The Law of Controlled Amnesia
Cash, code, and the limits of bearer finality — when money can be followed, reclaimed, and forfeited, and why “code is law” was never true
Cash is built to forget. A banknote carries no memory of who held it last, no record of how it was earned, no ledger of the hands it passed through. That amnesia is not a defect. It is the point. A currency that remembered everything would not circulate; it would litigate. Every shopkeeper would become a title investigator, every transaction a due-diligence exercise, every tenner a small lawsuit waiting to happen. Money works precisely because it is allowed to travel without a biography.
And yet the law sometimes needs the biography back. It needs to know who owned the notes, who took them, who paid them away, whether they were stolen, mistaken, mixed, or derived from crime, and whether the person now holding them knew any of that. The whole legal architecture of money is a negotiation between these two demands: the demand that cash forget enough of its past to circulate, and the demand that it not forget so much that theft, fraud, and criminal proceeds become legally purified by movement.
This essay is about that negotiation, and about an argument that the negotiation does not stop at the edge of the physical world. The fashionable claim of the last fifteen years — that distributed ledgers create a new kind of money beyond the reach of courts, that “code is law” — is the old cash problem dressed in new clothes. It fails for the same reason the strong version of cash-finality fails. Bearer finality has never been absolute. It is a legal tolerance, extended because ordinary exchange needs it, and that tolerance shrinks as value, opacity, and systemic risk grow. Blockchains do not escape this. They relocate it.
Three claims organise what follows. First, that cash is not a thing but an institution: a legally tolerated degree of bearer finality, not an unlimited right to move value without accountability. Second, that this tolerance is scalar — a five-dollar coffee and a ten-million-dollar suitcase are both “cash” in the trivial sense that the notes remain currency, but they are not the same institution, and the law does not treat them as such. Third, that digital systems inherit this structure wholesale: a smart contract can move a token, but it cannot decide whether the transfer was authorised, lawful, voidable, fraudulent, or subject to a court order. Code may settle the ledger. Law settles the rights.
What cash is, and what it is not
Begin with the ordinary categories: notes and coins, denominated in a sovereign unit of account. Most people think they know what “legal tender” means, and most people are wrong. In England and Wales, Bank of England notes and Royal Mint coins are legal tender, but legal tender does not oblige a shop to accept them. As the Bank of England’s own public guidance explains, the concept is much narrower than the folk understanding: it bears on the discharge of debts, not on whether a retailer must take your cash for a sandwich. A creditor who is properly tendered legal tender in satisfaction of a debt, absent a contrary contractual term, cannot then succeed in a claim for non-payment. That is what legal tender does. It does not make cash compulsory; it makes a proper tender a defence.
Four concepts get collapsed here that ought to be kept apart: money, cash, legal tender, and payment method. A debit card, a cheque, a contactless tap — these are ordinary payment methods, and none of them is legal tender. The distinction is not pedantry. It matters enormously for recovery, because the route by which value can be reclaimed depends entirely on what kind of thing the claimant is chasing: physical notes, a credit balance in a bank account, an electronic payment instruction, a crypto-token, or a contractual obligation to pay. Each travels through a different part of the law.
So what is cash, legally? It is a bearer asset with currency status — property in the hands of the holder, but property that also performs a public monetary function, and whose ordinary property rules are modified precisely because of that function. A watch or a painting carries its title history with it: a thief cannot generally pass good title to stolen goods, and the true owner can usually recover them or sue in conversion. Cash is different. It is meant to pass from hand to hand without inquiry into where it has been.
The pivot case is more than two and a half centuries old. In Miller v Race (1758), a stolen bank note made its way, through an honest chain, into the hands of someone who had given value for it without notice of the theft. Lord Mansfield held that the note was to be treated as currency, and that an action in the nature of conversion would not lie, because there is no property in currency once it has passed. His formulation has never been bettered: where money is stolen, the true owner cannot recover it once it has been “paid away fairly and honestly upon a valuable and bona fide consideration”; but before money has passed in currency, an action may be brought for the money itself. Commerce, Mansfield understood, could not function if every recipient of a banknote had to reconstruct its ownership history. The law therefore made a deliberate sacrifice: it gave up the owner’s claim against the honest recipient in order to keep money moving.
That sacrifice is the load-bearing idea. Cash is not merely a thing. It is a thing whose legal usefulness depends on being treated as final in ordinary circulation.
