Even With Free, Instant Transfers, Most People Still Don't Move Their Money

Take a case built specifically to remove every excuse. Same bank. Same risk. Same maturity. A tap on a phone screen moves money between accounts, and it lands the same day, for no fee. There is no application to fill out, no new provider to vet, no waiting period. If friction has been the real obstacle to people earning a better rate on their cash, this is the setting where it should finally show up as a solved problem.

It didn't. Most households still left their money in the lower-paying account anyway.

That finding comes from Fernando Cirelli and Arna Olafsson, using transaction-level data from a major Icelandic bank that serves roughly a third of that country's population, published as CEPR Discussion Paper 20612 in 2025. The one group that did respond was the wealthiest households, who moved money toward higher-yielding accounts roughly ten times more than the average household and earned about two percentage points more on their liquid savings as a result. Everyone else, across a national banking relationship with a third of a country's population behind it, barely moved at all.

Why this result is worse news than it sounds

A UK field experiment covering 124,000 savings-account holders already found that giving people direct information about a better rate, in one case a rate at their own current bank, produced almost no switching. That result ruled out ignorance as the explanation. It left open the possibility that the real obstacle was something more mundane: the actual mechanics of switching, the paperwork, the wait, the fifteen minutes of hassle.

The Icelandic data closes that door too. Nothing about switching in this setting was slow or costly. The obstacle that gets blamed most often in ordinary conversation, and the one a rate comparison site's whole design implicitly assumes it is fighting, was not present, and the non-response happened anyway.

Two different countries, two different obstacles ruled out one at a time (information in the UK, friction in Iceland), and the same basic result both times: most people don't act, even when acting costs them almost nothing.

What the wealthy did differently

The exception matters as much as the rule. Wealthy households in this dataset were roughly ten times more responsive to rate differences than the average household, and that responsiveness translated into hard numbers: about two percentage points more in annual return on their liquid savings. Business-cycle swings in how much money sits in low-rate accounts were driven almost entirely by this group actively reallocating, not by everyone adjusting a little.

Put a rough dollar figure on the gap, drawing on the same paper's own consumption-based framing (with the caveat below about applying it to US data): the researchers estimate that a typical, median household leaves interest income worth about 0.3% of its annual consumption on the table, while wealthy households leave closer to 2.5%. Using US Bureau of Labor Statistics figures for average annual household spending in 2024, that is roughly $236 a year for a typical household (0.3% of $78,535 in average annual expenditures across all consumer units) and roughly $3,759 a year for a wealthier one (2.5% of $150,342 in average annual expenditures for the highest income quintile). That second calculation combines an Icelandic behavioral finding with US consumption data as an illustrative estimate, not a measured US figure, and it should be read that way. Even treated as a rough estimate, the shape of it holds: the group with the most to gain in absolute dollars is also the group already best positioned to act, and mostly does. Everyone below that tier is left with a gap that persists whether or not the friction to close it exists at all.

The same paper offers a partial explanation for why wealth predicts action here, and it isn't simply "more money makes acting worthwhile." Survey evidence tied to the dataset found that financial literacy and accurate knowledge of inflation, not wealth by itself, are associated with stronger reallocation toward higher-yielding accounts. Wealth and financial literacy correlate closely enough in most populations that it's easy to mistake one for the other, but the mechanism the researchers point to is understanding, not balance size. That distinction matters for what a fix might look like: a wealthier household isn't simply less annoyed by the switching process. It more often already carries a working model of why a rate gap is real and durable, which is closer to a corrected belief than a bigger bank balance.

A US version of the same test

The closest US analogue isn't a different bank account. It's the cash sitting in a brokerage account, waiting to be invested. SwitchWize's own Brokerage Cash Sweep Index, an ongoing, independently sourced dataset that predates this research, tracks what major brokerages actually pay by default on that idle cash, and it turns up a version of the Icelandic result inside a single US firm's own product lineup.

At Merrill, the default bank-sweep option for a brokerage account under $250,000 pays 0.01% as of the most recent snapshot. Merrill's own premium cash option, available in the same account, at the same balance tier, with no new application and no new institution required, pays 2.89%. That is not a marketing rate at a competitor. It is the same firm, offering close to three hundred times more yield for what amounts to one election inside an account a customer already has open. UBS shows a similar pattern: a default sweep as low as 0.03% for a brokerage account under $250,000, against a same-firm premium option paying 3.15% at the identical tier.

