Death by a Thousand Withdrawals: Why Banks Finally Took Fintech Seriously — and Why That Is the Easy Part
Survival and prosperity are not the same word.
Executive Summary
For most of the preceding decade, the established institutions of financial services regarded their upstart digital competitors much as a large animal regards an insect: aware of it, occasionally irritated by it, but never genuinely troubled. That posture is changing. Somewhere in the last two or three years the register in boardrooms has shifted from amusement to unease, and the word now circulating to describe the challengers — fintech — carries a suggestion that would have been unthinkable not long ago: that the incumbents might not win by default.
This essay is an attempt to understand that shift rather than to celebrate or dismiss it. It asks why the industry felt safe for so long, what actually changed to make the threat feel real, and what the threat is really a threat to. The argument is that the danger has been badly named. It is not that a payments application or a peer-to-peer lender will “beat the banks” — most will not. It is that the modern bank is a bundle of loosely related businesses held together by cross-subsidy and habit, and the newcomers have discovered they can pull the profitable threads out one at a time without ever having to carry the whole tangled garment.
That reframing matters because it explains the most striking pattern of the moment: the gap between how seriously institutions now say they take the threat and how little their response actually disturbs them. Recognising a disruptor is the easy part. The hard part is that a serious response asks an organisation to attack the very cross-subsidies that pay everyone’s salary — and no committee has ever voted to do that willingly.
The Long Comfort
It is worth remembering how reasonable the complacency was, because dismissing it as mere arrogance misses the point. The people who waved away the early challengers were not fools. They were reasoning correctly from a set of premises that had held true for a very long time.
The premises went roughly like this. Banking is a business of trust, and trust accrues over generations, not funding rounds. It is a business of scale, and scale in deposits and capital cannot be conjured by a clever interface. Above all it is a regulated business, and regulation is a moat that widens every year: the cost of holding capital, of compliance, of the sheer machinery required to satisfy a supervisor, rises with each new rule, and every pound of that cost is a barrier a newcomer must somehow clear before it takes a single customer. After the events of the last financial crisis, that regulatory burden grew heavier still. To a senior banker in that world, the idea that a start-up with a few dozen engineers and no balance sheet posed an existential question was not just implausible; it was a category error.
And for years the evidence agreed with them. The early challengers were, by any honest measure, small. A peer-to-peer lender might originate in a year what a mid-sized bank booked in a quiet afternoon. A payments start-up processed volumes that rounded to nothing against the card networks. The mobile-money schemes that were genuinely transforming lives were doing so in markets the incumbents had never bothered to serve. It was easy — and, on the numbers, defensible — to file all of this under interesting, not important.
The complacency was not stupidity. It was the correct answer to the wrong question. The incumbents kept asking whether the newcomers could replace them, when the newcomers had no intention of trying.
The trap in that reasoning is one the strategy literature had already named. An incumbent judges a new entrant by whether it can do what the incumbent does, only worse — and by that test the entrant always fails, right up until the moment it wins. The newcomers were not worse banks. They were not banks at all. They had simply picked one thing the bank did, and decided to do only that.
What Actually Changed
If the complacency was rational, then something in the premises must have shifted for the mood to turn. Several things did, roughly at once, and it is their coincidence rather than any single cause that made the last couple of years feel like a threshold.
The first was trust, the very asset the incumbents counted as their deepest moat. The financial crisis and its long aftermath did something to the relationship between institutions and their customers that no marketing budget could repair. For the first time, “I don’t work the way the banks do” became a selling point rather than a warning. A generation formed its financial habits during the years when the banks were synonymous with bailouts and scandal, and that generation did not extend the benefit of the doubt.
The second was the phone. The smartphone did not merely give the challengers a channel; it dissolved the branch network’s meaning as a competitive advantage. For a century, distribution in retail finance meant physical presence, and physical presence meant capital that no start-up could match. Once the primary place a person met their money was a glass rectangle in their pocket, that entire moat drained overnight. The incumbents still owned the branches. It was no longer clear the branches were worth owning.
The third was the collapse in the cost of building things. The infrastructure that once required a data centre and a multi-year procurement could now be rented by the hour. A team could stand up a service, put it in front of customers, and scale it — or kill it — without ever making the kind of capital commitment that used to define entry into the industry. What had been a moat of capital expenditure became, quietly, a monthly bill.