Cash as tolerated bearer finality
Here is where the argument turns. Miller v Race is usually read as a statement about what cash is. It is better read as a statement about what the law is willing to tolerate. Finality is not a natural property of banknotes; it is a legal concession, granted because the alternative — perpetual recoverability — would destroy the very liquidity that makes cash worth having.
And concessions have conditions. The tolerance is at its strongest in ordinary retail and private exchange, where the social benefit of frictionless payment plainly exceeds the risk of concealed wrongdoing. It weakens — sharply — where the amount is abnormal, the context is suspicious, the transaction is structured to dodge scrutiny, the recipient is a regulated institution, the money is held rather than spent, or the value is linked to crime. The notes do not change. The institutional treatment does.
This is the home of the distinction that runs through the rest of the essay. A five-hundred-dollar cash payment for goods may be cash in the full economic and legal sense: ordinary, final, private, and practically uninvestigated. A ten-million-dollar transfer in physical banknotes is not, in any meaningful sense, “just cash.” It is a concentrated bearer-value event. It raises questions of source, purpose, beneficial ownership, tax, sanctions, laundering, coercion, and public order. The banknotes remain currency. The transaction no longer enjoys the governance tolerance that ordinary cash circulation receives.
Put the principle as a slogan: a small cash payment is presumptively transactional; a very large cash transfer is presumptively evidential. The first is something the law lets pass without asking questions. The second is something the law treats, by default, as a matter requiring explanation.
Figure 1. “Cashness” is scalar. The tolerance the law extends to bearer finality decays as transaction value and opacity rise, while the burden of verification and governance rises to meet it. The banknotes remain legal currency at every point on the curve; what changes is the institutional treatment of the transaction. (Schematic — illustrative of the argument, not an empirical measurement.)
This scalar view is not a moral claim that large sums are wicked. It is an institutional claim about where the law’s tolerance is rationally placed, and the next section explains why the placement makes economic sense.
Small cash solves transaction costs; large opaque cash creates verification costs
Standard economics gives money three functions: a medium of exchange, a store of value, and a connection to the unit of account. Cash performs all three, but it also does things that bank money and card payments do not fully replicate: it settles immediately and with finality, it works offline, it requires almost no operational infrastructure at the point of payment, it transfers by simple delivery, it preserves a measure of privacy, and it keeps working when the power and the networks fail. For small, immediate, ordinary exchange, these are not trivial advantages.
The institutional-economics tradition explains why an arrangement like cash exists at all. Ronald Coase’s insight was that the structure of economic institutions is shaped by transaction costs — the costs of search, negotiation, verification, and enforcement — and that arrangements emerge to economise on them. Douglass North built the broader account: institutions are the “rules of the game” that reduce uncertainty and make exchange possible across distance and anonymity. Oliver Williamson’s work on governance structures pressed the same logic into the question of how transactions are organised to manage the hazards specific to them. Cash, read through this lens, is a transaction-cost technology. It lets low-value payments clear without anyone bearing the cost of investigating provenance — a cost that, for a tenner, would dwarf the value at stake.
But the same features that make cash cheap at small scale make it expensive, and dangerous, at large scale. The advantages do not merely fade as value rises; they invert. Counting, storing, securing, transporting, insuring, and verifying large sums of physical currency is costly and risky. And the privacy that is a benign convenience at the till becomes, in bulk, opacity — the central concern of the systems built to police tax, sanctions, insolvency, and money laundering.
So the economic question driving the whole field is not whether cash is good or bad. It is a question of calibration: how much finality does cash need to function, and how much recoverability does the law need so that it does not become, in effect, a laundering subsidy? Recoverability protects victims and public order. Excessive recoverability destroys liquidity by turning every recipient into a title investigator — exactly the outcome Miller v Race was designed to avoid. The trade-off is real and it is two-sided, and the law’s answer is the scalar tolerance of Figure 1.
State the inversion plainly, because it is the engine of the argument: small cash solves a transaction-cost problem; large anonymous cash creates a verification-cost problem. The first is something society wants to be cheap. The second is something society needs to be legible.
The limits of cash, and why value changes legal character
It follows that the glib claim “cash is cash at every amount” is false in the only sense that matters. Legally, the same banknotes remain currency whatever the sum. Institutionally, the treatment is not the same. There are three limits at which the cashness of a transaction gives way.
