🇬🇧Licence Withdrawal and Wind-Down: What Happens to Client Money

Money sitting with an e-money institution, a payment institution, a money services business or a crypto platform is not a deposit. It is not insured, it carries no state guarantee, and legally it is the obligation of a private company backed by a segregation regime. While the licence holds, the difference between that account and a bank account is barely visible. It becomes visible on the day the regulator intervenes: a bank depositor is paid out by a guarantee scheme within weeks, while a non-bank operator's client enters a recovery process out of a segregated pool that runs for years and ends in a partial payment.

The scale is set out in the FCA's PS25/12: UK e-money institutions safeguarded roughly £26bn of client funds in 2024, up from £11bn in 2021, and payment institutions held about £6bn on any given day, while the share of consumers whose payment account sits outside a bank rose from 1% in 2017 to 12% in 2024. The same document carries the number that makes the mechanics worth understanding: for firms that became insolvent between Q1 2018 and Q2 2023, the average shortfall was 65% of the amount owed to clients.

Licence withdrawal, voluntary wind-down and insolvency are three distinct procedures with three distinct outcomes for the money. They are routinely conflated, including in operators' own announcements, even though which of the three has been triggered determines both how long recovery takes and how much of it there is.

Three exits the market calls by one name

Licence withdrawal is an administrative act. In the EU the grounds are exhaustively listed in Article 13 of PSD2: the institution has not used the authorisation for 12 months, has expressly renounced it or has ceased business for more than 6 months; obtained authorisation by false statements; no longer meets the conditions for grant; would threaten the stability of or trust in the payment system; or falls within a national ground. The authority must give reasons and publish the withdrawal in the register.

The point that gets lost in secondary coverage: withdrawal on its own does not make the company insolvent and does not start any distribution. It removes the right to provide services, nothing more. The reverse also holds — insolvency does not cancel the licence automatically. Ziglu entered special administration on 7 July 2025, and the FCA's own notice records that the firm continues to be FCA authorised; only the people in control have changed.

Voluntary wind-down is a solvent exit decided by the governing body while the money still covers everything. Under WDPG 3 of the FCA Handbook the wind-down period runs from the formal exit decision until the FCA cancels the firm's permission. This is the only one of the three routes in which the client normally recovers par: obligations are met in the ordinary course rather than through a collective procedure. The channel is larger than it looks — in 2025/26 the FCA cancelled the authorisation of 1,264 firms, largely through the "use it or lose it" powers introduced in September 2022 and aimed at firms not using their permissions.

Insolvency is the third route, and it is the one that produces shortfalls, queues and years of waiting.

The wind-down plan: a mandatory document almost nobody passes

A wind-down plan is an operational document, not a statement of intent. Under WDPG 3 it must set out the scenarios in which the firm becomes non-viable together with the governance and monitoring around them, the procedures for an orderly exit once the decision is taken, an assessment of financial and non-financial resources, and processes for identifying and removing obstacles. Non-viability arrives when the firm lacks adequate resources to conduct its regulated activities: material losses with no prospect of recovery, loss of key clients, failure of critical infrastructure. The financial side means daily cash-flow monitoring including the extraordinary costs of winding down — legal fees, redundancies, contract termination penalties. The non-financial side means premises, IT, key staff and external advisers remaining available throughout.

How well that is executed was answered by the FCA's multi-firm review of 26 June 2025. Across 2024/2025 the regulator examined a sample of 14 firms holding payments and e-money permissions and delivered an unhedged conclusion: none of the firms reviewed fully met its expectations. Plans were disconnected from the risk management framework, wind-down triggers were inconsistent with the risks the firms had themselves identified, liquidity was modelled from existing cash balances without stress testing, and — as the FCA notes separately — some firms did not understand the liquidity risks arising from their own safeguarding arrangements, including shortfalls.

