Faculty of law blogs / UNIVERSITY OF OXFORD

Who Actually Owns the Crypto-assets Held at a Crypto-asset Intermediary?


Time to read

4 Minutes


Henry Chong
CEO, Fusang Corp
Louise Gullifer QC (Hon)
Rouse Ball Professor of English Law, University of Cambridge

Most crypto-assets are not held directly: they are either held using commercial software or, particularly in relation to high value holdings, they are held in some sort of custodial arrangement. In our recent paper, we analyse the possible legal relationships that may exist between a crypto-asset intermediary and clients that hold crypto-assets with it. In our view, there are four possible legal relationships (trust, outright title transfer, quasi-bailment, and mere contract). Which is the correct analysis in a particular case will depend on the interpretation of the agreement between the parties. We argue that, where the common law applies to a custodial relationship it is most likely to be one of trust.

The most important consequence of the intermediary holding the asset on trust for the client is in terms of investor protection in the event of the intermediary’s insolvency, since the client’s assets would be unavailable to other creditors of the intermediary (as was held in the recent New Zealand case of Cryptopia). A trust relationship also imposes a default set of obligations on the trustee, obliging it to use the asset for the best interests of the beneficiary. Trustees’ duties can be modified to a considerable extent. A trustee can have a right of use and rehypothecation (but cannot have total free use: that would be an outright title transfer), and can reduce obligations to a minimum as the Citibank v MBIA case shows.

An alternative to a trust analysis is one of outright title transfer. Here, the intermediary can freely dispose of the assets it holds as it is the absolute owner, and is not under a duty to use the asset for the benefit of the investor. This carries consequences in terms of insolvency and tax treatment. On the intermediary’s insolvency, the client has no entitlement to the asset. Further, it is the intermediary on whom tax (eg capital gains tax) will be levied.  Given these consequences, we argue that it is unlikely that the parties will have intended title transfer, and this borne out by the wording in most client/intermediary agreements, which is usually ‘give us the asset, we will hold it on your behalf, and we promise that it is still your asset’.

Another possible legal relationship is one of ‘quasi-bailment’, which is essentially bailment with a caveat. Under the current common law in most jurisdictions it is impossible to create a bailment of an intangible asset, such as a crypto-asset. We use the term ‘quasi-bailment’ to reflect the fact that the principles relating to possession are applied by analogy with physical, tangible assets. Quasi-bailment is created by transfer of ‘quasi-possession’, by which we mean a degree of control of a crypto-asset which would amount to possession of a physical asset. If a quasi-bailment relationship is created, this would impose a duty on the bailee to take reasonable care of assets. We argue that a trust analysis is much more likely to apply than one of bailment. The parties are likely to have intended an arrangement that would provide commercial certainty, and the trust provides this certainty to a much greater extent. The English law trust has developed through centuries of case law, resulting in a concept which contains relatively well-defined boundaries, and yet is extremely flexible. This stands in stark contrast with bailment.  Not only is it very unclear whether bailment (or quasi-bailment) can apply to intangible assets, but the outer boundaries of the concept are far from settled, for instance in the context of a right of rehypothecation.

The final possibility is that the legal relationship between the client and intermediary will only sound in contract. This would occur where the intermediary is merely a ‘technology provider’ (such as Amazon Web Services), as opposed to a custodian. In addition to express contractual obligations set out in the investor-intermediary agreement, there would also be a duty of care and skill, which is implied in a contact for services. However, this arrangement does not create any property rights in the asset in favour of the intermediary. This analysis is unlikely to apply to a custodial relationship, where legal title and/or quasi-possession will be with the intermediary.

So why are crypto-asset intermediaries so reluctant to admit that there is a trust relationship, and either stay totally silent on the legal nature of their relationship, or claim that they have a mere contractual relationship (in which they disclaim any and all liability and duties to their client)? While trust might be the natural conclusion, it will also entail a number of regulatory consequences, quite apart from the legal duties that it implies. For example, most common law jurisdictions (with the notable exception of the UK), require a trustee acting in a professional capacity (as a repeated business, for financial consideration) to be licensed, irrespective of the nature of the underlying asset. Being licensed as a trust company and financial institution will force these intermediaries to maintain KYC/AML standards, which are impossible for many crypto-asset intermediaries to meet: most have no idea who many of their customers are, and where the money has flowed from. But if the crypto-asset industry wants to legitimise, becoming professional trustees and starting to genuinely operate in the best interests of their investors is an inevitability.

The precise legal characterisation of any consensual relationship depends on the interpretation of the agreement between the parties. Our paper identifies the criteria which inform this categorisation, as well as the consequences of each possible result. Our conclusion that the most likely analysis is one of trust chimes with the existing (sparse) case law, and also offers a route for intermediaries to bring themselves within the regulatory net and signal their legitimacy to potential clients.


Hin Liu is Legal and Business Associate, Fusang Corp and Lecturer of Law, University of Oxford.

Henry Chong is the CEO of Fusang Corp.

Louise Gullifer is the Rouse Ball Professor of English Law, Cambridge University.


With the support of