Nobody told him this was available to him John is a boat engineer. He designs the structures that keep vessels seaworthy — load, resistance, water pressure, the kind of work where “close enough” doesn’t hold up. We spent several weeks talking about his business before we talked about money at all: harbours, jurisdictions, a satellite connection that only ran in maritime mode. Somewhere in that conversation he went quiet, asked two precise questions, and then said something I haven’t stopped turning over since. “I didn’t know this was available to me.” He’d just realised that a professional life spent moving between countries and connectivity zones wasn’t only a source of friction. It was also, understood correctly, a source of financial options a business fixed to one office in one country doesn’t get. Nobody had hidden this from him. Nobody had ever framed his financial world in terms that connected to his actual life, so he’d spent years managing it the way you manage weather — absorbed, worked around, not thought about more than necessary. That’s a rational response to a bad offer. It’s also the most expensive decision most competent business owners make, and they make it without noticing. The UK keeps treating this as an information problem — a gap to be closed with a leaflet, a workshop, a “financial capability” module bolted onto onboarding. FutureDotNow puts the annual value of closing the UK workforce’s digital skills gap at £23 billion, a careful, well-sourced figure worth taking seriously on its own terms. Financial comprehension sits next to it, larger, and nobody has costed it at all. Parliament has just legislated a Financial Inclusion Strategy — November 2025 — and pencilled in primary-school financial literacy for 2028. That’s four more years of the current arrangement, for children who won’t be running businesses for another decade. It isn’t that owners lack information. Most could define APR without blinking. What they lack is permission to think of financial fluency as something that belongs to them rather than something they’re behind on — and the entire architecture of financial education, from the school savings passbook onward, has been built to communicate the opposite. (Somewhere in most banks’ compliance departments there’s a jargon glossary — “what is EBITDA,” that sort of thing — commissioned with good intentions and read, if the analytics are honest, by almost nobody. It’s the inclusion model in miniature: keep the language exactly as it was, offer a dictionary, call it generosity.) The gap has also changed shape underneath everyone. The Department for Science, Innovation and Technology’s research puts deliberate AI adoption among UK SMBs at around 16%. Ask a different question — how many are using AI at all, embedded in the accounting software, the payments platform, the CRM that quietly added a scoring feature eighteen months ago — and the figure is closer to 70%. Most owners running AI-assisted businesses don’t know they are. Nobody voted on the sequence by which their bookkeeping software started deciding which invoices to chase first. It just accreted, release note by release note, until it was load-bearing. The Financial Conduct Authority’s Consumer Duty, in force since July 2023, now obliges firms to support customer understanding rather than simply disclosing information and hoping it lands. That’s a genuine shift, and firms have spent two years building the compliance architecture around it. Whether it produces an owner who can read their own merchant statement is still an open question, and open questions administered thoroughly are not the same as urgency. None of this is a case against the technology. Open banking, live since January 2018, and algorithmic underwriting have done things a paper-based system never could. The problem sits one layer up, in who gets shown how any of it works, and on whose terms. Waiting for government to close that gap has an obvious cost, which is time nobody currently short of financial fluency has to spare. The businesses with the most direct stake in fixing it — and the most direct evidence of what its absence costs them — are the ones best placed to do it now, without a 2028 deadline. John messaged three weeks after our conversation. He’d found a tax adviser who specialised in maritime professionals. He’d restructured his invoicing — legally cleaner, operationally simpler than what he’d been doing for years. He’d found two AI tools that actually ran on maritime satellite, tools he’d never turned up before because he’d been searching the general market instead of his own. He said he’d started looking at his finances the way he looks at a boat: a structure with specific loads, which can be understood, and therefore steered. He supplied that metaphor himself, after the fact, better than I would have.