The first is transactional. Cash is at its best for low-value, immediate, ordinary exchange. Its usefulness declines with size because the physical burdens of handling it grow faster than the value it carries.
The second is evidential. Cash is hard to trace. For small values that is tolerable, because the administrative cost of tracing would exceed any benefit. For very large values, the difficulty of tracing is not a side effect — it is the precise thing that compliance regimes exist to counteract.
The third is governance. Large anonymous bearer transfers cut against tax systems, sanctions enforcement, insolvency rules, fiduciary duties, court orders, and anti-money-laundering law all at once. At sufficient scale, a bearer transfer is not merely an exchange between two parties; it is an event with consequences for everyone whose interests those regimes protect.
The institutional world already encodes this. Value-keyed rules are everywhere: the United States has long required currency transaction reports above a fixed threshold, and the European Union’s 2024 anti-money-laundering reforms introduce an EU-wide ceiling on large cash payments. These are not claims that large sums are criminal. They are recognitions that large opaque transfers impose systemic risk, and that the law’s response should scale with the risk rather than with the physical form of the instrument. A five-dollar note buying coffee is ordinary cash. A fifty-thousand-dollar payment may require explanation depending on context. A ten-million-dollar transfer — in notes, tokens, stablecoins, or bearer instruments — cannot be treated as ordinary cash circulation, because at that level the transaction implicates governance, not merely exchange.
When cash can be reclaimed: the private-law hierarchy
So when does cash cease to be merely money in circulation and become recoverable property? The private law of recovery is best understood not as a single rule but as a ladder, organised by cause of action, with the claimant’s prospects strongest at the top and weakest at the bottom.
Proprietary recovery sits at the top in principle. If identifiable notes or coins are taken and remain identifiable, the owner may claim the thing itself. But physical cash loses identifiability almost immediately: it is spent, substituted, deposited, mixed, and exchanged, and once it has, the specific-thing claim evaporates.
Following and tracing are the law’s response to that evaporation. To follow is to pursue the same asset into another’s hands; to trace is to identify the value of the original asset as it is exchanged into substitutes. The difficulty is fungibility. Once cash is mixed in a pile, deposited into a bank account, or swapped for another asset, the claim shifts from physical identification to value identification — a far harder exercise. Banque Belge pour l’Etranger v Hambrouck (1921) is the early landmark, allowing a claimant to trace misappropriated money through bank accounts where its identity could still be established. Tracing through accounts, and through mixtures, has been the battleground of this branch of the law ever since.
Unjust enrichment and the action for money had and received form the doctrinal centre, and the decisive modern authority is Lipkin Gorman v Karpnale Ltd (1991). A partner in a firm of solicitors, a compulsive gambler, stole large sums from the firm’s client account and lost most of it at the Playboy Club’s tables. The firm could not recover from the thief, who was insolvent, so it sued the club. The House of Lords held that the club had been enriched at the firm’s expense by the receipt of stolen money, and was liable to make restitution — even though the club had received the money innocently and in good faith. Crucially, the same decision recognised the defence of change of position: the club’s liability was reduced by the winnings it had paid out, because it had, in good faith, changed its position in reliance on the receipt. Lipkin Gorman did two things at once: it confirmed that innocence does not automatically bar a restitutionary claim, and it gave innocent recipients a defence calibrated to what they had actually, detrimentally, done.
Personal claims against the wrongdoer sit alongside all of this and are easily forgotten. The thief or fraudulent payer remains personally liable even where the cash itself cannot be recovered from an innocent third party. The loss of a proprietary route is not the loss of all remedy; it is the loss of one remedy against one person.
Claims against recipients with notice or fault mark the point where the analysis leaves the world of innocent circulation altogether. Where a recipient had knowledge, notice, suspicion, dishonesty, or breached anti-money-laundering obligations, the case moves toward knowing receipt, dishonest assistance, money-laundering liability, or statutory recovery. These are not defences to circulation; they are reasons the protection of circulation never applied.
Figure 3. The recovery hierarchy. Recovery is strongest against the thief and the volunteer, weaker against a recipient with notice, and weakest of all against the bona fide purchaser who gave value without notice — against whom the money “passes in currency” and the claim is defeated. The dividing line is not innocence alone. The strong protection requires innocence plus value plus absence of notice plus ordinary circulation. Bar lengths are schematic.