In the EU the document becomes a condition of entry. The PSD3 proposal adds point (s) to the documents supporting an application for authorisation as a payment institution: a winding-up plan in case of failure, adapted to the applicant's envisaged size and business model. The recital specifies that it must support an orderly wind-up under applicable national law, including continuity or recovery of critical activities performed by outsourced service providers, agents or distributors. PSD2 contains no such requirement — this is new, and it is one of the few procedural changes to authorisation in the package. Timelines for the whole package are set out in the piece on PSD3 and the PSR.

For token issuers MiCA requires two documents, and the numbering is worth getting right because it is frequently misreported. Article 46 of Regulation (EU) 2023/1114 is the recovery plan: measures to restore compliance with the reserve-of-assets requirements, expressly including liquidity fees on redemptions, limits on the amount redeemable on any working day and suspension of redemptions. Article 47 is the redemption plan: an operational plan for the orderly redemption of outstanding tokens, implemented on a decision by the competent authority that the issuer is unable or likely to be unable to meet its obligations — including on insolvency, on resolution and, in terms, on withdrawal of authorisation. The redemption plan must provide for the designation of a temporary administrator and equitable treatment of all holders. Both are notified within six months of authorisation or white paper approval; the authority may require amendments within 40 working days and the issuer must implement them within a further 40. Article 55 extends both obligations to e-money token issuers, counting six months from the offer to the public or admission to trading.

How the pool is built and who pays for the administration

The pool is not built from what is on the safeguarding account. It is built from what should have been there. The administrator reconciles segregated balances against liabilities on the internal ledger, and the difference — the shortfall — is shared pro rata among clients. That is the first reason recovery falls below par: if the operator failed to top up, everybody takes the haircut.

The second reason is cost. The independent review of the UK regime published on 17 December 2025 states the allocation rule plainly: costs attributable to the first objective — returning client funds — are payable from the client funds estate, while costs of the second and third objectives fall to the general estate, with a top-up mechanism where appropriate. The practical effect, as the review puts it, is that poor records, cross-border elements and legal complexity make costs high relative to safeguarded balances, and that hits consumer outcomes. Allied Wallet gives the order of magnitude: the court allowed £1,560,000 of recoverable costs and disallowed £35,000 — meaning almost everything the liquidator spent was paid by the clients out of their own pool.

The third reason is time. In the cost benefit analysis to PS25/12 the FCA assumed the reform would cut the average time to return funds from 2.3 years to 1.3 years; respondents objected that the reduction was not evidenced and that 1.3 years is in any case too long to wait, and the FCA conceded in reply that disorderly insolvencies are inherently slow. The independent review offers a separate estimate: delays to first distributions are often at least 12 months. In Xpress Money Services distribution took at least fifteen months. In the Ipagoo administration, opened in 2019, the notice of final distribution to e-money holders, with a proving date of 12 September 2022, appeared only in the fourth year.

The UK: a bespoke regime and its own government's verdict

The Payment and Electronic Money Institution Insolvency Regulations 2021 created a dedicated procedure — special administration for payment and e-money institutions, known as PESAR. The administrator has three objectives: return relevant funds as soon as reasonably practicable, including prompt return of post-administration receipts; engage in a timely way with payment system operators, the PSR, the FCA, the Bank of England and the Treasury; and either rescue the institution as a going concern or wind it up in the best interests of creditors. The administrator sets the priority between them, though the FCA may direct which objective takes precedence where the public interest requires it. The regime supplies tools ordinary administration lacks: bar dates, including hard bar dates sanctioned by the court; a bar on key suppliers — including the bank holding the safeguarding account — terminating for pre-appointment arrears; and transfer of the client book to another institution, with notice to clients and agents within 14 days.

Two things about this regime rarely survive into summary coverage. First, it is not compulsory. The FCA states in PS25/12 that entry into PESAR is not mandatory for a payments firm, and for that reason has deferred a separate consultation on rules for firms that enter some other insolvency procedure. A client therefore cannot assume that a failing operator will end up with an administrator under a statutory duty to return client money first.

