The tiptoe to crypto: An analysis and account of property in cryptocurrency

Abstract

This paper argues that cryptocurrency, including cryptocurrencies such as Bitcoin, should be understood as qualified property. We build up support for this claim in three stages: first, we outline the diversity of cryptocurrencies, a diversity which is underappreciated in the current literature. We then outline the importance of theoretical presuppositions which operate in the ‘cryptocurrencies as property’ debate. This is followed by a detailed critique of established positions. We explore the shortcomings of three commonly-held views in the debate: that cryptocurrencies are property, belonging to a third category of personal property beyond the choses in possession/choses in action distinction; that cryptocurrencies are property and belong to the category of choses in action; and, lastly, that cryptocurrencies are not property at all. We then develop a new analysis of property in cryptocurrencies, taking into account both the variety of cryptocurrencies which exist and capturing the doctrine operating currently in property law, including its theoretical underpinnings. Through understanding rights in property as relative, and property itself as a scale, we avoid the shortcomings of other popular accounts whose shortcomings we identify, and show that cryptocurrencies are a form of choses in possession – namely, qualified property.

Introduction

The reader, upon seeing the title of this paper, might be forgiven for exclaiming “Not another paper on crypto! Isn’t there anything else to write about?!” In our defence, we argue, as every good academic does, that the current literature suffers from three shortcomings that our paper aims to address, namely:

It is not comprehensive and does not appreciate that different forms of cryptoassets have different characteristics which affect their proprietary status;

It overlooks serious theoretical controversies concerning exactly what the law of property is about; and,

It assumes a uniform concept of property which simply doesn't exist in common law.

Our paper addresses these shortcomings by providing a comprehensive overview of cryptocurrency and blockchain technology, outlining competing understandings of property, critically analysing current understandings of cryptocurrencies’ proprietary (or non-proprietary) nature, and setting out a novel proprietary understanding of cryptocurrency – that cryptocurrencies amount to choses in possession and are a form of qualified or special property. It is important to note that although this paper has implications for the broader debate concerning whether digital files and assets, generally, amount to property, ¹ as well as whether cryptoassets such as NFTs (non-fungible tokens) are proprietary, its focus is on cryptocurrency with the authors’ views on the more general issue being reserved for future work.

By cryptocurrency, we mean crypto coins, which are mediums of exchange and the only user-functionality which they have is the power to alter a decentralised ledger which tracks transfers between different addresses; that is “all” that the user can do with crypto coins. Notable examples of crypto coins are Bitcoin and Ethereum (ETH). Crypto tokens, on the other hand, outstrip this functionality vastly, and can be ‘used in countless ways’ ² – they can be transferred and exchanged because they are programmed upon “Layer 1” cryptocurrency blockchain protocols – that is, crypto coins such as Bitcoin and ETH. They are also, however, capable of a far wider variety of actions which can be coordinated and scripted through the use of smart contracts. These uses include the generation and storage of art (e.g. NFTs) and voting in and controlling quasi-corporate entities (DAOs). As we do not consider crypto tokens in this paper, leaving them for future work, we will not discuss them further.

Blockchain technology and the varieties of cryptoassets

Blockchain technology is ultimately nothing but code or data. What makes it a unique information technology from the perspective of property law is the synergistic and complementary use of peer-to-peer networking, cryptography, and consensus mechanisms to create a distributed ledger system across a network of computers. ³ This stores information in a way which is typically decentralised (in the sense that no single party can unilaterally alter the information), transparent, immutable, and secure. Any one of those features may be lacking in any given blockchain, however. They can be permissioned or permissionless, with differing levels of restricted access to a chain on the basis of identity. They can be widely distributed and decentralised (across e.g. all users) or distributed across a smaller network of central servers. To participate in a chain, one typically requires a pair of cryptographic keys: a public key (visible to others) and a private key (kept secret). These keys determine one's identity on the network and correspond to an address through which one can transact on the network – e.g. by receiving and sending coins. ⁴ Cryptocurrencies come in a wide variety of forms and are therefore very diverse.

The majority of cases and legal literature regarding blockchain and property focus on ‘crypto coins’, as we term them. Since they are predominantly used as a means of storing and exchanging value, they are mostly contrasted with state-issued (or “fiat”) currency. While fiat currency depends upon the legal authority and power of a central institution (such as a reserve bank) to create and control money, blockchains verify and authorise the creation and transfer of ‘coins’ through what are called consensus mechanisms – ‘self-regulatory stack[s] of software protocols written into a blockchain's code’. ⁵ While this eliminates any central point of failure, vulnerabilities are nevertheless discovered and exploited as the market share and popularity of cryptocurrencies continue to grow. Therefore, despite cryptocurrency users often aiming to be ‘wholly independent of any legal system’, ⁶ many end up seeking the law's protection. The law, nevertheless, struggles to keep up with these technological developments, opening a gap between implementations of cryptocurrency technology and how they are understood by the law.

One such important gap is the underappreciated diversity of operation between different cryptocurrencies. While Bitcoin and ETH are known as decentralised, secure, and transparent coins, many new alternative coins and tokens (varieties of which are known as ‘alt-coins’, ‘meme-coins’, and ‘shit-coins’) have differing qualities. These different features are coded via smart contracts and can include (the ability to introduce) trading ‘taxes’, ⁷ restrictions or pauses on trading (often an element in ‘honeypot’ scams), ⁸ maximum transaction and token holding limits (sometimes only for specific addresses), ⁹ token ‘self-destruction’, ¹⁰ and external modification of the balance of any token holder (one form of a ‘rug pull’). ¹¹

Many of the above abilities are triggered by the smart contract automatically but many are also, depending on the code and smart contract, features which can be controlled by a specific address in the network – that is, a specific participant. This person is very often the creator of the token or the ‘owner’ of the contract. If the contract's owner has special access to the contract and can modify it unilaterally, there is a significant risk of token holders losing their assets and the whole situation imperils the token's status as property because the token is not sufficiently stable, as will be discussed below. Depending on the contract, the owner may have a variety of powers, many of which are outlined above. One particularly well-known kind of scam is a ‘rug pull’, where the contract owner unilaterally transfers all users’ tokens. ¹² To relinquish this kind of problematic level of control, legitimate coin developers ‘renounce' their contracts. This means that the above functions are not modifiable and no participant has special powers in the contract anymore. ¹³ However, ownership retrieval can be coded into the coin’s smart contract, to reverse renouncement. Similarly, ownership can be hidden in the code, so that a developer can still manipulate a token's smart contract, even if ownership appears to have been abandoned. Ultimately, all owner powers (even ‘hidden’ ones) are contained in the smart contract code and can be scrutinised by investors and algorithms.

In addition to the above smart contract features, the crypto coin chains themselves (e.g. Bitcoin and ETH) can have different features which could be relevant to a property law analysis. Most cryptocurrency blockchain protocols are stable and their histories are encoded into their very being; for this reason, if a developer or community wanted to alter any part of the rules, they must create a copy of the chain and its history, called forking, to encode different rules into it. ¹⁴ Precisely because the rules cannot be changed by a central authority, a ‘fork’ is required to create a new copy of the coin with the modified rules in place. This is, strictly speaking, a new coin and the old coin's existence continues – where developers and the broader community decide on the fork, they will follow the new ‘version’ of the coin rather than the old remnant (which can continue being used separately). Indeed, a process like this is what led to the co-existence of ETH and Ethereum Classic (ETC).