“Without knowledge”: innocent receipt, and the real shape of the defence
The hardest cases are framed by a single deceptively simple phrase — the recipient who took the money “without knowledge.” The phrase hides at least four different questions, and collapsing them produces most of the confusion in this area.
In private law, lack of knowledge does not, by itself, defeat restitution. As Lipkin Gorman shows, a recipient can be wholly innocent and still be enriched at the claimant’s expense; innocence matters, but it matters at the level of defences — above all, change of position — rather than as a complete answer to the claim.
In property law, the innocent purchaser for value is treated very differently from the volunteer. The volunteer — the donee who gave nothing — is exposed, because there is no countervailing reason to protect a recipient who has parted with no value. The purchaser is protected, because commerce depends on the ability to accept money without reconstructing its past. Value is doing real work here: it is the thing that converts mere innocence into a defence.
In criminal law and anti-money-laundering, the relevant mental states are knowledge, suspicion, or reasonable grounds for suspicion. A person who genuinely does not know and has no reason to suspect that cash is criminal property is in a fundamentally different position from one who knows or suspects and proceeds anyway.
In civil recovery under the Proceeds of Crime Act 2002, the statutory language is decisive, and it crystallises the whole analysis. Property obtained through unlawful conduct is “recoverable property,” and the Act allows it to be followed into the hands of others and traced into substitutes. But section 308 carves out the bona fide purchaser: where a person obtains the property “in good faith, for value and without notice that it was recoverable property,” it may not be followed into that person’s hands, and it ceases to be recoverable. The statute thus reproduces, in modern form, the very compromise of Miller v Race — and it tells us exactly what the strong protection requires. Not innocence alone. Innocence plus value plus absence of notice.
A recent decision sharpens the limits of that protection in a way that matters for the scalar argument. In R (World Uyghur Congress) v National Crime Agency (2024), the Court of Appeal considered whether goods alleged to be the product of forced labour could be cleansed of their taint somewhere along a supply chain by the payment of market value. The Court held that the provision of adequate consideration by one party does not prevent property from being criminal property in the hands of another party who has the requisite knowledge or suspicion; and it rejected the proposition that the chain could be broken “merely by the use of adequate consideration in any of the transactions involved.” The section 308 exception is real, but it is narrow and personal to the qualifying recipient. Taint is not laundered away by the bare fact of payment. This is the doctrinal expression of the scalar idea: the law does not let value pass clean just because money changed hands; it asks who knew what, and gave what, and in what circumstances.
The state’s claim: seizure, forfeiture, and the discipline of proportionality
Private recovery is only half the picture. The state has its own claim on tainted value, and it is a different kind of claim, serving a different end.
Under the Proceeds of Crime Act 2002, the state can pursue value along several distinct tracks. Conviction-based confiscation follows a criminal conviction and is defendant-focused and value-focused: it orders the defendant to pay a recoverable amount, rather than necessarily seizing specific notes. Civil recovery under Part 5 operates in rem, against the property itself, on the civil standard of the balance of probabilities, and crucially without any requirement of a conviction — indeed, civil recovery orders have been made against people acquitted in the criminal courts. Cash seizure and forfeiture allow authorised officers to seize cash on reasonable grounds for suspecting it to be recoverable property or intended for use in unlawful conduct, with detention and forfeiture then subject to court process. Restraint orders, investigative orders, and taxation powers fill out the toolkit.
The point of the public scheme is not to return property to a private claimant. It is to disrupt criminal benefit, to prevent future unlawful use, and to defend the integrity of the monetary system as such. That is why it is structured around property and suspicion rather than around private title and proof of a specific wrong.
And that structure carries an unavoidable governance tension, which is really a rule-of-law tension. Seizure begins with suspicion; forfeiture and recovery require proof and process. If the thresholds are set too low, lawful cash users are burdened and the mere act of holding cash becomes a ground for suspicion — a corrosive outcome for financial privacy and for the presumption that ordinary people may use ordinary money without explaining themselves. If the thresholds are set too high, cash becomes a convenient laundering medium and the integrity goal collapses. The whole legitimacy of the regime depends on whether suspicion, proof, process, and proportionality are kept in proper alignment. This is the public-law face of the same calibration problem the private law faces: enough recoverability to protect the system, not so much that ordinary use is criminalised by default.