Second, the government's own assessment is unflattering. The review led by Adam Plainer, published by the Treasury on 17 December 2025, accepts that PESAR is preferable to relying on the Insolvency Act 1986 alone, then delivers the verdict: its procedural complexity, high costs and reliance on court processes have led to delays and diminished outcomes for customers, particularly those who are vulnerable or hold low-value accounts. Five reform priorities follow — a clearer hierarchy of objectives favouring rescue and customer transfer, an out-of-court entry route, better use of interim distributions, contingency planning for the failure of a large provider, and separately the protection gap itself: PESAR creates no FSCS cover, because EMIs and APIs sit outside depositor protection. That is a material distance from how the regime was described at launch, as a mechanism prioritising the return of customer funds.

The cases catalogued in the review map the problem: JNFX (November 2025), Argentex (July 2025), Ziglu (July 2025), Blackthorn Finance (April 2025), Nvayo (February 2025), Contis Financial Services (January 2025), LCC Trans-sending (June 2024), Silverbird Global (March 2024), Rational Foreign Exchange (November 2023), Monneo (May 2023) and Xpress Money Services (February 2022), in which the court approved the first distribution plan under the regime.

What the English courts actually held about client money

Two judgments define the position, and both are read backwards in most commentary. In Re Ipagoo the Court of Appeal on 9 March 2022 rejected the FCA's case and held that the Electronic Money Regulations 2011 create no statutory trust over safeguarded funds. The headline reads as a defeat for clients. The substance is the opposite: the court simultaneously held that "asset pool" in regulation 24 must be given a wider meaning and include a sum equal to all relevant funds that ought to have been safeguarded and were not, with that sum added from the general estate; and that regulation 24 overrides the priority rules that would otherwise apply on the EMI's insolvency. E-money holders therefore rank ahead of general creditors for the whole amount owed, not merely for whatever was actually segregated. The counterweight sits at paragraph 467 of the same judgment: because the pool is enlarged, "costs of distributing the asset pool" in regulation 24(2) must correspondingly be read to include the costs of making the pool good. Clients pay for their own top-up.

Re Allied Wallet completed the structure: the court confirmed that unsegregated assets top up the protected pools and allocated the deficiency across multiple pools on the principle that equality is equity. Both companies entered their procedures before PESAR came into force — Ipagoo in 2019, Allied Wallet in March 2020 — and were decided under general insolvency law together with the EMRs. Treating them as special administration precedents, as some commentary does, is simply wrong.

Premier FX stands separately, as an illustration of safeguarding not being applied at all. The FCA's public censure of 25 February 2021 records that none of the 73 accounts the firm maintained in the UK between 2013 and 2018 was designated as a safeguarding account, and that only three had terms excluding the bank's right of set-off. As at 14 December 2020 the liquidators had received claims from 136 creditors totalling £9,202,400.77. The FCA states directly that none of those losses is covered by the FSCS: protection does not extend to authorised payment institutions providing money remittance, still less to activity outside the permission.

Ziglu adds the crypto dimension. Per the analysis of the administrators' report, the shortfall on the Boost product was £2.2m, and as at 15 August the administrators had recovered £2,595,819 of Boost assets from Coinbase against £4,841,000 owed to Boost customers, with roughly 19,400 non-zero balances identified overall. The FCA's notice fixes the status of that money without ambiguity: the crypto activity was unregulated, Ziglu was under no obligation to safeguard funds relating to it, and there is no guarantee of any return. Boost holders became ordinary unsecured creditors. How the UK perimeter is being redrawn is covered in the piece on safeguarding and the new crypto regime.