Money mules and the SMB: where laundered money meets your business
Money mules and the SMB: where laundered money meets your business Most founders treat fraud as something that happens to them: a scam invoice, a compromised card, a chargeback. Money laundering feels like someone else’s problem. It isn’t. Criminal proceeds have to leave the financial system somewhere, and a large share leaves through ordinary purchases from ordinary businesses. The cost is practical: an account frozen without explanation, a reserve imposed by your acquirer, a refund that quietly completes someone else’s laundering cycle. This lesson explains how a mule chain works, where an SMB sits inside it, and what to change in your own payment operations. How a mule chain moves money Laundering runs in three stages. Placement puts criminal money into the financial system. Layering moves it around to hide where it came from. Integration brings it back out looking legitimate. A money mule is a person or business whose account receives criminal funds and passes them on, knowingly or not. In a typical fraud, the victim’s payment lands in a first mule account, gets split and forwarded through further accounts, and is then cashed out. The FCA’s September 2026 multi-firm review puts numbers on the UK mechanics. Firms closed 238,396 suspected mule accounts in 2025. A public/private cell of 22 firms traced 140 fraud cases and found cash-out concentrated between the second and fifth account, by which point the money had been broken into smaller, less conspicuous payments. Card payments were the most common cash-out route: many low-value purchases, or larger payments to local businesses and retailers. Three terms carry the rest of the lesson. Cash-out is the point where criminal funds turn into goods, cash or crypto. Offboarding is a provider closing an account it suspects. De-risking is a provider withdrawing service from a type of customer it judges high-risk, often with little explanation, because the rules on disclosing suspicion reports restrict what it can tell you. Where your business sits An SMB touches a mule chain in three places. As a cash-out endpoint. When a mule spends criminal funds with you, the sale is real and the goods leave your premises. Your acquirer then sees your merchant flow as part of the pattern. High-value first purchases, payments split just under round numbers and refund requests to a different card are what monitoring systems flag. Refunds are the sharpest exposure: a refund to a new instrument turns spent money back into clean funds somewhere else, with your business as the laundering step. As an account holder. Business account closures for suspected muling were 10% higher in 2025 than in 2023, and challenger banks carried around half of them. Providers that onboard fastest attract mule accounts, and are quickest to close them. A legitimate business with an unusual month can get caught in the same net. Where it does not bite. Your rights as an APP fraud victim, your supplier payments and your borrowing are untouched by the mule findings. Diagnostic Answer yes or no. Do you sell goods that resell easily: electronics, gift cards, vouchers, luxury items? Can a customer get a refund to a card or account other than the one they paid with? Does more than one payroll cycle of your cash sit with a single provider? Could a cluster of high-value first-time orders go unnoticed for a week? Two or more yes answers mean your business is structurally attractive as a cash-out point, exposed to a freeze, or both. What to do this week Foundation: lock the refund route. In Stripe, Square, SumUp or whichever PSP you use, confirm refunds return only to the original payment method, and route every exception to one named approver. Applies to any business taking card or Open Banking payments. Layering: set one velocity rule. Using your PSP’s risk tools (Stripe Radar rules, for example) or a weekly export reconciled in Xero, flag first-time customers whose order exceeds three times your average ticket. The multiple is a starting heuristic, not a benchmark; tighten it if you answered yes to diagnostic question 1. Optionality: split the banking. Hold a secondary operating account with a different provider and keep one payroll cycle in it. Applies if you answered yes to diagnostic question 3. The test: could you pay staff and your key suppliers if the primary account froze tomorrow? Why this matters now Fraud controls are moving outward, from the victim’s bank to every account and merchant the money touches. The FCA now expects firms to look beyond the receiving account to payment characteristics and transaction context. The Economic Crime and Corporate Transparency Act 2023 lets firms share intelligence on suspected mules. The Home Office Fraud Strategy 2026–29 names mule networks directly. Each step widens the set of transactions a provider scrutinises, and SMB merchant flows sit inside that set. The 2028 financial literacy baseline assumes operators understand the system they trade inside, not just their own books. An operator who can explain their payment patterns, refund policy and banking concentration answers a provider’s question in a day. One who can’t learns what a frozen account costs. Tomorrow: how UK fraud reimbursement splits the cost between the sending and receiving firm, and what that means when your business is the one that got paid.