Digital cash is not cash because it moves digitally
Now extend the structure to the digital world, where the temptation to reason from labels is strongest and most dangerous. A system is not “cash” merely because its users can send units to one another peer-to-peer. The marketing word “digital cash” decides nothing. What decides things is a set of attributes: finality, control, governance, reversibility, traceability, legal status, settlement dependency, value stability, and recoverability. Test the system against those, not against its branding.
Run the test across five families. Physical cash gives direct possession and genuine bearer transfer. Bank deposits are not bearer anything; they are account-based claims against an institution, mediated and reversible through that institution and the courts. E-money and payment apps are contractual claims operated through regulated intermediaries. Cryptoassets are ledger-based assets whose operation in practice depends on a whole apparatus of protocol rules, validators or miners, software clients, exchanges, wallet providers, developers, courts, and social coordination. Central bank digital currencies, if issued, would be sovereign digital liabilities whose legal nature would depend almost entirely on design choices: account- or token-based, privacy-preserving or surveilled, programmable or not, offline-capable or not, intermediated or direct.
Figure 2. Four layers, six instruments. The same four questions — what form the value takes, what function it performs, who governs its alteration and reversal, and how it is recovered — receive different answers across the payment landscape. The governance row is the analytically decisive one: the darker the cell, the more concentrated the control over alteration, freezing, and reversal. Physical cash is the lightest; custodial stablecoins and design-led CBDCs are the darkest.
English law has already done much of the conceptual work. In AA v Persons Unknown (2019), the Commercial Court held that a cryptoasset such as bitcoin is property, applying Lord Wilberforce’s classic four-part test in National Provincial Bank Ltd v Ainsworth — that a property right must be definable, identifiable by third parties, capable of assumption by third parties, and have some degree of permanence. Bitcoin satisfies that test because it is rivalrous: the holding of a unit by one person prevents another from holding that very unit, a feature that arises not from law or convention but from the way the software works. The Law Commission’s Digital Assets report (2023) reached the considered conclusion that the common law can recognise a distinct, “third category” of personal property — beyond the traditional dichotomy of things in possession and things in action — capable of accommodating crypto-tokens, and recommended legislation to confirm it. That recommendation became the Property (Digital Assets etc) Bill, introduced to Parliament in September 2024, whose operative clause provides that a thing is not prevented from being the object of personal property rights merely because it is neither a thing in possession nor a thing in action.
None of this makes crypto “cash.” It makes crypto property — which is a different and, for our purposes, more important point. A digital object can be transferable and valuable without having the legal or economic status of cash; and conversely, any state design for genuine “digital cash” would have to reproduce cash’s institutional features — final settlement, broad acceptability, resilience, calibrated privacy, and lawful recoverability — rather than merely its peer-to-peer feel. “Digital cash” is not a technical category. It is a legal-governance category.
Blockchains are governance-bound, and code is not law
This brings us to the central modern claim, and to its refutation. The slogan “code is law” holds that a sufficiently decentralised protocol places transactions beyond the reach of courts: what the ledger records is final, full stop. The slogan is false, and the reason it is false is the reason the strong version of cash-finality is false. Blockchains do not abolish governance. They relocate it.
A distributed ledger as used by actual people depends, at every turn, on institutions and choices. Software gets updated; protocols get forked; validators and miners apply policies; exchanges list, delist, and freeze; wallet providers set defaults; node operators choose which implementation to run; issuers of stablecoins can freeze and blacklist; courts make orders; regulators enforce sanctions. Even where a base ledger is technically resistant to unilateral alteration, none of that resistance survives contact with the institutional layer. And the crucial analytical move is this: every one of those choices is governance. A fork is governance. A protocol upgrade is governance. Refusing to upgrade is governance. Exchange delisting is governance. Validator censorship is governance. A stablecoin freeze is governance. A court-ordered recovery is governance. Developer discretion is governance, where developers control the dominant implementation. Distributed control is still control. Fragmented governance is still governance. There is no point in the system at which “no one is in charge” in the sense the slogan requires.