The EU: no deposit guarantee, and half of one arriving in 2028

Start with what does not exist. The Bank Recovery and Resolution Directive applies, under Article 1, to "institutions" — credit institutions and investment firms — to financial holding companies and to branches of third-country institutions. Payment institutions and EMIs are absent from that list. There is therefore no bail-in, no bridge institution and no compulsory transfer of business: a failing EMI leaves its regulator with no resolution toolkit, only ordinary insolvency under national law. The nuance usually missed is that EMIs and PIs do fall within the CRR definition of a financial institution, so an EMI subsidiary inside a banking group that is consolidated for supervisory purposes can be pulled into BRRD scope through Article 1(1)(b). A standalone EMI never is.

Safeguarding itself sits in Article 10 of PSD2: funds are either kept unmixed with the funds of anyone other than payment service users and, if still held by the end of the business day following receipt, deposited in a separate account at a credit institution or invested in secure, liquid, low-risk assets; or covered by an insurance policy or comparable guarantee from a firm outside the group. The structural weakness is in the wording of the first limb: safeguarded funds are insulated against the claims of other creditors "in accordance with national law". The directive fixes the obligation but not the insolvency outcome, which stays national — which is why the result for a client of the same pan-European brand depends on where its licence sits. The account chain behind all of this is unpacked in the piece on correspondent banking and client-money protection.

Deposit guarantee cover never reaches money held at an EMI, because an EMI client is not a bank depositor. But the separate scenario — the bank holding the safeguarding account fails — was until recently answered differently in different Member States. The EBA's 2021 Opinion on the treatment of client funds under the DGSD recorded the spread: for payment institutions, pass-through cover applied in 16 Member States, did not apply in 10 and was unclear in one; for EMIs the split was 13 against 13, with one unclear. The claim that EU e-money is never covered by deposit insurance anywhere is accurate only for the failure of the EMI itself, and inaccurate for the failure of its bank.

From 2028 the spread disappears. Directive (EU) 2026/804 of 30 March 2026, published in the Official Journal on 20 April 2026, inserts a new Article 8b into the DGSD: Member States must ensure that "client funds deposits" are covered where the underlying clients are eligible for protection, the account is segregated in compliance with sectoral safeguarding requirements, and the clients are identified or identifiable by the financial institution before deposits are determined unavailable. The €100,000 coverage level applies to each underlying client, and by derogation the client's own deposits at the same bank are not aggregated into that calculation. The recitals name e-money institutions and payment institutions expressly. Transposition and application date: 11 May 2028. The protection still runs against the bank's failure, not the operator's.

The US: a statutory trust by state law, a token bond, and stablecoin holder priority

The American architecture rests on permissible investments rather than safeguarding. Under the CSBS Money Transmission Modernization Act a licensee must hold at all times permissible investments with a market value not less than the aggregate of all outstanding money transmission obligations. Section 10.03(c) is the strongest provision in any of the regimes surveyed here: permissible investments, even if commingled with the licensee's other assets, are held in trust for the benefit of the purchasers and holders of outstanding obligations on insolvency, on a bankruptcy petition, on a receivership petition, in any other judicial or administrative dissolution proceeding, or on action by a creditor who is not a beneficiary of the trust — and they are not subject to attachment, levy of execution or sequestration. In the US, commingling does not destroy the protection. In the UK, commingling is precisely what generated the litigation.

The surety bond is close to symbolic. Section 10.02(b) requires the greater of $100,000 or 100% of average daily money transmission liability in the state, capped at $500,000. For an operator with balances in the millions the bond covers a fraction of a percent and functions as a qualification barrier rather than a source of recovery. And the model is a model: per the CSBS legislative update of April 2026, some version has been enacted in roughly thirty states, with states adopting sections selectively — so the trust provision must be verified in the statute of the licensing state, not in the model text.