As the above indicates, cryptocurrency is not just a single thing to which a single, monolithic category of property law can be applied. The result is that the devil is very much in the detail. Indeed, the diversity of cryptocurrencies dovetails with the fact that property in the common law is ‘relative’ ¹⁵ with innumerable gradations ¹⁶ so that the proprietary ‘strength’ of cryptocurrencies, as with property generally, varies from coin to coin. ¹⁷

What do we mean by ‘property’?

Traditionally, property has been conceived of as the law concerning things; thus Gaius in the Institutes states that ‘[a]ll our law concerns [either] persons or things or actions’ ¹⁸ with Bracton restating this in the Middle Ages, ¹⁹ and Blackstone stating in his Commentaries that

the objects of our inquiry in this second book will be the jura rerum, or those rights which a man may acquire in and to such external things as are unconnected with his person. These are what the writers in natural law style the rights of dominion, or property. ²⁰

On such an approach, the ‘thingness’ – or, in the case of incorporeal things, the ‘thinglikeness’ ²¹ – of x is key to determining its proprietary nature, both in terms of establishing whether it is property at all – i.e. whether x ‘is the subject-matter of property’ ²² – and, if it is property, the degree to which it is so. ²³ Thus, where x has a mind of its own, such as a wild animal, or is transient, such as water, fire, light, air and so on, their thingness is such that they can only be property in a weak or qualified sense. ²⁴ On the other hand, land has the highest degree of ‘thingness’ given that it is ‘permanent, fixed and immovable’ ²⁵ so that one can ‘have a certain substantial property’ in it. ²⁶ In between these two extremes is an almost infinite number of things which all fit in at slightly different points on the ‘scale’ so that property can be said to have ‘an almost infinitely gradable quality’; ²⁷ nevertheless, the question is always: is x a thing and, if so, to what degree?

The ‘property as things’ approach can be contrasted with a ‘property as rights’ approach, reflected perhaps most clearly in Hohfeldian theories of property law. Hohfeld holds that any property ownership, possession, or other claim can be described using a set of four sui generis bilateral (in the sense of person-to-person) legal relations which (may) hold as between the claimant and every other relevant individual, simultaneously. This is because ‘all legal interests are “incorporeal” – consisting, as they do, of more or less limited aggregates of abstract legal relations’. ²⁸ – These legal relations only hold between individuals – there is no property in things or relationship of right as between an individual and a thing. Indeed, on such an approach, it does not matter, for property law, whether x is or is not a thing; rather, what matters is whether a person can have a privilege, power, or immunity in relation to it. ²⁹

The result is that those who adopt such an approach, on the one hand, and those who adopt a law of things approach, on the other, are like ships passing in the night – both the questions they ask and the answers they formulate are different. As we shall see below, this explains the fundamentally different approaches taken to cryptocurrencies by scholars and their inability to substantively engage with each other's arguments. It is for this reason that, although the point is generally overlooked, it is crucial for policymakers and courts to be aware of their own theoretical predicates as well as those operating in the literature before citing a particular authority to support their conclusions about cryptocurrency's proprietary nature. In rebutting such views, it is also necessary to both explain why the underlying philosophical predicate is problematic and, on occasion, to meet it on its own turf by explaining why, even on that view, it is incorrect.

In consequence, it is important for this paper to clearly state its philosophical predicates: it adopts a ‘law of things’ approach to property. It does this because, as shown above, this is the traditional approach of the common law. In addition, the debate concerning cryptocurrency's proprietary nature is ultimately about whether cryptocurrencies are suitable subject matter for property, and a Hohfeldian approach which necessarily eschews any analysis of ‘thingness’ or the nature of the thing cannot tell us whether cryptocurrencies are or are not suitable subject matter for property and why or why not. ³⁰ Examples of this can be seen in Robert Stevens' and Kelvin Low's approaches to cryptocurrency, discussed below, with the former ultimately requiring policy arguments to buttress a generally Hohfeldian argument as to why cryptocurrency should not be property, and the latter, in effect, arguing that as anything can be property, so can cryptocurrency.

Moreover, beyond merely capturing and justifying doctrine as it currently stands, a thing-focused approach is helpful when analysing cryptocurrency because it acknowledges that ‘property is something which has an existence in the real world. Specifically, property is a practice, a way of dealing with things in which people actually engage’. ³¹ This understanding of property accords with the reality of both legal and lay practice regarding property, as well as the practice of the crypto community. Just as ‘[n]orms in rem establish the general, impersonal practices’ that ‘allow strangers to interact with each other in a rule-governed way’, ³² the code and structure of blockchain technology and cryptocurrencies are similarly impersonal and facilitative, to a perhaps unique degree, allowing strangers to interact in a rule-governed way. Likewise, the ‘real world’ of cryptocurrency can be used to identify the actual thing which is property. The Law Commission does exactly this when it says that cryptocurrencies are ‘composite things’, ‘a notional quantity unit manifested by the combination of the active operation of software by a network of participants and network-instantiated data’. ³³

Brief overview of the current state of personal property law

As currently understood, the orthodox position in English and Commonwealth property law is that ‘[p]roperty in chattels personal may be [only] either in possession; which is where a man hath not only the right to enjoy, but hath the actual enjoyment of the thing; or else it is in action; where a man hath only a bare right, without any occupation or enjoyment’. ³⁴ Thus, Fry LJ in Colonial Bank v Whinney ³⁵ stated that ‘all personal things are either in possession or in action. The law knows no tertium quid between the two’. ³⁶ That view was subsequently upheld by the House of Lords ³⁷ and has been recently restated by the Court of Appeal in Your Response Ltd v Datateam Business Media Ltd. ³⁸ The distinction between choses in possession and choses in action matters because, among other reasons, choses in action cannot be the subject of the tort of conversion. Thus, the House of Lords in OBG V Allan ³⁹ refused to extend the tort of conversion to cover the latter because this would require ‘the … severance of any link whatever between the tort of conversion and the wrongful taking of physical possession of property … having a real and ascertainable value’. ⁴⁰

Despite clear reasoning and precedent, the bipartite division of personal property has come under significant pressure in recent years as a result of growing digitalisation, including the development of cryptocurrencies. We will explore the issue in depth below via an examination of the current proposed property law approaches to cryptocurrency, before then outlining our view that cryptocurrencies are choses in possession, and specifically a sub-set of them known as ‘qualified’ or ‘special’ property, which, as we will explain, best reconciles these existing authorities with the technological realities of cryptocurrencies and the structure of property law more broadly.

Current approaches to cryptocurrency's proprietary (or non-proprietary) nature

In general terms, there are three current approaches to cryptoassets:

Cryptocurrencies constitute a third category of personal property

Cryptocurrencies constitute choses in action

Cryptocurrencies are not property

Each of these approaches will be outlined and critiqued in turn below with reference to the leading judicial, scholarly, and policy proponents and authorities. Subsequently, we shall set out an overlooked fourth, and what we believe to be the correct, approach – that cryptocurrencies are choses in possession and specifically a form of qualified property.

Cryptocurrencies as a third category of personal property

The earliest application of the ‘third category’ approach to cryptocurrencies comes from the LawTech Delivery Panel’s ‘Legal statement on cryptoassets and smart contracts’ by the UK Jurisdiction Taskforce. The legal statement effectively adopts a functional approach, arguing that if cryptocurrencies meets the four-limbed test for property in Ainsworth ⁴¹ – namely, that it be ‘definable, identifiable by third parties, capable in its nature of assumption by third parties, and have some degree of permanence or stability’ ⁴² – it should amount to property, whether or not it fits into the existing categories of personal property.