- Trust stopped being a one-way advantage: the incumbents’ oldest asset became, for a cohort of customers, a liability.
- Distribution detached from physical presence, stranding the branch network’s value.
- The cost of entry fell from a capital commitment to an operating expense, opening the field to anyone with a good idea and a small team.
- Talent and capital began flowing the other way — the ambitious engineer and the growth investor now found the challenger more interesting than the institution.
None of these on its own would have been decisive. Trust erodes slowly; phones were already common; cheap infrastructure had been arriving for some time. But together they removed, more or less simultaneously, three of the four pillars on which the long comfort rested. Only the regulatory moat remained fully intact — and even there, supervisors were beginning to talk openly about competition as a goal rather than a risk, and about lowering the barriers to new entrants rather than raising them.
The Threat, Correctly Named
Here is where the popular framing goes wrong, and where the more interesting story begins. The language of the moment is the language of replacement and death — disruption, existential threat, the banks as dinosaurs. It is thrilling, and it is mostly incorrect. Very few of the challengers want to become banks, and fewer still will succeed at it. Most will fail, and a good number of the survivors will end up absorbed by the very institutions they set out to unsettle.
To stop the analysis there, though, is to make the incumbents’ original mistake in reverse. The threat is real; it has simply been misdescribed. To see it properly, you have to look at what a bank actually is.
A modern universal bank is not one business. It is a dozen businesses wearing one coat. It moves money, holds deposits, lends, exchanges currencies, advises, underwrites, safeguards, and settles — and it offers all of this to a customer as a single relationship, priced as a bundle. Some parts of that bundle are enormously profitable. Others are loss-leaders, tolerated because they hold the relationship together so the profitable parts can do their work. Cross-subsidy is not a flaw in the model; for a long time it was the model. The dull, expensive current account exists so that the lucrative overdraft, the foreign-exchange margin, and the lending relationship have somewhere to live.
The newcomers understood something the incumbents had stopped noticing: a customer experiences the bundle as a bundle, but its economics are anything but uniform. And so, one by one, the challengers went after the profitable, defensible-seeming threads — the fat margin on sending money abroad, the spread on a personal loan, the fee on a payment — and offered that single thread, done well and priced honestly, to a customer who had never been shown the seams.
| What the challengers take | What the incumbents keep |
|---|---|
| The visible, profitable slivers — payments, transfers, foreign exchange, unsecured lending, simple advice | The invisible, expensive foundation — the licence, the balance sheet, deposit insurance, settlement, the regulated core |
| The parts customers can compare and switch on a phone | The parts customers never see and rarely think about |
| The revenue that funds the cross-subsidy | The obligations the cross-subsidy was funding |
Seen this way, the danger is not death by a single blow. It is death by a thousand withdrawals. Each thread pulled out is survivable on its own. The trouble is that the threads the newcomers want are precisely the profitable ones, and the parts they leave behind — the regulated plumbing, the capital-hungry obligations, the loss-leaders — are exactly the parts that only made sense when the profits were there to subsidise them. An institution can lose no single business and still find, thread by thread, that the coat has become a pile of loose string.
“The newcomers are not trying to beat the bank at being a bank. They are quietly disassembling the bundle, and leaving the incumbent holding the parts that never paid their own way.”
The Honest Counter-Argument
A serious essay owes its subject the strongest version of the opposing case, and there is a strong one. It runs like this: the incumbents’ moats are real, and the unbundling story flatters the challengers.
The regulated core is not a nuisance to be routed around; it is the foundation everything else stands on. Deposit insurance, access to the settlement system, the licence itself — these are not features the newcomers can casually replicate, and many of the challengers, on inspection, turn out to be riding on an incumbent’s rails, an incumbent’s licence, or an incumbent’s balance sheet without always saying so. Trust, for all its post-crisis battering, still runs deep when a person is deciding where to keep the money they cannot afford to lose; novelty is charming for a payment and terrifying for a life’s savings. Scale still tells: the cost of acquiring a customer for a single-thread product is punishing, and a great many of the challengers are, beneath the elegant interface, businesses that have not yet shown they can make money. Incumbents, meanwhile, can copy a good interface far faster than a start-up can acquire ten million trusting customers or a banking licence.