Even a trust does not prevent a hole. When Nevada's Financial Institutions Division petitioned for receivership over Prime Trust on 27 June 2023 it put fiat obligations to customers at $85.67m against $2.904m of reserves — a shortfall of roughly 97% — plus a crypto deficiency of about $861,000. The Prime Core Technologies entities filed Chapter 11 petitions on 14 August 2023; the plan was confirmed on 21 December 2023 and became effective on 5 January 2024. In the Synapse collapse, end-user obligations ran to roughly $265m against approximately $219m actually held at banks; on 22 December 2025 the CFPB allocated $46.2m from its Civil Penalty Fund, around half the projected shortfall. Pass-through FDIC insurance does not reach clients when the fintech fails rather than the bank.

For stablecoins the GENIUS Act, signed on 18 July 2025, builds three layers. Section 11(a) gives holders' claims priority — ratably among themselves — over the issuer and every other claimant with respect to required reserves. The amendment to § 541(b)(11) of the Bankruptcy Code carves required reserves out of property of the estate while keeping the automatic stay over them, and § 362(d)(5) directs the court to use best efforts to enter a final order beginning distributions no later than 14 days after the required hearing. A new § 507(e) gives the residual claim — whatever the reserves failed to cover — first priority over any other claim, including administrative expenses.

That is the text. Adam Levitin, writing on 2 December 2025, argues the opposite conclusion: holders in fact rank fifth, because §§ 507 and 726 distribute only unsecured claims while secured claims are paid first under § 725 — placing repo and margin claims, DIP lenders, estate professionals through carve-outs, and custodial banks with setoff rights ahead of them. The dispute is unresolved, and the statute itself supplies indirect support for taking it seriously: section 11(h) requires federal regulators to study "existing gaps in the bankruptcy laws and rules" for issuers and report to Congress within three years. Section 4(e)(1) meanwhile bars payment stablecoins from being backed by the full faith and credit of the United States or subject to FDIC or NCUA insurance. The regime is covered in the pieces on the GENIUS Act and on stablecoins generally.

Canada: an order without a hearing and a nineteen-million-dollar hole

Section 20(1) of the Retail Payment Activities Act offers a provider three options: hold end-user funds in trust in a trust account used for nothing else; hold them in a prescribed account in a prescribed manner; or hold them in an account used for nothing else and hold insurance or a guarantee for an amount at least equal to the balance. Section 20(3) bars the account-holding bank from asserting set-off against those funds. Ongoing compliance obligations took effect on 8 September 2025.

What the RPAA does not create is a special insolvency regime. The Bank of Canada draws the boundary itself in its January 2026 FAQ: its role does not include administering claims during or following a PSP's insolvency, nor notifying insurance or guarantee providers. Deposit insurance is described as insufficient, because it protects against the bank's failure rather than the provider's. The whole of the protection reduces to whether the structure the provider chose actually functions.

XTM showed how that plays out. On 17 February 2026 the Bank of Canada issued a temporary order under section 94(4) of the RPAA requiring XTM and affiliated entities to cease retail payment activities immediately and prohibiting transactions or withdrawals from the platform's accounts; on 27 February the order was amended to let the platform operate under court and monitor supervision. One widely repeated description needs correcting: a section 94(4) order does not "expire after 30 days unless extended". Section 94(5) provides the opposite — a temporary order continues in effect beyond the 30 days if no representations are made, or if, representations having been made, the Governor notifies the provider that he is not satisfied there are sufficient grounds for revoking it. As at the date of publication XTM is the only entity on the Bank of Canada's enforcement register, and both entries against it are temporary orders.

On 2 March 2026 XTM obtained CCAA creditor protection in Ontario, with The Fuller Landau Group as monitor. Sources give the trust shortfall as at 30 September 2025 differently: the company's own release puts it at approximately $18.75m, while trade coverage reports $18.96m, against $13.96m at 31 December 2024. Roughly 120,000 tip recipients and some 3,700 businesses were affected. The $200,000 discrepancy does not change the conclusion: the hole grew for a year, and the system engaged only once it was in double-digit millions. The regime itself is covered in the piece on the RPAA in Canada.

