaction that happened in a foreign currency - and it’s unambiguous.
IAS 21, the international standard on the effects of changes in foreign exchange rates, states the rule plainly:
“A foreign currency transaction shall be recorded, on initial recognition in the functional currency, by applying to the foreign currency amount the spot exchange rate between the functional currency and the foreign currency at the date of the transaction.”
IAS 21, paragraph 21, IFRS Foundation
The “spot exchange rate at the date of the transaction.” Not the average. Not the end-of-period rate. The rate on the day the money changed hands. And the standard goes further - it explicitly warns against the exact shortcut the trip-splitter advice recommends:
“For practical reasons, a rate that approximates the actual rate… is often used, for example, an average rate for a week or a month… However, if exchange rates fluctuate significantly, the use of the average rate for a period is inappropriate.”
IAS 21, paragraph 22, IFRS Foundation
Read that last line again. When rates move meaningfully, an averaged or single rate is inappropriate - the standard’s own word. A blended rate is only a convenience, and only when the currency barely budged. On a real trip, over real days, currencies budge.
This isn’t an accountants-only convention, either. The US tax code reaches the same answer for anyone spending in a foreign currency: the IRS instructs taxpayers to “use the exchange rate prevailing when you receive, pay, or accrue the item.” Prevailing when you pay - that’s the transaction-date rate, in plain language, from the agency that audits it.
Sources: IFRS Foundation, IAS 21 “The Effects of Changes in Foreign Exchange Rates”, paragraphs 21-22; Internal Revenue Service, “Foreign currency and currency exchange rates.”
How much can a rate really move in a week?
Enough to matter. Exchange rates between major currencies drift continuously, and the drift is not a rounding error you can wave away. At the extreme, the European Central Bank described how violent that movement can get in a crisis:
“Foreign exchange volatility among major currency pairs has risen to its highest peak since 1973, reflecting broader and more rapid movements in the main bilateral rates than in previous episodes.”
European Central Bank, Monthly Bulletin, February 2009
That was an exceptional episode - but the everyday point still holds. Even in calm periods, the ECB noted, bilateral rates like the dollar-euro move on their own schedule, “displaying little association with the global business cycle.” They drift for their own reasons, on their own timing, which is why you have no reliable way to know in advance which days of your trip will land a favorable rate. The rate simply is what it is on each date - which is precisely why each date needs its own conversion.
And it never sits still. Foreign exchange is the most heavily traded market on earth - the Bank for International Settlements measures its turnover every three years in the Triennial Central Bank Survey - and it reprices continuously through every working day. The rate on your Monday dinner was quoted, filled, and gone before your Saturday car rental existed.
Sources: European Central Bank, “The recent exceptional rise in exchange rate volatility,” Monthly Bulletin, February 2009, Box 6; Bank for International Settlements, Triennial Central Bank Survey, 2022.
A worked example: the same ledger, two methods
Take a simple trip. Three of you travel together; you front three shared expenses over a week in a country whose currency is weakening against the dollar - it takes more of it to buy a dollar as the week goes on. The local amounts are identical - 30,000 each - but the rate you actually paid moved between them. (Rates below are illustrative, chosen to show the mechanic.)
You paid $612.41 - so a fair three-way split is $204.14 each. Now watch what a single end-of-trip rate does. By Saturday the currency has weakened to 150/$, so converting the whole 90,000 at that one rate values the trip at just $600 - each “share” becomes $200, and your two friends hand you $400. You spent $612.41. You just ate $8.27 more than your fair share, while each friend quietly underpaid by about $4.14. Had the currency moved the other way, the error would flip to their disadvantage - but you, the fronter, are the only one exposed to it at all.
The per-date method is the only one where the total that gets split equals what was actually spent. The average lands close here only because the rate moved smoothly - let it jump on one day and the average drifts too. Everything except the transaction-date rate is a small wealth transfer wearing the costume of convenience.
Who absorbs the drift - and why it’s always the same person
When one person fronts the shared expenses - the usual case - the loss isn’t random. It concentrates on whoever volunteered to be the group’s bank. They fronted the expenses, so they’re the only one exposed to the gap between the real transaction-date rate and whatever single rate the group later agrees on. Everyone else pays a clean, round number and moves on. (Spread the fronting across several people and the exposure spreads with it - but someone is always holding it.)
This is the same asymmetry that makes fronting a group trip quietly expensive, layered on top of currency. The fronter already carries the float - the weeks of waiting to be repaid. Per-date conversion at least makes sure that when the repayment comes, it’s for the amount they actually lost, not a rate-flattened approximation.
Paid every expense at its real date-rate. A single settlement rate over- or under-pays them by the full amount the currency moved across the trip.
Pays one tidy figure computed from one rate. Never sees the day-to-day movement, never feels the gap.
The amount split equals the amount spent. No one profits, no one subsidizes - the rate stops being a hidden lever.
How to actually do it (without a spreadsheet meltdown)
The honest method sounds like more work than it is. You don’t need to be an accountant - you need to capture the date, and let the rate follow from it.
Log each expense on the day it happens
The date is the whole game. An expense recorded the moment it occurs carries its own rate automatically. A pile of receipts sorted at the airport on the way home does not.
Use that date's rate, not today's
Your card statement is the ground truth - it already converted each charge at the transaction-date rate. For cash, central banks publish a daily reference rate for every working day (the ECB, for one, posts euro rates each afternoon) - use the one for that date. One rate per expense, keyed to its day.
Settle in one currency, converted per expense
Everyone can still pay you back in dollars. The point isn't the settlement currency - it's that the dollar figure is the sum of each expense at its own rate, not the whole ledger at one rate.
Settle while the dates are still fresh
Each expense’s dollar cost was fixed on its transaction date - waiting doesn’t change what your card already charged. What waiting erodes is the will to be precise: the longer the ledger sits, the more tempting it is to abandon the per-date detail and reach for one stale rate to “just even it up.” Settling promptly locks in the honest per-date numbers before that happens.
Rotate who fronts on long trips
If you travel together often, spread the bank role across trips. Per-date conversion removes the systematic loss; rotating removes the systematic burden of being the one exposed to it.
From a week of drifting rates to one honest number
The reason the one-rate shortcut wins is friction: converting each expense at its own date, by hand, across a week and three people, is genuinely annoying. So people don’t - and the fronter eats the difference. The fix isn’t to bolt currency accounting onto a bill-splitter - it’s to use the right tool for each half. A dedicated trip tracker like Splid is built to hold a multi-day, multi-currency ledger. splitty is built for the other half: the single shared bill in front of you, split fairly and settled on the spot.
Splitting a trip fairly across currencies isn’t about predicting the market. It’s about refusing to pretend six days happened at one rate. Settle each shared bill as it happens - a trip tracker like Splid for the running ledger, splitty for the bill in front of you tonight - and the drift stops being a tax on whoever held the card.
