REGULATING FINANCIAL INNOVATION: STARTING FROM SCRATCH

In this article Dr. Julian Roberts considers the benefits of going back to basics when regulating and litigating financial innovation.

KEY POINTS

The Law Commission’s Final Report (July 2023) sets out a welcome approach to the regulation of cryptoassets in the UK in that it seeks to identify the essential characteristics of cryptoassets by going back to basics rather than making “imperfect analogies” with, and shoehorning it into, existing law.

The English courts have also taken this approach – most recently the Court of Appeal in Tulip Trading v van der Laan [2023] EWCA Civ 83.

By contrast, the author posits that the regulatory and English court approaches to the treatment of derivative contracts have failed to break free from “imperfect analogies” with existing law; MiFID’s avoidance of systematic questions has, for example, allowed the courts to persist in treating derivatives contracts as no more than another form of sale.

INTRODUCTION

The UK is a major financial market. To retain that position in future, it must be able to accommodate and regulate innovation. That means not being fettered by imperfect analogies with existing law. It must, in other words, be capable of going back to basics when necessary.

There are two ways of approaching this.

In one approach, a comprehensive set of rules is set up and applied by a statutory body. Examples of this are the EU’s 2004 Markets in Financial Instruments Directive (MiFID), and the cryptoassets directive (MiCA) currently being implemented.

Another approach follows the common law, only codifying rules into statute once the courts – guided by academics and experts – have gathered enough hard results. The Sale of Goods Act of 1893 was a classic example. The Law Commission’s Digital Assets project is currently attempting something similar for crypto.

This article looks at two forms of innovative finance – derivatives and cryptoassets – and considers how successful the corresponding regulation has been. Derivatives were a prominent concern of rulemaking at EU level, yielding MiFID; cryptoassets have been a focus of the Law Commission and of a series of decisions in the English courts.

DERIVATIVES

Zero-sum deals are not sales

On their face, derivative deals (which feature prominently in MiFID – see Annex I, s C) typically look like some kind of exchange. For example, in a straightforward interest rate swap, one party pays a fixed rate of interest, while the other pays variable. Hence the term “swap”. Indeed, such transactions are often described as a “sale”.

However, this is misleading. First, in a swap, there is never any exchange. By virtue of “netting”, the parties’ payments are instantly set off against each other. Whenever a payment takes place, it only moves in one direction, depending, in a plain interest rate swap, on whether the “variable” rate is lower or higher than the fixed rate. Unlike in an exchange, where both parties increase their utility, in a swap only one party does. The other simply loses. This is a zero-sum transaction.

Second, the payments are not continuous (as is suggested by the reference to interest rates), but triggered by single events. In a typical interest rate swap, the variable rate of interest current on an agreed day is used to calculate who pays what for that period. This procedure is repeated as a series, every six months, for as long as the contract lasts (which may be very long – see the currently contentious LOBO (Lender Option Borrower Option) swaps).

Risk contracts

A swap does not function as a sale. It is, in reality, a risk contract, or series of risk contracts, whose outcome depends on a series of uncertain events extending into the future.

Risk contracts as such are not in themselves novel or unprecedented. The basic structure of a risk contract is familiar from betting. Such contracts may be harmless or even beneficial, as in the case of insurance, where the insured, by betting that some harm (for example, a fire) will take place, wins his bet (and hence compensation) if this indeed happens. Swaps can be used for this purpose too, for example to hedge commodities, FX and rates.

The problem with using swap contracts to hedge market risks is that a faulty hedge can massively increase risk. If, for example, the interest rate on a loan changes without corresponding changes in the “hedge”, the latter can start to function as a burden, not a prop. Bizarrely, this is especially likely when the rates on the loan are moving favourably, ie downwards. This is because falling rates cause the swap’s negative mark-to-market (MTM) (for the borrower, it is a bet on rising rates) to increase. That, in turn, breaches covenants in the loan, which allows the lender to restructure, and to do that on terms which no longer match the hedge. Result: default, despite the fact that by normal commercial standards the borrower’s business may be going well.