The taskforce argued that both definability and certainty were met by virtue of the ‘public parameter of a cryptoasset’ ⁴³ – namely, ‘encoded information about the asset, such as its ownership, value and transaction history’. ⁴⁴ The requirement for control and exclusivity is also ‘satisfied by the cryptographic authentication process, which permits the holder of the private key, and only that holder, to deal in the cryptoasset, and therefore to control it to the exclusion of others (subject to the powers of other keyholders where there are multiple private keys)’. ⁴⁵ Similarly, the Legal Statement holds that ‘cryptoassets appear to be as permanent as other conventional financial assets, which may exist only until they are, for example, cancelled, redeemed, repaid or exercised’. ⁴⁶ The final requirement, that of stability, was more complicated because total stability does not arise until consensus is achieved and the risk of forking remains, as discussed earlier. ⁴⁷ Nevertheless, the Panel maintained that ‘even without resolving those issues, cryptoassets are in our view sufficiently permanent or stable to be treated as property’. ⁴⁸

The statement distinguishes the orthodox authorities discussed above by arguing that whilst Fry LJ's statement ‘may carry the logical implication that an intangible thing is not property if it is not a thing in action[,] [i]t is not clear, however, whether Fry LJ intended that corollary, and it should not in any case be regarded as part of the reasoning leading to his decision’. ⁴⁹ Similarly, the House of Lords' approval of his statement ‘did not explicitly address the issue of exhaustive classification between things in action and things in possession and said nothing about the definition of property’ ⁵⁰ and, in any event, ‘to the extent that the House of Lords agreed with Fry LJ on the classification issue, that seems to have been on the basis that the class of things in action could be extended to all intangible property’. ⁵¹ Similarly, OBG and Your Response did not ‘go so far as to hold that intangible things other than things in action could never be property at all, only that they could not be the subject of certain remedies’. ⁵² Lastly, the statement noted certain recent decisions in which ‘the courts have found no difficulty in treating novel kinds of intangible assets as property’. ⁵³

Although prima facie convincing, there are significant issues with the Legal Statement's arguments that cryptocurrencies are a third kind of property. Firstly, as quoted above, Fry LJ is quite clear that all personal property is either in possession or in action, ⁵⁴ and the latter are things that cannot be occupied or enjoyed except via a right of action. ⁵⁵ This directly conflicts with the interpretation proposed by the Legal Statement, Law Commission, and others, so that their interpretation is simply untenable.

Secondly, it does not follow from Lord Blackburn's statement that ‘it was hardly disputed that, in modern times, lawyers have accurately or inaccurately used the phrase “choses in action” as including all personal chattels that are not in possession’ ⁵⁶ that all intangible property amounts to choses in action. This is because the statement says nothing about the nature of choses in action as such; it merely states that anything which is personal property but is not a chose in possession must be a chose in action. Thus, if choses in action must be enforceable by a right of action, as held by Fry LJ, Lord Blackburn's statement continues to make sense and, indeed, arguably makes more sense given that Fry LJ's view was upheld rather than repudiated by the House of Lords. As for OBG and Your Response, they may well not have explicitly held that there were only two forms of personal property but the decision was necessarily predicated on the point.

Lastly, none of the cases mentioned by the statement as recognising property that would amount to a tertium quid do so explicitly. Thus, in Swift v Dairywise Farms Ltd ⁵⁷ the Court of Appeal is at pains not to firmly conclude on the nature of a milk quota, prefacing its holding with the statement:‘[w]hatever the true nature of milk quota – and, in particular, whether or not it can properly be regarded as an asset within any existing category recognised by English domestic law’. ⁵⁸ Similarly, in Armstrong v Winnington, ⁵⁹ which concerned an EU carbon emissions allowance (EUA), Mr Stephen Morris QC, sitting as a deputy High Court Judge, merely stated ‘[f]or reasons set out below, ultimately I do not consider that it matters whether an EUA is a chose in action or merely some form of “other intangible property.”’ ⁶⁰ The authorities discussed by the judge in Armstrong likewise all concerned sui generis rights created by statutes. It is difficult to see how one can extract a cohesive ‘third category’ of personal property law from such disparate statutory schemes. The difficulty can be seen from the fact that, as will be discussed below, none of these sui generis rights would fall into the ‘third category’ of personal property as defined by the Law Commission.

Notwithstanding these issues, the reasoning of the Legal Statement was adopted by the English High Court in AA v Persons Unknown. ⁶¹ In that case, the Court held that cryptocurrencies did not fit into the two traditional categories of personal property given that ‘they are not choses in possession because they are virtual, they are not tangible, they cannot be possessed [nor are] they … choses in action because they do not embody any right capable of being enforced by action’. ⁶² Nevertheless, the Court held that cryptocurrencies could still amount to property given that they met the test for property laid down in Ainsworth of being ‘definable, identifiable by third parties, capable in their nature of assumption by third parties, and having some degree of permanence’. ⁶³

Most recently the UK Law Commission has analysed the issues posed by cryptocurrencies in its Report on Digital Assets. It represents by far the most detailed analysis of the property issues posed by cryptocurrencies to date. The report adopts a ‘law of things’ approach to property law; thus, in explaining the differences between choses in action and choses in possession, it states that

one distinctive feature of things in possession is that they are things which exist regardless of whether anyone lays claim to them, and regardless of whether any legal system recognises or is available to enforce such claims. In contrast, things in action have no independent form and exist only insofar as they are recognised by a legal system. ⁶⁴

This, in turn, explains why the Law Commission rejects the idea that cryptocurrencies are choses in action, as ‘the software and data together manifest a “thing” that is necessarily independent of any rights that may be claimed over that thing’ with the result that ‘[c]rypto-tokens would continue to exist even if the law [did not recognise them as property]’ and thus they cannot be choses in action. ⁶⁵ The Law Commission also rejects the idea that digital assets could amount to choses in possession on the basis that ‘[t]he legal concept of possession is traditionally limited to tangible things’ and thus digital assets cannot be choses in possession. ⁶⁶

In consequence, the Law Commission adopts a ‘third category’ approach to cryptocurrency and cites many of the same cases to support its view that such a category is in fact recognised by English law. ⁶⁷ To determine whether a thing falls into this third category the Law Commission sets out three requirements:

the ‘thing’ ‘is composed of data represented in an electronic medium, including in the form of computer code, electronic, digital or analogue signals’;

it exists ‘independently of persons and…of the legal system’; and,

it is ‘rivalrous’. ⁶⁸

The first requirement was abandoned by the Law Commission following the consultation paper on the basis that the issue was not so much the data present in a cryptocurrency but rather what the data allowed its holder to do. ⁶⁹ We will therefore not discuss it further.

The second requirement is intended to ensure that ‘only those things that are properly identified as distinct things, existing independently from any particular person’ ⁷⁰ can amount to third-category things. The second limb of the second requirement also prevents choses in action, and statutory property rights, such as IP rights, from amounting to third-category things. ⁷¹ It is appropriate to note here that it is fundamentally contradictory for the Law Commission to exclude ‘statutorily created property rights’ ⁷² from amounting to third-category things whilst at the same time heavily relying on cases concerning such rights to justify its view that the common law recognises such a category.