Which of your flows look like a cash-out
Which of your flows look like a cash-out Three checks on card acceptance, refunds and banking concentration, ordered by urgency. DAILY SIGNAL · PART 2 · FOUNDER REACTION If fraud proceeds exit through card spend at local businesses, the SMB question is which of your own flows look like a cash-out, and what your bank sees when they do. The signal, briefly This morning’s SIG-A read the FCA’s money mule review: 238,396 suspected mules offboarded in 2025, cash-out concentrated in the second to fifth account, and card payments the most common exit, often as larger payments to local businesses. The control perimeter is widening from the mule account to the merchant flow. SMBs sit inside it without choosing to. What this changes operationally Two categories move: payments and customers. On payments, card acceptance and business accounts at challengers and e-money providers now sit under heavier monitoring. Challenger banks held half of offboarded business accounts in 2025, and EMIs closed most suspected mule accounts within six months of opening. The practical risk is an unexplained account freeze or a sudden acquirer reserve. On customers, high-value card payments from first-time buyers, split transactions and requests to refund to a different card are the patterns worth a second look. Cash, credit, suppliers, hiring and contracts are unaffected. There is no new rule on SMBs, no change to reporting obligations, and nothing that alters your APP reimbursement position as a sender. Founder actions, ordered by urgency This week — ops lead, PSP dashboard (Stripe, Square, SumUp, Dojo). Pull the last 90 days of card payments above your normal ticket size from first-time customers. Decision criterion: any cluster of high-value first purchases, or payments split just under a round number, gets a named review before it becomes a pattern your acquirer finds first. By Friday — finance lead, refund policy. Confirm refunds go back only to the original card. If your PSP settings allow refunds to a different card, a bank transfer or store credit on request, switch that off or route it to founder sign-off. Refunding to a different instrument is a well-worn laundering exit. This month — founder, banking. If your primary business account sits with a challenger or e-money provider, open a secondary operating account elsewhere and hold at least one payroll cycle in it. Decision criterion: could you pay staff and your top three suppliers through a ten-working-day freeze on the primary account? What not to do Don’t bolt blanket friction onto every new customer’s checkout. It costs conversion, and the FCA finding is about patterns, not first purchases. And don’t move the whole business account to a high-street bank on the strength of one review: retail banks and building societies still accounted for 56.1% of suspected mules offboarded in 2025. The risk is concentration, not provider type. Tomorrow’s Friday regulatory anchor picks up the policy trail, including what the PSR’s December APP consultation is set to cover.
The till is the last visible point
The till is the last visible point SIG-A (Signal Analyst) The FCA’s money mule review traces fraud proceeds out of the banking system through ordinary card spend — and into SMB merchant flows. The FCA’s money mule review shows fraud proceeds leaving the banking system through ordinary card spend at local businesses, which makes the SMB till the last visible point in the chain. What happened On 23 September the FCA published a multi-firm review of money mule activity, built on a survey of 35 retail banks, building societies, challenger banks, PIs and EMIs, and a public/private cell of 22 firms that traced 140 cases across seven fraud types. Firms offboarded 238,396 suspected mules in 2025, up from 184,935 in 2023. Cash-out clustered between the second and fifth mule account. Card payments were the most common exit route. Why it’s the signal The headline number is about banks. The operationally important finding is where the money leaves. The FCA found card payments used either as many small transactions or as larger payments to local businesses and retailers — spend that reads as normal consumer behaviour. By the time cash-out starts, funds have already been broken into smaller, less conspicuous payments. Detection pressure therefore moves downstream, from the receiving account to the merchant flow. Two further data points sharpen it: challenger banks accounted for 50.1% of business accounts offboarded for suspected muling in 2025, and EMI offboarding rose 164.6% year on year, with 74.1% of EMI closures inside six months of account opening. What it tells us The story isn’t that banks are closing more mule accounts. It’s that the control perimeter is widening to include the businesses on the receiving end of the spend. Three things converge. The FCA expects firms to look beyond the initial receiving account, to payment characteristics and transaction context. It is issuing an alert through the National Economic Crime Centre with the cell’s detail. And the Economic Crime and Corporate Transparency Act 2023 gives banks, PIs and EMIs a statutory route to share intelligence on linked accounts. SMB business accounts and card acceptance sit inside that net — as potential mule accounts and as cash-out endpoints. Neither is a status an SMB chooses. Watch list NECC alert to firms. Watch for acquirer and EMI monitoring changes that follow the cell’s findings. PSR APP consultation. Changes to the reimbursement regime, scheduled for December 2026. FCA supervisory follow-up. Mule controls flagged for ongoing supervisory work in the review’s next steps. This afternoon’s SIG-R turns the cash-out finding into three checks an SMB can run on its own card and account flows before Friday.