Regulatory response to derivatives problems

There is nothing intrinsically wrong with derivatives such as interest rate swaps. On the contrary, in a modern economy they are essential for managing commercial risks. Nor is there anything reprobate in market actors using them for speculation; indeed, this serves to inject liquidity into the market. The difference from the older form of risk contract lies in the fact that derivatives rely on computational resources which have only become available since the 1990s. These make it possible to set market prices, but they are inaccessible to anyone without a (prohibitively expensive) data terminal. Consequently banks can structure swaps containing an upfront profit which, for retail customers, is invisible. Banks have been tempted to conceal this profit even though – given the zero-sum structure – the potential for a conflict of interest is substantial.

This problem is obscured by the vocabulary of “sale”. In contracts of sale, there is no automatic duty of disclosure – caveat emptor. This is different in risk contracts. In principle, it is up to the parties to agree specifically and individually on the entire content of the risk, and that includes a duty of disclosure. In insurance, this is the principle of uberrima fides – utmost good faith.1 In betting, it is the requirement of “equal knowledge or equal ignorance” between the parties, as enunciated by Lord Mansfield in Jones v Randall (1774).

Given that the “bets” in derivatives are not single but extend over a potentially long series, the same principle should apply a fortiori. It is essential for both parties to know the current market price, not only at the outset, but as long as the series continues. Indeed MiFID (and thus COBS – Conduct of Business Sourcebook) from the outset required firms to warn customers as soon as the mark-to-market value of any “contingent liability transactions” (which includes swaps) fell substantially:

“Investment firms that hold a retail client account that includes positions in leveraged financial instruments or contingent liability transactions shall inform the client, where the initial value of each instrument depreciates by 10% and thereafter at multiples of 10%.” (COBS 16A.4.3)

Information of this kind, had it been provided, would have enabled borrowers to terminate falling swaps in time. However, due to regulations disallowing corporates (and hence many small businesses) from invoking COBS, and due also to the fact that this provision is buried deep in subsidiary rules, it has never been invoked in litigation.

Given the lack of any wider discussion, the courts have largely stuck with the notion that derivatives are sales, and resisted the obvious affinity with risk contracts.2 When a decade ago the “hedging products” scandal broke, regulators confined themselves to making banks provide a measure of compensation.3 This was a purely local instrument (the provision of redress) with no analytical basis.

Overall, regulation of derivatives, both in England and in the EU (which had plenty of its own derivatives scandals), never broke free from the “imperfect analogies” warned against by the Law Commission.4 Fortunately, as we shall now see, the situation with crypto is much better.

CRYPTOASSETS

Cryptoassets as property

Cryptoassets have been criticised as “the largest pyramid scheme ever yet dreamed up” (Stevens, 2023). They are puzzling because there is no obvious reason for their having any value. High trading costs and execution delays make them unsuited to everyday use, and they are beset by scammers. So what exactly – in law, or otherwise – are they? In particular, are they property?

The creator of Bitcoin (which is still the paradigmatic cryptocurrency) started from a claim that traditional electronic payments cannot guarantee immutability. Electronic payments are, in principle, reversible. If disputes arise, mediators must intervene to sort them out, and that creates a need for personal trust:

“With the possibility of reversal, the need for trust spreads. Merchants must be wary of their customers, hassling them for more information than they would otherwise need. A certain percentage of fraud is accepted as unavoidable.”5

This can be avoided by using physical currency, in which possession on its own establishes and concludes the transaction. But physical currency has disadvantages in terms of security and location which in many situations greatly outweigh the advantages. The disadvantage of traditional electronic transfers, on the other hand, is that they rely on the law of obligations as a condition of their effectiveness. If you want to match the advantages of a classic physical currency such as gold, while retaining the speed and convenience of electronic transfer, you have to be able to create, electronically, an object of the transaction which is independent of any personal obligations surrounding it. There may or may not be issues of a contractual or tortious nature surrounding that act, but the goal is to make the act itself (as in handing over gold) conclusive and effective without regard to those surrounding issues.

So, if Bitcoin can, by electronic means, collapse the act of payment into transfer of possession alone, then it will combine the advantages of both electronic and physical currency.