The third requirement would appear to be the most important given the Law Commission's repeated reference to it throughout the paper. This is also accepted by the Law Commission itself when it states in its supplemental report that ‘we consider that rivalrousness is a particularly important feature of things that are appropriate objects of property rights’. ⁷³ The (earlier) final report provides that ‘[a] thing is rivalrous if the use or consumption of the thing by one person, or a specific group of persons, necessarily prejudices the use or consumption of that thing by one or more other persons’. ⁷⁴ The Law Commission goes on to argue that cryptocurrencies are rivalrous because

[t]he use or consumption of the crypto-token by one person, or a specific group of persons, necessarily prejudices its use or consumption by one or more other persons. The technology ensures this. In other words, the technology ensures that a crypto-token cannot be double-spent. ⁷⁵

It is however arguable that the rivalrousness of cryptocurrency more generally is artificial and technologically contingent ⁷⁶ – it is not in the nature of the thing itself. Thus 51% (and even 34%) attacks by hackers are possible and have, when successful, allowed for limited double spending; likewise, forking allows for the duplication of such assets. Note that while it could be argued that forking does not duplicate a cryptocurrency but merely creates a new one, this position underappreciates that a forked currency is an exact mirror of the original chain – e.g. all transfers and wallet holdings are recorded in both. Anyone who held coins on the original chain now holds exactly twice as many, with identical holdings recorded on the new chain; this is an essential feature of forking – ‘two identical blockchains exist until changes to one of the chains are made’. ⁷⁷ Moreover – as has been outlined above – not all cryptocurrencies are rivalrous, let alone rivalrous to the same extent; thus some ‘shitcoins’ may not be rivalrous at all because, for example, the developer may have unlimited power to duplicate or otherwise interfere with them, other altcoins may be only partially rivalrous, and others still may initially be rivalrous but then be modified by their developers to become non-rivalrous or only rivalrous to a limited extent (as where the developer retains a power to self-destruct the cryptocoins or to unilaterally alter balances of the coin held in others' wallets). These technological realities suggest that cryptocurrencies in general are not as rivalrous as the Law Commission believes them to be – nor can the matter of rivalrousness be determined in the abstract – and we believe this suggests that a more nuanced, case-by-case approach is required.

It is true that the Law Commission accepts in its supplemental report that not all digital assets are rivalrous, distinguishing between an ordinary Word document (which is not rivalrous) and cryptoassets. However, again, the Law Commission merely asserts that cryptocurrencies are necessarily rivalrous ⁷⁸ and does not appear to be aware of the reality that certain cryptocurrencies will not, in fact, be rivalrous for the reasons we have outlined above. ⁷⁹ These weaknesses do not necessarily undermine the Law Commission's draft bill – but that is because the bill is anaemic: it merely proposes that just because x is not a chose in action nor a chose in possession, it is not necessarily not property. If this wording seems clumsy it is because the bill is phrased as a double-negative; in practice, its effect is that third-category things can amount to property but the bill does not actually mandate such recognition. ⁸⁰ It seems odd that several years of work and hundreds of pages of reports have been expended to justify such an unambitious outcome. However, from the authors’ perspective, the minimal nature of the proposed statutory intervention is a positive (despite the extensive resources expended) because the Law Commission's arguments in favour of the existence of such a third category are, as we’ve discussed above, undermotivated and mandating such recognition in a less minimal and more intrusive fashion would risk fundamentally unsettling centuries of common law development for very little gain. As it is, even if the bill were passed, courts could, and perhaps should, continue to classify purported third-category things into one of the generally accepted two categories of property, i.e. choses in action or choses in possession. In the specific context of cryptocurrency, we, as will be seen below, argue that they should be conceived of as a specific type of choses in possession, namely qualified property.

Cryptocurrencies as choses in action

The possibility that cryptocurrencies are choses in action has been pioneered in Singapore as a result of the work of Kelvin Low. His argument is alluringly simple: the category of choses in action is not restricted to those things which can be enforced by court action; rather, ‘the categories of intangible property and chose in action are quite simply co-extensive’. ⁸¹ Thus, anything which amounts to intangible property is necessarily a chose in action. Low's arguments are novel and well-made but suffer from two primary shortcomings.

The first is that it is arguably not supported by authority. Fry LJ in Colonial Bank held that company shares were choses in action because ‘of the share itself … there can be no occupation or enjoyment … No action is required to obtain … the right to the dividends or other accruing benefits in the case of the share; but an action is the only means of obtaining … [these] benefits’. ⁸² In other words, the reason that they are choses in action is not merely intangibility but also because the benefits of such things can only be obtained via a court action; the requirements are cumulative and meeting one alone is insufficient for x to amount to a chose in action. The second issue is that if a right of action is not necessary for property to amount to a chose in action, it is difficult to explain why the category is typically referred to as ‘choses in action’ rather than simply ‘intangibles’. In other words, the longstanding nomenclature of property law itself necessitates a link between such property and rights of action.

Low also makes a novel contribution regarding what exactly is proprietary vis-à-vis cryptocurrencies, arguing that it is ‘the legal right to have their bitcoins, or more accurately their unspent transaction output or UTXO, locked to their chosen public bitcoin address on the blockchain’. ⁸³ Low justifies this view on the basis that it prevents the law from recognising pure information as property which would ‘serve as a grave impairment of the free flow of information and the freedom of expression’. ⁸⁴ This view fits quite well with the technological functioning of cryptocurrencies, where it is only the private key holder's right to transfer, or control, the relevant cryptocurrency that matters. However, for this right to be proprietary in a system which views property law as the law of things, it must first relate to a thing; what, on Low's account, is the thing to which it relates?

Low does not answer this question but instead argues that a strict “thing-focused” approach is incoherent as English law recognises bank money as a chose in action – ‘specifically a debt owing by the bank to its customer’ – and ‘yet there doesn’t appear to be any separate thing that is the subject of a debt other than the debt itself’. ⁸⁵ Low therefore suggests a broader definition of property as referring to ‘the law's recognition of and willingness to enforce a holder's rights to exclude others from a resource, whether tangible or intangible, without necessarily providing any clues as to its exigibility’. ⁸⁶ However, it is not apparent that things are as clear-cut as Low makes out. If one accepts that “property” has two primary meanings, ‘both as things (be they intangible or intangible) and as rights in relation to things’, ⁸⁷ and that it can vary in strength depending on its similarity to tangible things or strong proprietary rights in relation to things, ⁸⁸ then much of the problem outlined by Low goes away. Moreover, Low's broader definition of property is unable, as Hohfeldian approaches generally are for the reasons outlined earlier, to account for the undoubted reality ‘that some things simply cannot be the subject-matter of property’. ⁸⁹ Lastly, such an understanding of property fundamentally disturbs the existing taxonomy of the common law which is predicated on the idea that the subject matter of a proprietary right necessarily affects the ‘strength’ and nature of that proprietary right. The terms “real”, “personal”, and “qualified” property cannot be explained, or understood, without an acceptance of the view that the subject matter of a purported proprietary right is fundamental in determining the extent and true nature of that right. The point will be discussed further below where the paper's novel view of cryptocurrency as a form of qualified property is outlined.