The English courts have already stared directly at this question. In Tulip Trading Ltd v van der Laan (2023), a company that owned a large quantity of bitcoin, but had lost the private keys controlling it, sued the developers of several networks — including the developers associated with BTC and the Bitcoin Association for BSV — arguing that they owed it fiduciary and tortious duties to take steps that would restore its access. At first instance the claim was struck out. The Court of Appeal reversed, holding that there was a serious issue to be tried: it was arguable that developers who exercise real control over a network’s software, making discretionary decisions on behalf of all participants, occupy a role with fiduciary characteristics. Birss LJ framed the claimant’s case as the contention that “the decentralised governance of bitcoin really is a myth,” and held that whether decentralisation is genuine is a question of fact to be established at trial on evidence, not a legal shield to be assumed in the developers’ favour. His most quoted line is the one that matters most for the slogan: “the internet is not a place where the law does not apply.”
The state of the authority should be stated precisely. There is, as yet, no English decision holding that developers owe such duties; there is a Court of Appeal decision that the question is arguable enough to be tried — that the issue turns on facts about how control is actually exercised over a network, and that those facts cannot simply be assumed away by invoking the word “decentralisation.” That is itself a significant ruling for the thesis here. It means the question of who governs a blockchain is a justiciable legal question, to be answered by examining where control really sits, rather than a matter the protocol can place beyond the court’s reach by description alone. Whether or not a fiduciary duty is ultimately made out on any given set of facts, the court’s willingness to entertain the question is the point: the governance of a ledger is something the law will look into.
The argument for the proposition does not, in any event, depend on the outcome of any single piece of litigation. It rests on the distinction between two kinds of finality. Technical finality means the system treats a transaction as settled according to its own rules. Legal finality means the law declines to unwind, reallocate, compensate, seize, or otherwise disturb the value consequences. These are not the same thing, and conflating them is the whole error. A technically final transfer can still generate a proprietary claim, a restitutionary claim, criminal liability, a tax consequence, fiduciary liability, sanctions exposure, or a court-ordered remedy. Code can make a transfer happen. It cannot determine whether the signer was coerced, mistaken, hacked, defrauded, insolvent, mentally incapacitated, acting outside their authority, laundering proceeds, breaching a trust, or subject to a freezing order. Those are questions of right, and questions of right are decided by law. Code may settle the ledger; law settles the rights.
Where digital cash is recoverable: from possession to control points
If blockchains are governance-bound, then the recovery framework built for physical cash carries over — but with one structural change. For physical notes, recovery is possession-based: you chase the thing or its traceable substitute. For digital value, recovery shifts to control points — the institutions and parties with the practical ability to freeze, transfer, or surrender the asset.
Digital value can be reclaimed in several ways, none of them defeated merely by the assertion of technical finality. It can be recovered as property, where the legal system recognises the asset as ownable — which, as we have seen, English law now does. It can be recovered by tracing, where movement through addresses, accounts, exchanges, bridges, and substitutes can be established. It can be recovered through restitution, where a recipient has been unjustly enriched. It can be recovered through court orders against intermediaries — exchanges, custodians, wallet providers, stablecoin issuers — who hold the practical levers. And it can be recovered through criminal confiscation or civil recovery where the value represents proceeds of crime, the digital case being no different in principle from the cash case.
The most instructive recent decision shows both the promise and the limits, and it refuses to flatter either side. In D’Aloia v Persons Unknown (2024), the High Court held — for the first time at a contested trial rather than an interim hearing — that a stablecoin, Tether (USDT), is property under English law: neither a thing in possession nor a thing in action, but a distinct form of property capable of being traced and of forming the subject matter of a trust. That is a significant doctrinal win for victims. And yet the claimant lost against the exchange he was pursuing. He had been defrauded of around £2.5 million in stablecoins, which moved through fourteen “hops” before some portion reached a wallet on the exchange Bitkub. The court accepted that a constructive trust could in principle bind an exchange that received identifiable misappropriated crypto, and that the asset could in principle be followed through a mixture. But on the facts, the claimant’s expert evidence — a “black box” tracing methodology that could not be transparently audited — failed to prove that his coins, rather than someone else’s, had reached the wallet. The claim failed not because crypto is beyond recovery, but because recovery, like all recovery, depends on proof.
That is the honest shape of digital recoverability, and it points in both directions at once. The law is willing to treat digital value as property, to trace it, to impose trusts on those who receive it with notice, and to order intermediaries to give it up. But the practical difficulty of tracing fungible value through mixtures and across jurisdictions is real, and it is often decisive. Recoverability is genuine; it is also hard. And critically, it scales with control: a custodial stablecoin is far easier to freeze than a bearer banknote, a transparent ledger far easier to trace than a privacy coin, a programmable CBDC reversible by design if a state chooses to build it so. The cruder claim — that digital assets are simply unrecoverable — is false. The accurate claim is that recovery has migrated from the holder’s pocket to the system’s control points, and that whoever holds those points answers to the courts.