All of this is true, and it is why the language of extinction is overblown. But notice what the counter-argument concedes even as it reassures. It admits that the profitable slivers are contestable. It admits that the incumbent’s enduring advantages cluster in the parts of the business that cost money rather than make it. A moat around the plumbing is real protection only if the plumbing is where you earn your living — and for the modern bank, it is not. The counter-argument does not refute the unbundling thesis. It refines it: the incumbents will very likely survive, but they may survive into a shape they would not have chosen, keeping the obligations and ceding the margins. Survival and prosperity are not the same word.
Why Seriousness Is Not Enough
Which brings us to the most revealing pattern of all, and the one that should trouble the institutions most: they now take the threat seriously, and it is not helping.
The evidence of seriousness is everywhere. There are innovation labs and accelerators, hackathons and partnerships, a newly minted digital officer in the executive photograph. Sums are being committed that would have been unthinkable when the challengers were still filed under interesting, not important. And yet, watch what these responses actually touch, and a consistent shape emerges. They cluster at the edge of the organisation, carefully insulated from its core. The lab experiments; the core continues exactly as before.
The reason is structural, not a failure of nerve or intelligence, and it follows directly from the unbundling diagnosis. If the newcomers are attacking your profitable threads, a genuine response means competing with them — which means offering your own customers the honestly priced version of a product you currently sell at a fat margin. It means, in plain terms, cannibalising your own most profitable lines before someone else does it for you. And that is a decision no individual executive is structured to make.
- The head of the profitable line is measured on this year’s margin, and is asked to volunteer that margin for a benefit that accrues, if at all, to someone else in some future year.
- The cross-subsidy means the loss-leaders quietly depend on those margins, so cutting them exposes costs the whole institution would rather not confront.
- The innovation unit, deliberately kept at arm’s length so it can move freely, has no authority over the core lines it would need to change — and so it produces demonstrations rather than decisions.
The result is a kind of theatre that is entirely sincere. Everyone involved believes they are responding to the threat. The accelerator is real, the partnerships are signed, the digital officer works genuinely hard. But the organisation has arranged its response so that nothing it truly depends on has to change, and it has done so not through cynicism but through the ordinary physics of incentives. The antibodies of the core business attack any initiative that threatens a current margin, and they do so automatically, because that is precisely what they were built to do.
This is the true content of the phrase the gap between transformation intent and transformation reality. It is not that institutions do not mean it. It is that meaning it, at the level of the executive committee and the annual report, is fully compatible with changing nothing that matters at the level of the profitable line. Intent lives in the language; reality lives in the cross-subsidy; and the two can coexist indefinitely.
What a Serious Response Would Require
It would be dishonest to end an essay of this kind with a confident recipe, and the honesty of not-knowing is part of the subject. No one at this moment can say which challengers will endure, which incumbents will adapt, or what the industry will look like once the reshuffling settles. What can be said is what a serious response would have to be willing to do, whether or not any given institution proves willing to do it.
It would have to name the cross-subsidy out loud. An institution that cannot say, internally and plainly, which of its products lose money and which of them pay for the losses cannot reason about which threads it can afford to lose. Most cannot say this with any precision, and the vagueness is not accidental; the bundle is more comfortable when no one looks too closely at its seams.
It would have to give the response real authority over the core, not a comfortable distance from it. An innovation unit that cannot change the terms of a profitable product is a research department, not a competitive response. The uncomfortable truth is that a serious effort has to be allowed to hurt the core on purpose — and that means it has to be sponsored by someone with the standing to overrule the head of the line it is disrupting.
And it would have to accept, at the most senior level, that the choice is not between changing and staying the same. It is between disrupting your own margins on your own terms and having them disrupted for you on someone else’s. The incumbents who come through this well will not be the ones who took the threat most seriously in the language of strategy. They will be the ones who were willing to lose the profitable thread themselves, deliberately, before it was taken — and that is a species of courage that org charts are not designed to produce.
The challengers, for all the noise around them, have done the industry an accidental service. By pulling at the threads, they have made visible a structure the incumbents had stopped seeing: that the bank was never one business, that its profits and its obligations had drifted apart, and that a bundle held together by habit is only as durable as the habit. Whether that lesson is learned in time is the open question. What is no longer in doubt is that the toy in the corner was never a toy. It was a mirror, and the institutions have finally begun to look into it.