Twenty years of financial education. The number hasn’t moved.
Twenty years of financial education. The number hasn’t moved. The Decimal Currency Board had a mascot. Dickie Decimal, a cartoon coin with a face, deployed across conversion charts and public information films to make a new currency feel familiar before it arrived on 15 February 1971. It didn’t work. And the reason it didn’t work is still running inside every financial education programme in Britain today. The preparation wasn’t lazy. Halsbury recommended decimalisation in 1963, the Act passed in 1967, and the Board spent years on a well-funded campaign. The arithmetic was never hard. What went wrong was the order of operations. The change was decided, then delivered, then explained — and only the last of those three involved anyone who actually had to use the thing. A woman who’d priced a Manchester market stall in shillings for thirty years got a pamphlet on a Monday morning and the assumption that comprehension follows exposure. Fifty-five years on, that sequence hasn’t changed. Ask anyone whose accounting software added an AI forecasting feature last year without telling them what it was doing. Who was this built for? Here’s the question the financial inclusion sector rarely asks. Not why people fail to understand money — who was financial education designed to serve? Not them. The Trustee Savings Bank ran savings schemes in British primary schools for most of the twentieth century. A child brought in a few pennies a week, a teacher wrote it in a passbook. What was being taught was habit rather than arithmetic: deposit regularly, don’t touch it. With a quieter lesson underneath about trusting whoever holds the book. Fine for a seven-year-old. It produced generations of dependable savers. It was also a curriculum built around what makes a good customer, not what makes someone able to judge whether the rate they’re offered is any good. That wasn’t a conspiracy. Banks existed to allocate capital efficiently, efficiency needed customers who behaved predictably, and predictability needed you to understand the product — not the system the product sat inside. The model survived nationalisation and the postwar credit boom. It survived the Big Bang of 1986, then the internet, without anyone revisiting the premise. When FinTech arrived talking about democratisation, it walked into an educational landscape nobody had redesigned, and found a fresh set of products not to explain. The register problem Financial education in this country is almost always offered as a correction. Targeted at people identified as lacking something. Described in the language of gaps and vulnerability. Delivered with an implicit message the curriculum never says out loud: you’re here because something went wrong. The targeting is usually accurate. The framing comes along with it anyway. Nobody aspires to remediation. A small thing that gives the game away: the phrase “financial capability.” I’ve never heard a business owner use it about themselves. It’s a term the sector uses about other people, which tells you most of what you need to know about who it was written for. What the sector published about itself In September 2025 the London Foundation for Banking & Finance published, with Bayes Business School, the most thorough independent review of UK financial literacy provision in a decade. Accelerating Progress: Financial Capability in the UK is an unusually candid document. Its central finding is that effective interventions from reputable providers aren’t shifting the numbers — individually or together — at anything close to the pace required. It records literacy gaps reaching 45% where demographic factors intersect, and recommends that tailoring provision to individual circumstances matters as much as scaling it. Two decades of investment. Competent programmes, professionally delivered, by organisations with no cynical motive whatsoever. And the Money and Pensions Service still finds 39% of UK adults — 20.3 million people — saying they don’t feel confident managing their own money. If the content were the problem, that number would have moved. What works, and why almost nobody copies it M-Pesa never ran a financial literacy programme. Safaricom launched it in Kenya in 2007 by starting from a narrower question than the education sector asks: what do people need to do with money, given the constraints they live under? Send it to each other, safely, without a bank account. The interface mirrored behaviour users already understood from informal transfers. The agent network — small shops and kiosks acting as human access points — supplied the translation layer. Formal financial inclusion in Kenya went from 26.7% in 2006 to 82.9% by 2019. Nobody completed a module first. The learning sat inside the thing being used. Nobody’s suggesting we import a Kenyan mobile money system into the West Midlands. The point is narrower: the design principle has been proven at national scale, and adopting it costs very little beyond a willingness to stop leading with deficit. If you’re running a business, this is your problem too The remedial frame isn’t only a policy failure. It’s the reason most competent owners have never seriously invested in their own financial fluency — they’ve only ever been offered it as catch-up, and turning that down is a rational response to a bad offer. Meanwhile the ground moves. DSIT research puts deliberate AI adoption among UK small businesses at roughly 16%. Ask instead how many are using AI at all — inside the accounting software, the payments platform, the CRM that added a scoring feature in a release note nobody read — and it’s nearer 70%. Most owners running AI-assisted businesses have no idea that’s what they’re doing. It’s the 1971 pattern again, at higher speed and with better graphics. One number for scale, and it’s worth being precise about what it covers: FutureDotNow, working from Lloyds’ Essential Digital Skills research, puts the annual value of closing the UK workforce’s digital skills gap at around £23 billion. That measures a specific thing — whether working adults can perform twenty tasks designated as essential. Financial comprehension overlaps with that and isn’t the same gap. Using a supplier portal is some distance from