Crypto: where segregation exists and where it does not

Article 70 of MiCA requires crypto-asset service providers to make adequate arrangements to safeguard clients' ownership rights, "especially in the event of the crypto-asset service provider's insolvency", and to prevent the use of clients' crypto-assets and funds for their own account. Client funds other than e-money tokens must be placed with a credit institution or a central bank by the end of the business day following receipt, in an account separately identifiable from the provider's own. Paragraph 5 disapplies those duties for CASPs that are themselves e-money institutions, payment institutions or credit institutions, which are already inside their own sectoral regime. The weakness mirrors PSD2: the regulation harmonises the segregation duty but not the insolvency consequence, which remains national. For e-money token issuers, Article 54 additionally requires at least 30% of funds received to sit in separate accounts at credit institutions, with the remainder in secure, highly liquid instruments denominated in the currency the token references. The wider frame is in the pieces on MiCA and the CASP licence.

The UK has gone further and is closing the gap that Ziglu exposed. On 30 June 2026 the FCA published its cryptoasset regime policy statements — PS26/9 on admissions, disclosures and market abuse, PS26/10 on stablecoin issuance, PS26/11 on regulated cryptoasset activities, PS26/12 on the prudential regime and PS26/13 on Handbook application. Stablecoin backing assets and custodied client cryptoassets fall under a statutory trust, redemption is at par, and the regime becomes fully operational on 25 October 2027.

Reading the signals before a withdrawal

Regulatory pressure is visible publicly for around a year before the end. The places to look are registers and enforcement pages: the FCA publishes firm restrictions on the Financial Services Register, the Bank of Canada maintains a dedicated enforcement decisions register, and EU authorities must publish withdrawals under Article 13(3) of PSD2. Ziglu ran the standard sequence — on 23 May 2025 the FCA restricted certain products, on 17 June the firm agreed to cease payment and crypto activity, and on 7 July it entered special administration. Nvayo followed the same shape: FCA restrictions, a failed appeal to the Upper Tribunal, then an application for special administration.

Externally visible early markers: delays on outbound payments and "technical maintenance" on withdrawals; a change or loss of auditor; accounts filed late; loss of correspondent relationships, a subject covered in the piece on bank account closure; departure of the compliance head; and the launch of high-yield products, as Boost was at Ziglu, by an operator whose core economics do not close. Internally, the markers are the plan's own triggers: liquidity falling below the risk-appetite threshold, a shortfall on the daily reconciliation, an approaching expiry on the safeguarding insurance.

The practical conclusions are short. Keep operating balances at a non-bank operator, not savings. Diversify across several operators licensed in different jurisdictions rather than across several products of one operator. Check which specific service is regulated: at Ziglu the e-money was safeguarded and Boost was not, and that distinction was the difference between a pool share and unsecured creditor status. Keep evidence — statements, transfer confirmations, balance screenshots — because it will be needed to prove a claim before a bar date. Navigate by the map of licensing regimes and the FCA authorisation map rather than by marketing language about "bank-grade protection".

What the operator must do when it stops serving clients

From 7 May 2026 a UK operator must meet the supplementary safeguarding regime confirmed in PS25/12: reconcile at least once each business day; maintain a resolution pack — the documents and records that would allow a timely return of funds on insolvency; arrange an annual safeguarding audit where relevant funds exceed £100,000; submit a new monthly regulatory return; and notify the FCA without delay if internal records are materially out of date, inaccurate or invalid, if a reconciliation cannot be performed, if a discrepancy cannot be remedied, or if at any time in the previous year there was a material difference between the amount safeguarded and the amount that should have been. There is a separate rule on insurance: a contingency plan must be in place at least three months before a policy or comparable guarantee expires, and absent a replacement the firm must be ready to move to segregation.

The same policy statement contains a turn that dropped out of most 2025 commentary: the FCA decided not to implement the post-repeal regime — the statutory trust over relevant funds and receipt of funds directly into a designated safeguarding account. The regulator said it is not proposing to implement those proposals without further consideration and consultation, and will revisit once a full audit period under the supplementary regime has run. Statements that a statutory trust arrives in the UK in May 2026 are wrong: what arrives is records, reconciliation, audit and reporting. The trust is deferred.