MACHINES WRITE THE LEDGER

Interfering third parties (“trusted intermediaries”) are, in the Bitcoin account, what makes traditional electronic payment systems slow and insecure. How to get rid of them?

The answer is decentralised finance (DeFi). DeFi gets rid of centralised servers in banks and other large institutions, and instead distributes financial record keeping among an open network of participants whose individual computers all mirror the same financial “ledger”. By means of DeFi, the Bitcoin ledger, in this example, is repeated thousands of time in computers all over the world. This is the first bulwark against attempts to manipulate the “ledger”.

The second is a system to ensure that each additional entry in the “ledger” is valid, and, once validated, irreversible. This is achieved by cryptographic techniques. One ensures that there is only one unique way to draw on ledger entries. This is the “private key”, which gives its possessor exclusive power to transfer any bitcoins present at a line of the “ledger”. The private key is sufficient and necessary to access and move the bitcoins. Without it, the bitcoins at that line of the “ledger” are lost, for ever.

The second cryptographic technique ensures that transactions in the ledger (for example, when the possessor of the private key transfers her bitcoins to a payee’s address) are validated and set immutably. This is achieved by taking all the transactions that have taken place within a set interval (ten minutes) and putting them into a “block”. The participating computers are then invited to find the unique cryptographic key (a “hash”) which corresponds to the totality of the data in the block. This can only be done by “brute force”, ie by trying out millions of answers until the right one is found. Once found, the block’s hash is then incorporated into the next block of transactions. This hash then becomes part of the unique hash of every subsequent block, meaning that nothing can be altered retrospectively without altering the entire chain (rewriting the “ledger”) back to that point.

Since the process of finding hashes takes time and a great deal of computing power, the chain (the ledger) is resistant to any attempt to interfere with its onward growth. Manipulation would typically take the form of modifying a past transaction, for example by converting a payment to address A into a payment to address B.

According to Bitcoin, this will be prevented by the mechanics of the system. Specifically, the attacker would need to go back and recalculate a new hash for the block containing the transaction he wants to modify, and then continue this procedure for all subsequent blocks. But all this time the legitimate computers on the network will be continually calculating hashes for later blocks on the basis of the original, unmanipulated transaction.

It is therefore unlikely, or impossible, that the attacker will be able to calculate new, illegitimate hashes quickly enough ever to catch up.

This applies, at least, as long as the attacker has less computing power – fewer CPUs (central processing units) – than the legitimate users. With 50% or more of the CPUs, the legitimate chain will always be ahead of, and longer than, the attacker’s chain. This system of validation is called “proof of work”, which can be conceived of as majority rule by machines instead of humans:

“Proof-of-work is essentially one-CPU-one-vote. The majority decision is represented by the longest chain, which has the greatest proof-of-work effort invested in it. If a majority of CPU power is controlled by honest nodes, the honest chain will grow the fastest and outpace any competing chains. To modify a past block, an attacker would have to redo the proof-of-work of the block and all blocks after it and then catch up with and surpass the work of the honest nodes.” (Nakamoto)

Thus, the bitcoin ledger, once set in motion, processes incoming transactions automatically and without human intervention. This makes the blockchain inert and independent – just like physical currency, and just like property.

CRYPTOASSETS AS PROPERTY;

THE LAW COMMISSION CRITERIA

Despite some academic dissent, there is wide acceptance in the courts that Bitcoin is property.

Under English law, the difficulty with treating cryptocurrencies as property is that they do not belong to either of its two recognised forms, namely: (i) things in possession; and (ii) choses in action. This is because they are intangible (which excludes (i)), and, by intention at least, they are not claims against persons (which rules out (ii)).

For this reason, the Law Commission has suggested that cryptoassets can and should be recognised as belonging to a third category, with the following characterics:6

(i) it is composed of data represented in an electronic medium, including in the form of computer code, electronic, digital or analogue signals;

(ii) it exists independently of persons and exists independently of the legal system; and

(iii) it is rivalrous, ie 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.

These suggestions have been useful in a number of court decisions dealing with the question of how to deal with crypto scams whose perpetrators are unknown. One of these is Tulip Trading v van der Laan [2023] EWCA Civ 83.