Embedded Lending: How the Marketplace Model Works
Embedded Lending: How the Marketplace Model Works More UK business loans now arrive through a banking app than through a direct application to a lender. A challenger bank’s interface shows the offer, the account and the repayment schedule — but a separate company, plugged in through an API, is often the one underwriting the loan, funding it, and holding the credit risk. For a founder comparing options, that distinction rarely appears above the fold. It matters anyway: who you’re contractually bound to shapes what happens when a payment fails, a covenant is triggered, or terms need renegotiating. This lesson explains how the embedded lending model works, and what to check before you sign. The mechanism Embedded lending works through a straightforward architecture: a bank or platform with an existing customer relationship — a challenger bank, an accounting platform, a payments processor — integrates a specialist lender’s product directly into its own interface via an API. The customer sees a loan offer inside their banking app or software dashboard and applies without leaving that interface. But the entity actually assessing creditworthiness, extending the funds, and carrying the loan on its balance sheet is the specialist lender behind the integration — commonly a business-focused non-bank lender such as iwoca, rather than the bank or platform itself. This differs structurally from a bank originating and holding its own loan book. In a direct model, the bank underwrites, funds and owns the credit risk end to end. In the embedded model, the interface and the balance sheet are split across two companies, connected by an API rather than a single institution. The commercial logic runs both ways: the specialist lender gains distribution without building consumer-facing banking infrastructure, and the bank or platform offers a lending product without carrying the credit risk itself. For the customer, the visible experience — application, decision, repayment schedule — looks identical regardless of which model sits underneath it. What this means for you The distinction matters at three specific moments. First, at underwriting: the criteria used to approve or decline your application are the embedded lender’s, not the bank’s — a rejection inside a banking app reflects the lender’s model, not the bank’s. Second, at a problem: if a payment fails, you need a covenant waiver, or you want to restructure, the right call is to the lender managing the loan, not the bank’s general customer service line — misdirecting the call costs time you may not have. Third, at pricing: an embedded offer can carry a partner margin that a direct application to the same lender doesn’t, so the convenience of a single interface isn’t always the cheapest route. Where it doesn’t matter: your day-to-day banking relationship, your current account terms, and your existing facilities with other lenders are unaffected by any single embedded partnership. This is a distribution question, not a signal about the overall UK credit market or about whether now is a good time to borrow. Diagnostic Before your next application, run this check on any loan offer that appears inside a banking app, accounting platform, or payments dashboard rather than through a lender’s own site: Does the loan agreement name a lending entity different from the platform showing the offer? Is there a direct-application route to that same lender, and does it quote a different rate? Do you know which company’s customer service line to call if the repayment schedule needs to change? If you can’t answer question three with confidence, you don’t yet know who you actually borrowed from. What to do this week First — foundation. Pull the loan agreement for any active facility that arrived through a banking app, accounting platform, or payments dashboard rather than a direct lender application, and identify the actual contracting entity in the small print. Do this regardless of whether the loan is performing well; the point is knowing who to call, not fixing a problem. Second — layering. Update your facility or loan register — whether that’s a simple spreadsheet or a dedicated tool — to record the true lending entity against each facility, separate from the platform brand it appeared under. A five-minute fix now saves a frantic search later. Third — optionality. If you’re weighing a new embedded offer, request a direct quote from the underlying lender where one exists. Compare the two rates before accepting the embedded convenience; the difference, if any, tells you what the marketplace margin is costing you. Why this matters now Embedded finance is not a temporary trend — it’s becoming the default distribution model for SME credit in the UK, and the pattern will only deepen as more banks and platforms choose partnership over building lending capability in-house. The 2028 financial literacy implementation date assumes a baseline of operator competence that includes knowing who actually holds your credit relationships, not just which app you log into. Founders who build the habit now — checking the contracting entity, tracking it properly, comparing embedded pricing against direct quotes — will find it compounds as embedded products spread from lending into insurance, payments and working capital tools over the next two years. The founders caught out won’t be the ones who borrowed poorly; they’ll be the ones who never knew who they’d borrowed from.