Despite these issues, the Singapore High Court appears to have adopted Low's reasoning in ByBit Fintech Ltd v Ho Kai Xin and others, ⁹⁰ stating that:

The argument that crypto assets should not be classified as things in action rests on the origin of this category as rights enforceable by action (in the sense of litigation in court) against persons … There is no individual counterparty to the crypto holder's right. But over time the category of thing in action has expanded [and] this diversity suggests that the category of things in action is broad, flexible, and not closed. ⁹¹

This would appear to incorporate Low's reasoning but the Court then goes on to state that ‘the holder of a crypto asset has in principle an incorporeal right of property recognisable by the common law as a thing in action and so enforceable in court’, ⁹² which appears to adopt the narrower view of a chose in action as being a property right enforceable only by legal action. We suggest that this is because, as we have discussed above, property in the common law is ultimately understood with reference to ‘a materialist archetype of property’; ⁹³ thus even when discussing dephysicalised ‘things’, common lawyers ‘understand property in terms of their experiences of material things’. ⁹⁴

Cryptocurrencies as non-proprietary

The view that cryptocurrencies are not property at all is a third view in the literature, one which is put forward succinctly and strongly by Robert Stevens. Stevens argues that Bitcoin, and cryptocurrencies more generally, are not property as they can neither be physically possessed nor does a crypto holder have any ‘cause of action to enforce a primary right or one to correct any infringement’, ⁹⁵ given that ‘no tort applicable to things’ applies to them and ‘the holder of the key has no contractual rights against anyone [nor any] statutory rights of enforcement’. ⁹⁶ Adopting a Hohfeldian approach to property, he argues that the end result is that the ‘correct holder of a key to a Bitcoin wallet’ does not have any ‘rights (or privileges, powers or immunities)’ in relation to the Bitcoin, ⁹⁷ and thus they do not hold any property rights.

As an initial point, it is worth noting that the traditional view that digital assets cannot be possessed has come under heavy attack in recent years, with courts in both New Zealand and Canada recognising that they can both be possessed ⁹⁸ and be the subject of conversion claims. ⁹⁹ The rejection of this possibility in England and Wales has also come under significant scholarly criticism ¹⁰⁰ and it is plausible that it does not accurately capture the state of the law (we address this point in further detail below where we set out why we believe such assets can be possessed). However, even if the traditional view outlined by Stevens is correct, this tells us nothing about why cryptocurrencies should not be recognised as property – for example, by statute – or why they could not be recognised if the legal position regarding digital assets changed. To plug this gap, Stevens sets out several policy arguments arguing that ‘[a]lmost all cryptoassets are unproductive and many are positively harmful. For most of their forms, our legal system should not be seeking to facilitate them but, alongside other jurisdictions, attempting to eliminate their use where possible’. ¹⁰¹

There are two significant problems with this argument. Firstly, it is fundamentally an argument from public policy and the role of public policy in property law is a limited one. Thus, whilst in extreme cases – such as those concerning the human body and body parts ¹⁰² or slavery ¹⁰³ – public policy can play a role in limiting or prohibiting the recognition of property rights, in general such rights ‘are determined by fixed rules and settled principles’ so that ‘decisions of legal policy, have no place in the law of property’. ¹⁰⁴ Stevens must therefore show that recognising cryptoassets as property raises strong questions of public policy. However, he does not do so to a sufficient degree; the policy issues he mentions are that cryptocurrencies facilitate criminal activity and that they are ‘enormously environmentally damaging’. ¹⁰⁵ Secondly, even assuming that these policy points constitute sufficient grounds to refuse cryptocurrency proprietary status, this would only apply to certain cryptocurrencies. In the case of environmental damage, only those networks that use the computation-intensive ‘proof of work’ consensus mechanism would be problematic, whereas many use newer processes ¹⁰⁶ that were designed to be more energy efficient and environmentally friendly. ¹⁰⁷ Likewise, the level to which cryptocurrency facilitates criminal activity depends on the specific cryptocurrency (as Stevens acknowledges in his article) and the immutability and transparency of blockchain ledgers are both features of cryptocurrency which would arguably be very unattractive to criminal enterprise. Indeed, the popularity of cryptocurrency in criminal circles will depend on the extent to which cryptocurrency is acknowledged, engaged with, and regulated by law. The solution then, as it was for bonds, shares, and financial securities more broadly, as well as cash, which were all at one stage, and to some extent still are, used for criminal enterprises, is regulation rather than prohibition or non-recognition.

Lastly, Stevens argues that, at best, the holder of a private key to a cryptowallet can be analogised to an individual who ‘know[s] where a pot of gold is buried’ – such a right may be valuable, but ‘if another independently of me discovers the same fact I have no claim against them. I have no right exigible erga omnes (or indeed, without more, against anyone) to information as such. It is for this reason that we do not refer to information, without more, as “property”’. ¹⁰⁸ The result is that if someone (who we will call A) were to correctly guess the private key and access an individual's cryptowallet, this would be ‘the equivalent of another independently discovering where gold is buried’ and thus ‘there would not without more be any claim’. ¹⁰⁹ However, we suggest that the holder of a private key (B) can be better analogised to a gold miner who has not only staked a claim to prospect a particular area ¹¹⁰ but is actually on the land and has begun to extract the gold. This is because possession of the private key enables total control over the cryptocurrency which the key ‘unlocks’, thereby allowing the cryptocurrencies to be accessed and transferred. This in turn communicates to the world, or at least to the rest of the crypto community, that x crypto asset is ‘yours’ with the result that they will treat you as its owner. ¹¹¹ A holder of a private key therefore has a lot more than mere knowledge of where a pot of gold is buried. Thus, if A not only achieves the statistical miracle of guessing the exact numbers and letters which make up a private key, but also achieves the further miracles of coincidentally determining that they form a private key and knowing which chain they correspond to, before going ahead and taking B's cryptoassets, the appropriate analogy cannot be someone who has merely independently discovered where gold is buried. A more fitting image would be one where A has torn out the flags staking out B's claim (either whilst B was away or by physically repelling B from the claim), sabotaged gold mining operations, and then proceeded to misappropriate the gold to themselves. It is obvious that in such circumstances, B should be able to bring a claim against A.

Cryptocurrencies as choses in possession and qualified property

As noted earlier, we suggest that cryptocurrencies are choses in possession, specifically a sub-set of them known as ‘qualified property’. Our argument will therefore proceed in two stages: first, we will show why cryptocurrencies are best understood as choses in possession; second, we will show why they are a form of qualified property.

Cryptocurrencies are choses in possession

We suggest that the answer to the conundrum of whether cryptocurrencies amount to property, and if so how, is to recognise the reality that digital assets are not ‘simply information’ ¹¹² but rather are ‘“things” in which a person [can have] an “interest”’ ¹¹³ given that they ‘can be transferred, stored, and, indeed, possessed’. ¹¹⁴ The simple reality is that, at the time when it was determined that intangibles could not be possessed, the human mind could not conceive of the technological developments that would allow for the existence of blockchain technology and cryptocurrencies which, whilst non-physical, nevertheless ‘do manifest themselves in the physical world, albeit in a way that humans are unable to perceive’ ¹¹⁵ and in a way which allows for rivalrousness (a quality physical things inherently possess). It is true that ‘[t]his physical manifestation at the level of digital bits and bytes is not permanent and changes with every transaction. Nonetheless, we identify what is going on as a particular digital token, somewhat like how we give a name to a particular river even though the water contained within its banks is constantly changing’. ¹¹⁶ The parallels with the concept of qualified property are extremely strong here as such water would be considered qualified property at common law. ¹¹⁷

Our argument has some similarities to that of Jessica Lai who suggests abandoning the difference between tangible and intangible things with regard to the tort of conversion. ¹¹⁸ Lai argues that ‘whether property can be converted or not should not depend on whether it is tangible (or tangible enough), as this is extremely difficult to legally delineate. This is because everything has some degree of tangibility or effect on tangible things ... [moreover] the development of technology might challenge our notions of tangibility more and more’. ¹¹⁹ Instead, Lai proposes a functional analysis so that ‘[i]f the property at issue has the characteristics of a public good (non-excludible and non-rivalrous), it should not be capable of conversion’ and vice-versa. ¹²⁰ Similarly, we argue that it is irrelevant whether or not a particular cryptocurrency is physical or possesses sufficient physicality – for example, via being on a hard drive; instead, the question is two-fold: 1) is it property? And, if so, 2), can it, functionally or from a common-sense perspective, be possessed? If the answer to both questions is yes, then it should fall into the category of choses in possession notwithstanding its lack of physicality.