The substance of ceasing to serve clients is the same across every regime surveyed here: advance notice to clients stating the cessation date and the withdrawal process; maintained withdrawal access until the last day; transfer of the client book to another licensed operator where feasible — in the UK regime a transfer carries notice to clients and agents within 14 days and novation of the agreements; notification of the regulator; notification of banks, insurers and guarantors — the banking stack behind a licensed operator usually breaks before the procedure closes; and retention of records sufficient to reconstruct every client's balance, since record quality is what determines both the duration and the cost of any subsequent procedure. How these duties sit inside the wider operating structure is covered in the piece on the compliance stack, and selling a licensed company instead of winding it down in the piece on change of control.

Common mistakes

Treating safeguarding as the equivalent of deposit insurance. Safeguarding is segregation, not a guarantee. It creates no source of payment for a shortfall and does not cover the cost of the procedure. The 65% average shortfall among failed UK firms follows directly from the fact that segregation depended on operator discipline.

Conflating licence withdrawal with insolvency. These are different events with different consequences. Ziglu remained FCA authorised while in special administration. Withdrawal from a solvent company usually ends in par recovery; insolvency of an authorised one does not.

Assuming the special regime as a given. PESAR is not compulsory, and the FCA says so plainly. Falling into an ordinary procedure means no statutory duty to return client funds first.

Believing a wind-down plan exists because a file exists. The FCA examined 14 firms and found none that fully met its expectations. Plans without liquidity stress testing and without triggers tied to the firm's real risk map do not work when the event arrives.

Holding savings where only an operating balance belongs. Yield products at a non-bank operator frequently sit outside the regulated perimeter. That is exactly what happened to Boost holders: no separate segregation, and unsecured creditor status.

Relying on a US surety bond. The $500,000 per-state cap in the model act covers a token fraction of a large operator's obligations. The source of recovery is permissible investments and the statutory trust, not the bond.

Not checking whose licence carries the account. A pan-European brand may operate through a subsidiary in another jurisdiction, and Article 10 of PSD2 refers insolvency protection back to national law — so the outcome depends on the licensing country rather than on the logo in the app.

Scenarios

Recovering money from an operator already in a procedure. Identify the appointed administrator or liquidator through the regulator's register and the official notice, submit a balance confirmation before the bar date, and preserve statements and transfer confirmations. Expect at least 12 months to a first distribution, and in some cases three years or more. Filing with the FSCS in the UK is pointless: e-money and payment services are outside the scheme.

Assessing an operator's risk before anything happens. Check the permission type and, separately, which products it actually covers. Read the regulator's enforcement register and the restrictions section of the firm's entry. Establish where client funds are placed and across how many banks. For EU operators, factor in that from 11 May 2028 pass-through DGS cover of €100,000 per underlying client becomes mandatory — but only against the bank's failure.

Winding down your own licensed entity. Take the solvent route rather than waiting for intervention: a governing-body decision, full performance of client obligations, notifications, transfer of the book, then an application to cancel the permission. The earlier the plan runs, the higher the chance of paying par and avoiding a procedure whose costs land on client money.

Launching a business that holds client money. Design the wind-down plan and resolution pack at application stage, not after authorisation: in the EU the winding-up plan is already part of the PSD3 authorisation dossier, token issuers file recovery and redemption plans within six months of authorisation, and in the UK the resolution pack is mandatory from 7 May 2026.

Holding stablecoins from a US issuer. Note that the GENIUS Act priority attaches to required reserves and is contested as to real-world ranking where secured claims exist. Establish who the issuer is — a depository institution, a subsidiary of one, or a nonbank — because that determines whether the matter lands with the FDIC, a state regulator or a bankruptcy court.