The law of controlled amnesia
Return, finally, to the negotiation we began with, because the whole field reduces to it.
Cash is best understood not as a form of money but as a legally tolerated form of bounded bearer finality. That tolerance is strongest where value is low, exchange is ordinary, and the social benefit of frictionless payment exceeds the risk of concealed wrongdoing. It weakens as value, opacity, abnormality, and systemic consequence rise. A five-hundred-dollar cash payment is cash in the full sense: ordinary, final, private, practically uninvestigated. A ten-million-dollar transfer in banknotes, tokens, stablecoins, or ledger entries cannot claim the same treatment merely by wearing a cash-like form. At that scale the transaction is a governance problem, and the law treats it as one.
Digital systems do not escape this logic; they re-enact it. Blockchains, smart contracts, stablecoins, and CBDCs are not outside the law. They are governance-bound arrangements whose rules can be altered, contested, enforced, ignored, or overridden by the institutions and courts that surround them. Code can execute a transfer. It cannot determine whether the transfer was valid, recoverable, criminal, taxable, voidable, or final. Those determinations belong to law, and they always have.
None of this means cash should be abolished, and the governance counterweight is worth stating because it is easy to forget amid the enforcement logic. Cash is in retreat but it is not obsolete, and the law has recognised as much. In the United Kingdom, the share of payments not made in cash rose from 46% to 86% between 2002 and 2022, thousands of free-to-use ATMs were removed, and bank branches closed at pace. Yet around three million adults still rely on cash, a substantial share of small businesses still take most of their takings in it, and a global IT outage in 2024 reminded everyone what dependence on a single digital rail can cost. The Financial Conduct Authority’s access-to-cash regime, in force from 18 September 2024 under powers granted by the Financial Services and Markets Act 2023, now requires designated banks and building societies to assess and fill significant gaps in cash provision. Inclusion, privacy, and resilience are not nostalgia. They are reasons the institution of cash still earns its keep — and reasons the right answer is calibration, not abolition.
Figure 4. Cash is in retreat, but not obsolete — the governance counterweight. The collapse of cash’s share of payments sits alongside the persistence of a population that depends on it, which is why the policy response has been to guarantee access rather than to let cash disappear. Figures drawn from the FCA’s access-to-cash policy materials (PS24/8) and related FCA statements, 2024.
So the field has a shape, and the shape is a single principle worn by many doctrines. Cash is recoverable when the law can identify a claimant, a wrong, a proprietary or restitutionary path, a statutory basis, or criminal taint — without defeating the ordinary function of cash as circulating money. It is not recoverable merely because some prior holder suffered a loss, if the value has passed in good faith, for value, without notice, and in ordinary circulation. The same logic governs digital value, with recovery migrating from possession to control points but never disappearing.
The law of cash, in the end, is a law of controlled amnesia. Cash must forget enough of its past to circulate, but not so much that theft, mistake, fraud, and criminal proceeds become legally purified by movement. The strong claim made for distributed ledgers — that technical finality can manufacture the kind of forgetting that places value beyond the reach of courts — asks the law to forget everything. It will not. It never has. And the cases now coming through the courts, on stablecoins and tracing and developer duties, are the law saying so, one judgment at a time.
Notes and references
The standard applied here is set out in the note that follows. Primary legal authorities are cited in the conventional form; the holdings, and the quoted passages, were taken from the judgments and the legislation themselves, not from secondary summaries.
Cases-
Miller v Race (1758) 1 Burr 452; 97 ER 398 (KB) — bank note treated as currency; no property in currency once paid away in good faith for value; per Lord Mansfield.
-
Banque Belge pour l’Etranger v Hambrouck [1921] 1 KB 321 (CA) — tracing of misappropriated money through bank accounts.
-
Lipkin Gorman v Karpnale Ltd [1991] 2 AC 548 (HL) — action for money had and received; restitution for unjust enrichment against an innocent recipient of stolen money; recognition of the change-of-position defence.
-
AA v Persons Unknown [2019] EWHC 3556 (Comm) — cryptoassets are property, applying the property test in National Provincial Bank Ltd v Ainsworth [1965] 1 AC 1175 (HL) (per Lord Wilberforce); rivalrousness of bitcoin.