How then do we establish whether x cryptocurrency is property and whether it can be possessed? As discussed above, we believe that there are no ‘fixed, universal, and exhaustive criteria’, ¹²¹ given that there are ‘degrees of property’, ¹²² and thus we do not – indeed, cannot – provide a clear bright-line test for answering these questions. However, we do believe that one can still formulate non-binding ‘guidelines’ which should not be applied mechanically but be sensitive to the nature of the ‘thing’, and which ought to be helpful in most situations. These guidelines are the, albeit maligned, ¹²³ test in Ainsworth which holds that in order to amount to property a thing must be ‘definable, identifiable by third parties, capable in its nature of assumption by third parties, and have some degree of permanence or stability’. ¹²⁴ We say these are only guidelines because not all property may meet every limb of the test, at least to the same degree; thus, for qualified property, it is clear that it will possess substantially less permanence or stability than absolute property, given that it is ‘such as is not in its nature permanent, but may sometimes subsist, and at other times not subsist’. ¹²⁵

As for what amounts to the ‘thing’ in the context of cryptocurrencies, we agree with the Law Commission that the relevant ‘thing’ is the ‘notional quantity unit manifested by the combination of the active operation of software by a network of participants and network instantiated data’. ¹²⁶ We also agree with the Law Commission that cryptocurrencies, in principle, meet the Ainsworth guidelines as they are definable ‘by reference to the rules of the system in which [they] exis[t]’ ¹²⁷ – the blockchain clearly permits their identification by third parties; it also similarly ensures their transferability. Moreover, even if they can only be owned for as long as they are possessed, they have some degree of permanence or stability which again is guaranteed by the blockchain.

However, as discussed in detail earlier, there is significant variety in the crypto world and thus certain – indeed, perhaps even many – cryptocurrencies may possess attributes which completely fail the Ainsworth guidelines. This is most obviously true of so-called ‘shitcoins’ ¹²⁸ where, for example, the developer retains backdoor access to the smart contract or blockchain protocols and, for example, seizes an individual's cryptocurrency, renders it unsellable/inalienable, or imposes significant taxes (such as diverting 90% of traded tokens to their own wallet before transferring the remaining 10%). ¹²⁹ Such cryptocurrencies would not amount to property because, although property is a sliding scale, they fall so short of the Ainsworth guidelines – especially the requirements of being capable in their nature of assumption by third parties and of being permanent and stable – that they are not ‘things’ in any sense and so cannot be property at all.

The question of whether cryptocurrencies can be possessed is even more complex because it has been said that ‘the word “possession” is one for which English law has never worked out a completely logical and exhaustive definition’. ¹³⁰ Nevertheless, we can say that, in general terms, ‘a man is said to possess or to be in possession of anything of which he has the apparent control, or from the use of which he has the apparent power of excluding others’. ¹³¹ Thus, possession ‘involves…two distinct elements, one of which is mental or subjective, the other physical or objective’. ¹³² Cryptocurrencies per se pose no particular issues regarding the mental element and thus, for our purposes, we can state that ‘[p]ossession is control’. ¹³³ The primary complexity of possession as regards cryptocurrencies arises from the fact that, in general, it has been thought that possession necessarily entails physical possession so that intangibles cannot be possessed.

Thus Lord Scarman stated in the case of R v Boyesen ¹³⁴ that possession ‘denotes a physical control or custody of a thing plus knowledge that you have it in your custody or control’. ¹³⁵ More recently, Lord Browne-Wilkinson stated in JA Pye (Oxford) Ltd v Graham ¹³⁶ that ‘there are two elements necessary for legal possession: (1) a sufficient degree of physical custody and control….; (2) an intention to exercise such custody and control on one's own behalf and for one's own benefit’. ¹³⁷ Most relevantly, as noted above, in OBG v Allan, ¹³⁸ the House of Lords appeared to rule out the possibility of extending the concept of possession to intangible things. ¹³⁹ The Court of Appeal built on this in Your Response ¹⁴⁰ and drew a clear distinction between ‘practical control’ and possession. It stated that ‘[p]ossession is concerned with the physical control of tangible objects; practical control is a broader concept, capable of extending to intangible assets and to things which the law would not regard as property at all’. ¹⁴¹ Although this appears to rule out the possibility of digital assets being possessed, we suggest that the Court of Appeal's holding is wrong both because it conflicts with general principles and because it conflicts with precedent in other areas of law.

Firstly, as a matter of principle, if possession is effectively about control, it is difficult to see why it should be limited to physical control. If it is possible to exercise the same level of control over a non-physical thing as a physical thing then it seems arbitrary, and even incoherent, to refuse to accept that the former can be possessed while acknowledging that the latter can. ¹⁴² Instead, one should adopt a more open-ended definition of possession such as that of Tony Honoré, who ‘famously described possession as “to have exclusive physical control of a thing, or to have such control as the nature of the thing admits.”’ ¹⁴³ Salmond is even more categorical in rejecting the idea that possession necessarily relates to the physical power to exclude, arguing instead that it ‘is the security of enjoyment, and there are other means of attaining this security than personal presence or power’. ¹⁴⁴ The end result is that

[i]n practice indeed possession is often said to be a social rather than a physical fact, in the sense that a person will be held to possess a thing if he has the sort and extent of control that society … would regard as appropriate to the kind of thing and the circumstances of the case. ¹⁴⁵

It was on the basis of such views, specifically those of Honoré, Green, and Randall, that the High Court of New Zealand held that one can ‘apply the concept of possession to digital assets… includ[ing] all forms of information stored digitally on an electronic device, such as emails, digital files, digital footage and computer programmes’. ¹⁴⁶

Secondly, the concept of possession has already been extended to cover digital files by English law in certain limited contexts. Thus, it is a criminal offence to possess indecent photographs of a child under s 160(1) of the Criminal Justice Act 1988 or to possess extreme pornographic images under s 63(1) of the Criminal Justice and Immigration Act 2008. In formulating the test for possession under these statutes, the Court of Appeal in Cyprian Okoro (No 3) v R ¹⁴⁷ devised a test almost identical to that of possession for physical things stating that ‘(1) the images must have been within the appellant's custody or control, i.e. so that he was capable of accessing them; and (2) he must have known that he possessed an image or a group of images’. ¹⁴⁸ Similarly, the Supreme Court of Canada held in R v Morelli ¹⁴⁹ that ‘to commit the offence of possession [of child pornography] … one must knowingly acquire the underlying data files and store them in a place under one's control’. ¹⁵⁰ Similar decisions have also been rendered in Australia and New Zealand, ¹⁵¹ with the latter also criminalising the ‘possession of objectionable materials’ more broadly. ¹⁵² The reality then is that the law already accepts that digital assets or files can be possessed in certain circumstances; the question is merely whether this must then be limited to statutory schemes existing in splendid isolation or can be extended by the common law to other digital assets, such as cryptocurrencies. We suggest that the latter option is the most coherent and principled.

Thirdly, the authorities we have discussed above, aside from Your Response, do not necessarily reject the idea of possessing non-physical things; they merely decline to consider it. This is because, as we have discussed above, the possibility of the creation of ‘digital things’, which could be controlled in a similar way to physical things, was not within their contemplation due to the state of technology at the time the case was heard. ¹⁵³ The obvious exception to this is OBG but we suggest that it should only be considered an authority for the view that conversion cannot be extended to choses in action.