Tulip Trading and the software developers

The achievement claimed by Bitcoin is to produce a payment token which, in Law Commission terminology, is not only rivalrous, but also self-administering and independent of human manipulation.

This has been challenged as a “myth”, for two reasons. First, the blockchain is, indisputably, serviced by ongoing updates, bug fixes and even major “forks” which are merged into the software. Second, these updates do not appear spontaneously in the net community but are implemented by a relatively small group of select individuals who, in the last resort, wield great power. These individuals, in the upshot, constitute the “trusted third party” that Bitcoin allegedly does without. Indeed, claims one influential commentator, they are in law fiduciaries and owe fiduciary duties to the holders of Bitcoin (Walch 2019).

This argument has been taken up by the Court of Appeal in Tulip Trading v van der Laan.

In Tulip, the claimant says it owns approximately 110,000 tokens located at two blockchain addresses. These include Bitcoins (BTC) worth over £4bn. The files in which the claimant stored its private keys were hacked in February 2020, with the result that it can no longer access its bitcoin. Nor, it would seem, can the hackers: the bitcoin is still lying untouched at the two addresses.

The identity of the attackers is unknown. However, in Tulip the claimant is suing a number of individuals supposedly involved in developing software for the bitcoin system. The claimant says: (i) it is within the power of the developers to allow a person who no longer has access to his private keys nonetheless to access his Bitcoin; and (ii) the court can and should tell the developers to do so.

The Court of Appeal in Tulip confirms that the Bitcoin is to be treated as property, and indeed this is the basis for its finding that the developers may be fiduciaries. The court recognises that Bitcoin, though intangible, “exists outside the minds of individuals” and to that degree “exists”. However, it also infers that this existence is a matter of software which is controlled by developers.

“Bitcoin ... has properties which exist outside the minds of individuals, but those properties only exist inside computers as a consequence of the bitcoin software. There is nothing else. And crucially, asserts Tulip, it is the developers who control this software.” (per Birss LJ at 72)

The private key as independent property

The argument that individuals developing Bitcoin software are fiduciaries seems problematic and indeed failed at first instance. As the defendant developers in Tulip Trading have countered, Bitcoin software is open source. It is stored on GitHub, which is a public repository; in principle anyone can suggest modifications and bugfixes, and the choice of which ones to adopt is “democratic”, dependent on whether a majority of users decides to follow. There is no mechanism for forcing a change against the wishes of the community. There is no individual, “trusted” or otherwise, who can be enjoined to take any particular action. (Even were they in a position to do so, it is unclear what action developers could take in the Tulip case without, in effect, destroying the chain, and, in consequence, the value intended to be recovered.7)

Beyond this, the claimant’s argument in Tulip Trading is based on a questionable analogy. The claimant asserts that private keys to Bitcoin addresses are comparable to the keys to a safe. As keys, they are, it is claimed, distinct from the value contained in the safe and, in the last resort, incidental to the property itself. Hence interfering with the keys (by, as it were, breaking open the lock to recover the coins stuck inside) does not interfere with what we are really interested in, namely the property itself.

This argument, if valid, would negate the initiative Bitcoin and other cryptocurrencies are based upon. Private keys are not analogous to the keys to the safe. A Bitcoin private key is not a means to possessing something else but is itself all that there is to possess. The blockchain address can be derived from the private key; and, once the address is established, any credit balance is publicly visible. Thus, should you stumble across a private key lying in the road (noted on a piece of paper, perhaps), you by that fact alone possess the property. It is exhaustive of the property at issue – there is literally nothing else to possess apart from the private key. So, odd though it may seem at first glance, knowledge of a private key is possession of property, analogous not to possession of a safe key, but to possession of the coins themselves.

For the Court of Appeal in Tulip, the attraction of looking to the developers as fiduciaries lies in the fact that the claim becomes more readily justiciable. As Birss LJ comments, the court’s decision in the Tulip case will, at best, “only be in personam and not in rem, but that is true in any or at least most property disputes” (at 83). The difficulty with tackling the developers as originators of the software, however, is that it undermines the conclusion that cryptoassets are property. Targeting the developers, it is submitted, needs to be treated with caution.