Visa–Mastercard interchange settlement clears court — merchant fee structure changes ahead
Not long after news of the merchant card fee settlement began circulating, a client who runs a growing trade supply business got in touch with a question that caught my attention: “If merchants are being compensated for historical card fees, how do I know what I’m paying today is fair?” Despite processing significant annual card volume, he couldn’t explain the difference between interchange, scheme fees, and acquirer charges. His finance team could see the monthly bill. They couldn’t see the pricing decisions hidden inside it. A US federal judge has just handed small merchants exactly this kind of decision, at scale, for the first time. On 9 June, Judge Brian Cogan in Brooklyn gave preliminary approval to a settlement between Visa, Mastercard and the roughly 12 million merchants who sued them over interchange fees — a fight that has run for 21 years. Final approval is not expected before 2029, and an appeal is already being flagged by merchant lobbyists who think the deal still favours the networks. But the terms are significant enough to plan around now. The settlement scraps the “honour all cards” rule that has bound US merchants for decades. For the first time, a retailer can accept a customer’s standard consumer Visa while declining their premium travel-rewards card, or reject commercial cards altogether. Interchange on standard consumer credit is capped at 1.25 per cent for eight years; the broader effective rate drops by 10 basis points over five. Merchants also get wider latitude to surcharge. None of this is trivial — merchants paid a weighted average of 2.35 per cent on Visa and Mastercard transactions in 2024, according to the Nilson Report — but the interesting part isn’t the rate. It’s the decision-making the rate now requires. Here is the part the settlement’s architects seem to have half-noticed and then underfunded. Buried in the terms is a line for a “merchant education program about payment acceptance and cost management.” One line, no detail on scope, no indication of who delivers it or to whom. The card networks have effectively conceded that the people now empowered to categorise cards, price surcharges and steer customers toward cheaper payment methods have no particular reason to know how to do any of that. Then they wrote a single clause about it and moved on. I have watched this exact gap play out before, because the UK and EU already ran this experiment. Brussels capped interchange at 0.3 per cent for credit and 0.2 per cent for debit back in 2015, more than a decade before Washington got anywhere near it. If a hard cap were sufficient to close the capability gap, UK merchants would by now be sophisticated readers of their own payment costs. Most aren’t. Acquirer statements remain close to unreadable for a business owner without a finance background; blended pricing still obscures what any given card actually costs to accept; surcharging rules under the UK’s Payment Services Regulations get applied inconsistently or not at all, often out of simple uncertainty about what’s permitted. A rate cap changes what a merchant is owed. It does nothing to change whether they know they’re owed it, or how to collect. That gap sits in an unglamorous place — the layer between the card network and the merchant’s bank account. Merchants aren’t billed by Visa or Mastercard directly; a processor, a PayFac, or increasingly a software platform sits in between, and a headline rate reduction has to survive that intermediary’s own margin before it reaches anyone’s till. US merchants are about to discover what UK merchants already know: the acquirer or PayFac relationship is where savings quietly disappear, and almost nobody renegotiates that relationship because almost nobody has been taught how. This is where Vogue Boost’s founding premise comes from, not the other way round — we built a fintech upskilling company because the pattern kept repeating across every part of financial infrastructure I’d worked in. Regulators and networks fix the plumbing. Nobody teaches the people standing next to the taps how to use them. None of this argues against the settlement, appeal risk and all. Ending “honour all cards” is a genuine transfer of leverage from networks to merchants, and merchants should take it. But leverage unused is not leverage at all, and a merchant who doesn’t know she can now decline a loss-making commercial card is no better off than one operating under the old rule. The three-tier card categorisation this settlement introduces — commercial, premium consumer, standard consumer — is itself a decision layer most small business owners have never had to navigate, and there’s no natural constituency responsible for teaching them. So the question for policymakers on both sides of the Atlantic isn’t whether interchange caps or settlement terms are generous enough. It’s who is accountable for making sure the businesses affected can actually act on them. The FCA, the Payment Systems Regulator and their US counterparts have spent years negotiating the price of the plumbing. It’s time someone was made responsible for the people who have to use it.