In consequence, we suggest that there is no reason not to accept that cryptocurrencies can in principle be possessed. This is because holders of cryptocurrencies (excluding ‘shitcoins’ for the reasons outlined earlier) meet the requirements for possession as they exercise ‘custody or control’ over them given that ‘having a token deposited in one's wallet address means that [the holder] has positive and negative control’. ¹⁵⁴ The wallet holder has the positive ability to transfer their cryptocurrencies and ‘obtain any benefits associated with holding the coin’ ¹⁵⁵ as well as the negative ability ‘to exclude people from accessing and transferring the asset because they have a private key’. ¹⁵⁶ As this is arguably ‘the maximum degree of control allowed by the asset’ it can be considered the ‘digital equivalent of [the first limb of] “possession”.’ ¹⁵⁷ As for the second limb – the mental element – cryptocurrencies do not pose any unique issues in this regard and thus determining whether it is met will, as is the case in the law as a whole, have to be done on a case-by-case basis. We would therefore formulate the test for the possession of digital assets, and possession generally, as follows: an intention to possess, and such control as is necessary to secure enjoyment of the thing. ¹⁵⁸ This builds on the work of Honoré and Salmond, who recognise that not all things will be capable of the same level of control and that what really matters is that one possesses sufficient control, vis-à-vis others and society as a whole, ¹⁵⁹ to secure enjoyment of it. Having explained why cryptocurrencies should be considered choses in possession, we can now turn to an analysis of why we say that they are a form of ‘qualified property’ and what exactly we mean by this.

Cryptocurrencies are a form of qualified property

As noted earlier, English law divides all personal property into choses in possession and choses in action. Another division is often overlooked, however: that of ‘absolute’ and ‘qualified’ property. ¹⁶⁰ Blackstone describes ‘qualified, limited, or special property’ as being ‘such as is not in its nature permanent, but may sometimes subsist and other times not subsist’. ¹⁶¹ Its most well-known application is with regard to wild animals, ¹⁶² however, it can also apply to other ‘things’. Thus, Blackstone states that ‘[m]any other things may also be the objects of qualified property. It may subsist in the very elements of fire or light, of air, and of water’. ¹⁶³ This follows from the fact that

[a] man can have no absolute permanent property in these, as he may in the earth and land; since they are of a vague and fugitive nature, and therefore can admit of a precarious and qualified ownership, which lasts so long as they are in actual use and occupation, but no longer. ¹⁶⁴

This qualified property right allows an individual to bring an action against a person who ‘disturbs another and deprives him of the lawful enjoyment of these’ but ‘the property in them ceases the instant they are out of possession’. ¹⁶⁵ In this sense, qualified property is what can be called a possessory proprietary right; thus, in the same way that a thief or someone who finds a jewel, can, so long as they possess it, enforce their right of property ¹⁶⁶ over a stolen (or found) article against everyone but the true owner, ¹⁶⁷ so can an individual maintain their qualified property right in or over x against the entire world as long as they maintain possession of x. Moreover as ‘[p]ossession, in the relevant sense, is not merely evidence of absolute title [but] confers a title of its own, which is sometimes called a “possessory title” [which] is as good as the absolute title as against, it is usually said, every person except the absolute owner’, ¹⁶⁸ possessory proprietary rights are truly proprietary rights; they are not merely indications of, or the result of, legal presumptions about absolute ownership. ¹⁶⁹

A qualified property right can also arise not due to the nature of the thing but rather the nature of the person holding the right; this is the case, as Blackstone explains, in bailment and pledge. ¹⁷⁰ It has also been held by the English courts to apply to a bankrupt regarding ‘anything which he acquires after the bankruptcy’ ¹⁷¹ and has been used as a catch-all term to apply to any limited or special form of property. Thus, a sanitary authority's ownership in a public lavatory ¹⁷² and an owner of land's ownership in a drain ¹⁷³ were both held to be limited and thus qualified property.

The concept of qualified property is a concrete example of the crucial, but oft-overlooked, truth that property in the common law is not a homogeneous concept, ¹⁷⁴ with the result that attempting to define it ‘by reference to fixed, universal, and exhaustive criteria is delusive’. ¹⁷⁵ Instead, we should accept that there are ‘degrees of property’ ¹⁷⁶ or, perhaps more accurately, ‘degrees of thingness’. Consequently, as noted by Peter Turner, ‘property varies in strength’, with ‘strength’ being determined ‘according to how similar a given item of property is to a familiar tangible thing’. ¹⁷⁷ One could also consider property rights to be stronger, or weaker, depending on their similarity to ‘typically strong proprietary relations which humans have with respect to tangible things: ownership of land and ownership of goods’. ¹⁷⁸

Property is therefore a sliding scale with stronger property, such as land – as well as property rights over or in it, such as absolute ownership – at one end and weaker property such as wild animals or water – as well as property rights over or in them, such as qualified ownership or other possessory proprietary rights – at the other. The point has recently been affirmed in eloquent terms by the majority of the High Court of Australia in Hocking v Director General of the National Archives of Australia, ¹⁷⁹ where they noted that

[p]roperty is not ‘a monolithic notion of standard content and invariable intensity’; it is not ‘a term of art with one specific and precise meaning’. It is ‘a term that can be, and is, applied to many different kinds of relationship with a subject matter’… Accordingly, property is not for all purposes to be equated with ‘full beneficial, or absolute, ownership’. Indeed, a proprietary relationship can have the quality of relativity. ¹⁸⁰

The specific relevance of the concept of qualified property for cryptocurrencies is that many of their attributes are similar to those of things held to be qualified property, such as ferae naturae, air, and water. This is because once the private key (mere information – a string of letters and numbers), which allows access to one's cryptowallet, is out of one's ‘possession’, then the control, use, and enjoyment of the cryptocurrency is likewise gone. However, for as long as one holds the key then the cryptocoins are theirs, and even if their ownership was lost as a result of wrongful interference – for example, by hacking, theft, and so on – they could bring a variety of claims against such wrongdoers. Thus, for example, (and albeit in a statutory context) an individual who owned an eel weir in a river which managed to catch several eels but was subsequently interfered with so that the eels escaped, could sue for the loss of the eels. ¹⁸¹ Similarly, in the same way that an owner of doves could sue someone who shot them for trespass to goods, ¹⁸² a holder of cryptocoins could likewise sue an individual who misappropriated cryptocoins through hacking or otherwise wrongfully obtaining them, in trespass. Claims in conversion may also, in appropriate circumstances, be possible, although the law on the point appears to be unsettled. ¹⁸³

However, just as fish in a lake are no longer the qualified property of A if A's creditor, B, exercises their security over the land by selling it to C (even if the sale says nothing about the fish) who takes possession of it, cryptocoins would no longer be the qualified property of A if A, or his creditor B, sells his key to C, as it is this key which allows possession of the cryptocoins. ¹⁸⁴ It is unclear why the outcome should be any different if A merely loses his key, as in such circumstances A has lost possession – and, since qualified property is merely a possessory proprietary right, this necessitates loss of ownership on A's part. Moreover, because of the ‘sticky’ nature of the consensus mechanism, which we explore in further detail below, it is not possible for the software developers of a particular chain to release a software patch to allow for the return of misappropriated coins, even if the majority of individuals on a particular chain agree with such action. Thus, coins once lost are lost forever, and it is not possible to seek specific restitution or delivery of them, whether by detinue ¹⁸⁵ (where that cause of action still exists or is at least referred to) or conversion. ¹⁸⁶ The result is that, in general, whether the value of the coins has risen or fallen, their former owner will only be able to sue for their value as of the date of the detinue/conversion ¹⁸⁷ or trespass. ¹⁸⁸

Beyond these general circumstances, our analysis uniquely highlights and accounts for the importance of the distinction between custodial wallets (which operate like banks, whereby a user gives the custodian – often a crypto exchange – control over the private keys and access to their wallet in exchange for convenience and peace of mind) and non-custodial wallets (where the holder of cryptocoins maintains full control over and responsibility for their cryptowallet and their private key). ¹⁸⁹ In the case of a custodial wallet, our analysis means that as ‘the third party [the custodian] has full control over the cryptoassets’; ¹⁹⁰ through possession of the private key, they now own the cryptoassets with the coinholder merely having contractual claims against them for repayment of the cryptoassets on demand. Thus, custodial wallets operate like banks because just as, when depositing money into an account, a ‘customer [of a bank] does not acquire any interest in or charge over any asset of the bank or over all of the assets of the bank [but merely] acquires a chose in action, namely the right on request to payment by the bank of the whole or any part of the aggregate amount of principal and interest which has been credited or ought to be credited to the account’ ¹⁹¹ so too does the ‘customer’ of a custodial wallet.