Note: the ontology of cryptoassets

The question of what exactly a cryptoasset is (its ontology) has exercised academic commentators. Critics assert that it is merely “information” (a line of numbers and letters indicating an address), which does not have real being at all (Stevens, 2023). Against this, it is argued that, though “ideational”, currency quite properly invokes something which goes beyond its merely physical being (Fox, 2022).

Is this a problem?

We think we are familiar with markets. But in fact, there is much that is mysterious about them. We invest objects with a value which has little bearing on their usefulness, their scarcity, or on any other objective property. This applies to tokens like gold. It also applies to commodities – ie to any item whose price is set not by use value or even scarcity, but by “the market”.

Marx famously called this “fetishism”.

A fetish is a magical property ascribed to an otherwise everyday object. This sounds pejorative. But in fact, it is a commonplace of philosophical analysis. Adam Smith targeted something similar when he described the “hidden hand” governing markets. Hegel addressed it in his theory of “second nature”. Institutions, customs, and for that matter markets and prices, are human constructions, but no less real for that. We quite properly treat them as natural and inevitable, for without them the ingenuity of human life would be impossible.

It should not alarm us that cryptoassets are “ideational”. The market is essentially ideational, but that has not stopped it from becoming one of the most effective of all human creations. There is no point in denying the reality of such entities and institutions. Far better to recognise what they essentially are – their ontology – and then to allow rules to emerge on that basis.

CONCLUSION

The technologies for both derivatives and crypto did not exist 50 years ago. The innovations that have subsequently emerged raise fundamental questions: what are these novel entities? Can they be assimilated to fit existing rules, or should legal analysis start at a more systematic level?

As far as derivatives were concerned, the response was left to the statutory regulators, who produced MiFID. Despite its length, this massive compendium eschewed systematic questions entirely. In practice, it turned out to be a toothless monster, as banks throughout the EU gleefully sold toxic swaps to local authorities, SMEs and other ill-equipped clients.

One may have similar reservations today about the EU’s MiCA initiative. This imports volumes of COBS-style rules for traders, but, as far as crypto itself is concerned, confines itself to stablecoins and other peripheral issues.

By contrast, in the UK, the Law Commission’s “common law” approach to crypto takes detailed account of the interactions between academic debate and court decisions. The “Final Report” which appeared in July engages with the new technology in systematic and powerful ways, and is to be welcomed.

References:

Fox, David (2022), ‘Digital Assets as Transactional Power’, JIBFL 37, p 1.

Stevens, Robert (2023), ‘Crypto is not Property’, LQR 11 April.

Walch, Angela (2019), ‘In coders we trust: software developers as fiduciaries in Public Blockchains’, in Hacker (ed.) Regulating Blockchain.

1 “In policies of insurance, whether marine insurance or life insurance, there is an understanding that the contract is uberrima fides, that if you know any circumstance at all that may influence the underwriter’s opinion as to the risk he is incurring, and consequently as to whether he will take it, or what premium he will charge if he does take it, you will state what you know.” Brownlie v Campbell, 925 (1880) 5 AC 954 (HL 1880) per Lord Blackburn.

2 WW Property Investments Ltd v NatWest [2016] EWCA Civ 1142 (Briggs and Christopher Clarke LJJ 29 November 2016).

3 FSA, Interest Rate Hedging Products – Pilot Findings, 31 January 2013.

4 Law Commission, Digital Assets: Final Report, 28 June 2023.

5 Satoshi Nakamoto, ‘Bitcoin: A Peer-to-Peer Electronic Cash System’, October 2008.

6 Law Commission, Digital Assets: Final Report, 4.5.

7 Norton Rose Fulbright, ‘Tulip v van Der Laan: The Future of Cryptocurrency in the United Kingdom’ (Norton Rose Fulbright, 16 May 2023).

Julian Roberts is a barrister practising from Temple Square Chambers. He appeared for the claimant in WW Properties v NatWest.

Email: jroberts@temple-consulting.co.uk; jroberts@templesquarechambers.com