Europe is about to let in the stablecoins nobody understands
Europe is about to let in the stablecoins nobody understands Tether has never held a MiCA licence. Europe’s largest crypto exchanges list it anyway, routing round the rule rather than through it, because European users keep asking for it and European law has no answer for what happens when they get it. That gap is what the European Commission opened a consultation to close on 20 May, and it is due to report by the end of August. The proposal under discussion would let non-EU stablecoin issuers operate inside the bloc, provided their home regimes are judged equivalent to Europe’s own. It is a sensible fix to a real problem. It is also the wrong problem to be fixing alone. What MiCA never asked The Markets in Crypto-Assets Regulation, when it came into force in 2024, built a reasonably coherent perimeter around issuers, reserves and redemption rights — for firms based in the EU. It said almost nothing about what happens when the reserves, the redemption promise and the legal recourse all sit in Delaware or the Cayman Islands, and the user sits in Frankfurt. The review now under way exists because that omission became untenable once the United States passed its own stablecoin framework, the GENIUS Act, and dollar-backed tokens started moving faster than European law could track them. So the Commission is doing the obvious thing: building an equivalence regime, so a Tether or a Circle can operate in Europe if their home jurisdiction meets EU standards. Patrick Hansen at Circle put the current situation plainly — European users are, in his words, either unprotected or cut off. Fair enough. But notice what the fix targets. It targets the issuer. It says nothing about the person deciding whether to hold, recommend, or process one of these instruments in the first place. The compliance officer who has to decide There is no functional way to explain to a client why a dollar-backed token redeemable in Delaware carries different risk from a euro-denominated one licensed in Paris. That is not a knowledge gap you close with a leaflet. It is a structural one, because the products changed faster than the training did. Multiply that by whatever number of newly admitted, differently domiciled stablecoins come through an equivalence regime once it’s live, and the problem doesn’t shrink. It compounds. A compliance officer at a mid-sized UK bank will soon have to assess counterparty risk across issuers governed by two, three, possibly more overlapping regulatory regimes, each with its own redemption mechanics and its own definition of what “backed” actually means. Nobody has built the training for that. Nobody is planning to. There’s a parallel worth a sentence, no more: this is roughly the same shape of problem HMRC created with Making Tax Digital, where the compliance machinery got upgraded and the humans required to operate it didn’t. Regulators are good at building rails. They are not, on the whole, in the business of teaching people to walk them. Equivalence is not understanding The Commission’s equivalence test — is a foreign regime “as good as” ours — answers an institutional question. It does not answer a human one. A stablecoin issued under an equivalent regime is not therefore a stablecoin your average IFA, or your average retail saver, has any better grasp of. Equivalence tells the regulator the plumbing is sound. It tells the person holding the asset nothing at all. ESMA’s own parallel review of custody and operational resilience, running from July through the first half of 2027, is a tacit admission of this. If the infrastructure needs eighteen months of scrutiny before anyone trusts it, the case for pausing on capability is at least as strong. Yet capability doesn’t appear on the consultation’s list of questions — I’ve read it twice to check. The reckoning None of this is an argument against opening MiCA to non-EU issuers. American money is not going to wait for Brussels to feel ready, and pretending otherwise just pushes European users back into the unlicensed grey market this review is meant to close. But a regulator that spends three years building an equivalence regime for issuers, and zero months building capability for the people who have to apply it, has solved half a problem and called it done. The Commission has until the end of August to decide what the next phase of MiCA looks like. It should ask, alongside every question about reserves and redemption, who in the market is actually equipped to price the risk it’s about to let in — and what it intends to do about the ones who aren’t.