The result is that as on the insolvency of a bank a customer ‘can only prove in the liquidation of the bank as an unsecured creditor for the amount which was, or ought to have been, credited to the account at the date when the bank went into liquidation’; ¹⁹² the same should be true on the insolvency of a custodial wallet provider. There is already US authority to this effect; ¹⁹³ however, in that case the court held that this was because the custodial wallet customers had explicitly signed over ownership. In our view this is unnecessary – merely giving possession and control of the assets to the custodial wallet holder is sufficient to transfer ownership in exchange for a contractual right to receive payment on demand of the cryptocurrency from the custodial wallet provider. It may be that in certain cases, if one can prove that the custodial wallet holder was holding the assets on trust for their customers, those customers of a custodial wallet could still possess proprietary rights over their coins, ¹⁹⁴ although as mere segregation of the funds is insufficient this is likely to require explicit provision in the contract between the customer and the custodial wallet provider. ¹⁹⁵ Admittedly, the idea of a trust over qualified property raises some unique issues, particularly as there does not appear to be any authority directly on the point; however, as other possessory proprietary rights can be held on trust – for example, a thief has been held to hold his possessory rights on trust for the true owner ¹⁹⁶ – there does not seem to be any good reason why qualified property could not also be held on trust. ¹⁹⁷

Our theory does not, therefore, preclude the creation of trusts over cryptoassets, whether in the context of custodial wallets or otherwise, but merely provides a helpful default rule. If a customer of a custodial wallet provider has handed possession of the cryptoassets over to the custodial wallet provider, whether de jure or de facto, via disclosure of the private key, then the custodial wallet holder is now the owner of the cryptoassets, and the customer merely holds a claim against them according to the terms of their agreement. ¹⁹⁸ There is nothing unusual or unfair in this, as it parallels the position customers of banks ordinarily find themselves in, and there is no reason why customers of custodial wallet holders, particularly given the well-known high-risk nature of crypto, should be in a better position than the customers of banks. Individuals who do not wish to trust custodial wallet holders can, in the same way that individuals who do not trust banks can, choose to retain custody themselves – while that may eliminate some risks, it will also create others, along with being generally less convenient. Thus, just as, in most countries, hiding money under the bed will not generally be safer or more convenient than paying it into a bank, neither will the use of a non-custodial wallet.

Finally, limiting the recognition of cryptocurrency as property in this way also captures the widely accepted sentiment in the crypto community that whoever holds the private key to x address owns any coins there, ¹⁹⁹ captured by the phrase ‘not your keys, not your coins’. ²⁰⁰ Access to the key provides complete and irreversible control. Similarly, losing the key (where this is not the result of wrongful interference by another) means falling short of the heightened responsibilities of self-custodianship ²⁰¹ and losing any right of complaint if one's coins are also lost and completely irretrievable; that is inherent in the system.

Admittedly, this understanding of cryptocurrencies has been challenged; thus in Tulip Trading v Bitcoin Association ²⁰² the English Court of Appeal held that it was at least arguable that software developers could, and should, release a software update on a blockchain to allow stolen coins to be returned to their rightful ‘owner’. ²⁰³ However, the problem with such views is that ‘developers can only propose changes to the protocol, whereas the implementation of Bitcoin Improvement Proposals (“BIPs”) requires a consensus to be reached by other participants, especially Bitcoin users’. ²⁰⁴ Thus, unless consensus can be achieved, it will be impossible to revise the blockchain and the general rule ‘not your keys, not your coins’ applies. Moreover, even if, as was the case following the 2016 hack of ETH, the majority of users agree to revise the chain in order to ‘undo’ the misappropriation of funds, if a significant minority refuse to do so, a ‘hard fork’ may occur. In consequence, even if on the ‘revised’ chain the hack or theft may be undone, on the ‘unrevised’ or ‘forked’ chain, the hacker or thief will retain the misappropriated funds. This, in fact, is exactly what happened on what became known as the ‘Ethereum Classic’ (ETC) chain after the 2016 DAO hack attack on ETH. ²⁰⁵

Moreover, relying on developer software updates directly undermines the unique facets of blockchain cryptocurrencies which allow for understanding them in a proprietary way. The more a chain is decentralised, the more robust the existence of an asset which cannot be double spent and is in fact rivalrous (since it is beyond any single user's or group's control), the more likely it is to fork in unpredictable ways if a software update is released. Our account is sensitive to these features of the technology, rather than expecting law to have unrealistic reach in being able to dictate outcomes to a hard-coded system. If a particular chain is no longer decentralised but rather concentrated in the hands of one or very few individuals, ²⁰⁶ this may undermine or, in extreme cases, even eliminate their rivalrousness, and that in turn will undermine or even destroy its proprietary nature (at least as regards those who do not hold the now centralised power on the chain). This is an inherent feature of the technology, which the law must therefore reflect. Likewise, it is the technological operation of cryptocurrencies themselves that necessitates their recognition as ‘qualified property’ because it is simply not possible to own them absolutely. ²⁰⁷ Instead, one can only own them to the extent that one exercises possession over them via holding their private key or, in other words, ‘your keys, your coins’ and vice-versa.

Conclusion

This paper provides a unique contribution to the extensive literature concerning cryptocurrencies in three main ways. Firstly, it offers an overview of the different types of cryptoassets which goes into more depth than other academic engagements with the topic, demonstrating how their diverse natures affect the extent to which they can, or cannot, be considered property. Secondly, it has outlined various underlying philosophical approaches taken to property law, including the ‘law as things’ approach which we argue best reflects the practice and history of common law property. Thirdly, it has outlined three alternative approaches to the proprietary status of cryptocurrencies and explained why, arguably, each of them falls short – for example, because they do not “gel” with the technological realities of cryptocurrencies, or because they cannot explain the practice of common law property law or are not compatible with its history and core structure. Fourthly, it has outlined a novel understanding of cryptocurrencies as a specific sub-category of choses in possession – namely, a qualified, or limited, form of property. This approach best reconciles the technological realities of cryptocurrencies with the practice and understanding of property in the common law. Specifically, we have argued that cryptocurrencies can amount to qualified property so long as they meet the Ainsworth guidelines and are capable of being sufficiently controlled to secure enjoyment of them. Determining whether a particular type of crypto asset, such as Bitcoin or a ‘shitcoin’, meets both these requirements has to be done on a case-by-case basis in order to respect both the ‘infinitely gradable quality’ ²⁰⁸ of property and the equally infinite technological diversity of cryptocurrencies.