-
Tulip Trading Ltd v van der Laan [2023] EWCA Civ 83 (CA) — arguable (serious issue to be tried) that software developers owe fiduciary/tortious duties to coin owners; whether a network is genuinely decentralised is a question of fact to be established on evidence at trial; “the internet is not a place where the law does not apply” (Birss LJ). The Court did not decide that such a duty exists. The fiduciary test applied was that in Bristol and West Building Society v Mothew [1998] Ch 1 (CA).
-
D’Aloia v Persons Unknown Category A & Ors [2024] EWHC 2342 (Ch) — first trial-level holding that a stablecoin (USDT) is property under English law; constructive trust against a receiving exchange arguable in principle; tracing through mixtures available in principle; claim nonetheless failed on the facts for want of transparent, reliable tracing evidence.
-
R (World Uyghur Congress) v National Crime Agency [2024] EWCA Civ 715 (CA) — adequate consideration anywhere in a supply chain does not, by itself, cleanse property of criminal taint or break the chain; analysis of POCA s 308.
Legislation-
Proceeds of Crime Act 2002, Part 5 (esp ss 240–241, 266, 278, 304, 305, 308) — recoverable property; following and tracing; the bona fide purchaser exception in s 308 (”good faith, for value and without notice”). Part 7 (esp ss 327–329, 340) — money-laundering offences and the meaning of criminal property. — Financial Services and Markets Act 2023; Financial Services and Markets Act 2000, Part 8B — statutory basis for the FCA access-to-cash regime.
-
Property (Digital Assets etc) Bill (introduced to Parliament 11 September 2024) — operative clause: a thing is not prevented from being the object of personal property rights merely because it is neither a thing in possession nor a thing in action.
Official and regulatory materials-
Law Commission, Digital Assets: Final Report (Law Com No 412, 2023), and Digital Assets as Personal Property: Supplemental Report and Draft Bill (2024) — recommendation to confirm a distinct “third category” of personal property for certain digital assets.
-
Financial Conduct Authority, Access to Cash (Policy Statement PS24/8, July 2024) and associated FCA statements (2024) — rules in force 18 September 2024; fourteen designated firms; cited figures on the decline of cash payments (46% → 86%, 2002–2022), ATM removals, SME cash reliance, and the c. 3 million adults who rely on cash. — Bank of England, public guidance on legal tender and banknotes — legal tender concerns the discharge of debts and does not compel retail acceptance.
Scholarship (cited as relied upon, not independently vouched for)
— Angela Walch, ‘In Code(rs) We Trust: Software Developers as Fiduciaries in Public Blockchains’ in Philipp Hacker and others (eds), Regulating Blockchain: Techno-Social and Legal Challenges (OUP 2019) — cited because the Court of Appeal in Tulip Trading expressly referred to it; it is reported here as a source the court relied upon, not as a work whose full argument has been independently re-analysed for this essay.
Economics (canonical works, cited for their established central propositions)-
R H Coase, ‘The Problem of Social Cost’ (1960) 3 Journal of Law and Economics 1; and ‘The Nature of the Firm’ (1937) 4 Economica 386 — transaction costs as the determinant of institutional arrangements.
-
Douglass C North, Institutions, Institutional Change and Economic Performance (CUP 1990) — institutions as the “rules of the game” that reduce uncertainty in exchange.
-
Oliver E Williamson, The Economic Institutions of Capitalism (Free Press 1985) — governance structures as responses to transaction hazards.
A note on sources and the standard applied
This essay is built, deliberately, on a backbone of primary legal sources — decided cases and statutes — together with official and regulatory publications. Those are the load-bearing citations, and their holdings and quoted words were taken from the judgments and legislation themselves. Where a normative or evaluative claim is made (for instance, that “calibrated finality” is the right governance posture, or that cashness is best understood as scalar), it is presented as the essay’s argument, not as a demonstrated fact; the descriptive legal propositions are what the cited authorities establish.
Two further points of candour. The economics references (Coase, North, Williamson) are cited for their established central propositions, which are not in dispute, rather than for fine-grained empirical findings; readers should treat them as canonical signposts, not as freshly re-derived results. And the single secondary academic source named — Walch — is cited only because the Court of Appeal relied on it; it is reported as the court’s reliance, not as a work whose full text has been independently analysed here.