FinTech — 29 July 2026 (UK + US)
HMRC digitised the tax system. It forgot to digitise the taxpayer.
HMRC digitised the tax system. It forgot to digitise the taxpayer. Making Tax Digital’s first deadline lands on 7 August. Most of the people it affects still don’t know it’s coming — and the lesson reaches well beyond tax. On 7 August, roughly 864,000 sole traders and landlords will find out whether Britain’s tax system was built with them in mind. That is the day the first quarterly update under Making Tax Digital for Income Tax falls due — HMRC’s most significant change to self-assessment in a generation, and the first hard deadline of a reform years in the making. On paper, the requirement is modest: a digital summary of income and expenses, filed through approved software, replacing part of the old annual paper trail. Research published by Sage in the weeks before the deadline found that only 37 per cent of those affected could correctly name the date. I recognise this shape. Thirty years spent investigating financial crime, running technology organisations through regulatory upheaval, and building companies at the point where fintech meets the people who actually have to use it, teaches you one lesson above the rest: systems fail not at the point of design, but at the point of use, among the people nobody consulted while building them. Making Tax Digital is not a software rollout gone slightly wrong. It is what happens when an institution digitises its own machinery and calls that the whole job. The evidence is not subtle. Research by IPSE and Sage last year found that only three in ten sole traders had a clear understanding of what MTD actually requires; a third were still keeping records on paper, two-thirds used spreadsheets, and more than half tracked income straight from bank statements. When Xero surveyed the same population as the April mandate took effect, 41 per cent said they were not ready, more than a quarter admitted they were behind schedule and unsure whether they would meet the August deadline, and one in seven had taken no action at all. Lloyds Banking Group’s research, drawn from sole traders already above the £50,000 qualifying threshold — the very group required to comply now — found that 55 per cent still had work to do. HMRC’s logic for the reform is sound as far as it goes. The department estimates that avoidable errors and simple failures to take reasonable care account for up to £9 billion of the tax gap each year — the shortfall between tax owed and tax actually collected — and it believes real-time digital records will close much of it. But the scope only grows from here. The qualifying income threshold falls to £30,000 in April 2027 and to £20,000 in April 2028, pulling roughly 900,000 more people into a regime that the current, better-resourced cohort is still struggling to meet. Each drop reaches further into thinner margins, less accounting support and lower digital confidence — precisely the population least equipped to absorb a compliance shift alone. This is the part policymakers consistently underfund. HMRC built approved-software standards, a penalty framework, and a soft-landing year in which no penalty points apply for late quarterly updates. What it did not build, at anything like the same scale, is a programme to raise the digital and financial capability of the people the reform depends on. Awareness campaigns are not capability. A leaflet explaining a deadline is not the same as teaching someone how to keep digital records, read what their software is telling them, or trust a number they did not calculate by hand. Britain treats financial and digital capability as a private responsibility, something individuals should already have or acquire in their own time, even as it builds public infrastructure that assumes they do. I see the same failure across every domain where financial infrastructure goes digital, from open banking to the early architecture of digital money: the technical build gets funded, tested and delivered on schedule, while the capability to use it safely is left to the market, or to chance. That gap is the subject of the book I am completing on how states are choosing, often by default rather than design, who controls the movement of money and who is left to keep up. Making Tax Digital is a small, domestic, unusually well-documented version of that same choice. Britain has decided, again, that digitisation and capability-building can proceed on separate timetables. They cannot. None of this argues for slowing MTD down. Quarterly digital records are a reasonable ask, and the tax gap they target is real money that other taxpayers effectively subsidise. But every threshold drop between now and 2028 will bring in people with less capacity to cope than the cohort filing this August, and government cannot keep treating that as someone else’s problem to solve after the fact. If HMRC and the Treasury are serious about a digital tax system, they need to fund capability with the same seriousness they fund compliance software — measured, resourced and owned, not left to accountants, banks and fintech firms to patch together after the deadlines have already landed. The 7 August deadline will pass quietly for most of Westminster. For most of the people it actually applies to, it will not.