This is not as frictionless as a tap on a phone screen. Electing the higher-paying option typically requires an active step, in some cases an advisor conversation, rather than happening automatically the way Iceland's account-to-account transfers did. That difference matters and should not be minimized. But it is a meaningfully smaller lift than opening a new account at a new institution, which is the switch a typical bank comparison asks a saver to make, and the default still captures the overwhelming majority of that cash. A fuller pass on this data, including a proper pass-through analysis across the rate cycle, is planned as a separate piece of research; this is a preview using data that already exists, not the finished version of that work.

The honest complication

Two of the findings behind this research pull in different directions, and naming that tension is more useful than smoothing it over. Separate research on 154 US credit unions found that the largest, highest-balance depositors are the least sensitive to rate differences of any group, holding their balances as a medium-run store of liquidity rather than actively managing them for yield. The Icelandic data says the opposite about wealthy households: they are the most responsive group by a wide margin.

Both findings can be true at once because they are not describing the same thing. A single large balance sitting in one account, built up after a windfall and drawn down unevenly, is a different object than a wealthy household's whole liquid-asset position spread across multiple accounts and actively managed. Size of one balance and sophistication of an entire financial life are correlated but not identical, and this research doesn't resolve which one is doing the work when the two point in opposite directions. That is a real, open question, not a footnote to explain away.

Banks are not passive bystanders to any of this

None of this behavior sits outside a bank's awareness. Xu Lu and Lingxuan Wu, in research awarded the 2026 Ieke van den Burg Prize for work on systemic risk, measured depositor inattention directly across millions of US accounts by comparing how quickly people move money that arrives on a predictable schedule versus money that shows up unexpectedly. Banks serving more inattentive depositors set lower rates, pass through less of any Federal Reserve rate move, and see less money leave even as their own rates fall further behind competitors.

Read alongside the Icelandic finding, this closes a loop rather than opening a new one. It isn't only that most households fail to act on a rate gap even when friction and information are both removed. It's that the institutions holding their money have already measured how likely that inaction is, for depositors like them specifically, and priced accordingly. The gap this research keeps finding isn't a market failure nobody noticed. On one side of it, it's a forecast that keeps coming true.

What doesn't generalize from Iceland

Extending an Icelandic finding to US household behavior needs real caution, for reasons more specific than "results might differ in other countries."

Iceland's entire banking system collapsed in 2008, when its three largest banks failed within days of each other, an event with no equivalent in recent US banking history at that scale. Depositor behavior in a market that lived through that is not automatically representative of a market that didn't, and the direction of the effect isn't obvious either way: a banking collapse could make depositors more cautious and inert, or it could make them more actively distrustful of any single bank and thus more willing to move money around. This research doesn't settle which.

The dataset also covers one bank serving about a third of one country's population, which is a large sample by the standards of banking research but still a single institution's customer base, not a random sample of households across a fragmented market the size of the United States. And the accounts compared sit within that one bank, meaning the study tests whether people move money between products at the same institution, not whether they switch institutions entirely, which is the specific behavior a bank comparison site exists to encourage. Those are related questions. They are not the same question, and treating them as interchangeable would overstate what this evidence actually shows.

What this means for a comparison site

Between the UK finding and this one, the two most common explanations for deposit inertia, not knowing and not being able to be bothered, have each been tested directly and each came up short. What is left is belief: an assumption, formed before anyone looks at a single rate, that acting won't be worth it or comes with a catch. Neither a clearer number nor an easier switch reliably changes that assumption on its own.

That doesn't make a rate comparison, or the Bank Gap Index this site publishes, worthless. It makes clear that the comparison was never the whole job. The wealthiest households in the Icelandic data didn't need a friendlier interface to act; something else about how they already relate to their own money made the decision easy for them.

If understanding, not balance size, is what actually separated the households that moved from the ones that didn't, that points toward a different kind of fix than a cleaner rate table or a faster transfer button. It suggests spending less effort on making the switch itself easier, since the Icelandic case already had that solved, and more on the narrower job of replacing an untested assumption with a tested one before a number ever appears on screen. A friction-by-friction, time-stamped account of what an actual switch involves does that more directly than a bigger, bolder version of the same static comparison, because it answers the specific doubt rather than restating the opportunity. So does a pre-committed trigger, decided once in a calm moment rather than re-litigated every time a new rate shows up, since re-litigating the same pessimistic belief on repeat visits is exactly the pattern this research keeps finding. Both ideas are still being scoped rather than shipped, and neither is guaranteed to work; the honest position, consistent with everything above, is to test them against real switching behavior rather than assume either one closes a gap that has now resisted two different fixes in two different countries.

This article is educational and is not personalized financial, tax, or legal advice. Verify current rates, terms, and FDIC insurance status directly with any institution before acting. Full methodology and sources below.

Related reading: Why showing people a better rate doesn't make them switch, the Brokerage Cash Sweep Index, and the Bank Gap Index's academic foundation.