If banks want things to change, says the RLN, things will have to stay as they are
- Future of Finance
- May 1, 2024
- 53 min read
Updated: Jan 29
A Future of Finance interview with Tony Mclaughlin, Emerging Payments and Business Development at Citi
The Regulated Liability Network (RLN) embodies an idea of the future of money that, unlike most conceptual novelties in the field, has become more voguish rather than less since it was first unveiled in a white paper of November 2022. In fact, the RLN can lay claim to have pioneered an approach to scaling the tokenisation of assets that has captured the interest of supranationals and central banks. The white paper may have coincided with the International Monetary Fund (IMF) advancing the idea of an “X-C platform”11 but it appeared months before the Bank for International Settlements (BIS) outlined its notion of a “unified ledger” or “single programmable platform”2 and the Monetary Authority of Singapore (MAS) announced it was working with four banks on Global Layer One (GL1), an open digital infrastructure to host tokenised financial assets and applications. But it would be a mistake to label the RLN as avant-garde. It is based in a sound understanding of the classic theory of computation and aims unashamedly to preserve fiat currencies and their twin variants of commercial and central bank money as the foundations of the financial systems of the future. Its design for a common settlement infrastructure for tokenised money also bears an uncanny resemblance to the way payments are settled today, in terms of intermediation as well as technique. Which is why RLN might just be adopted widely once banks understand its design. Dominic Hobson, co-founder of Future of Finance, spoke to Tony McLaughlin, Managing Director, Emerging Payments and Business Development at Citi Treasury and Trade Solutions, and one of the 11 industry leaders that contributed to the development of the original idea of the RLN.
International Monetary Fund, A Multi-Currency Exchange and Contracting Platform, prepared by Tobias Adrian, Federico Grinberg, Tommaso Mancini-Griffoli, Robert M. Townsend, and Nicolas Zhan, Working Paper WP/22/217, November 2022. ↩︎
BIS Annual Economic Report, 20 June 2023, Chapter III, Blueprint for the future monetary system: improving the old, enabling the new, pages 85-118. ↩︎
Key Insights
The Regulated Liability Network (RLN) is a design for a new financial market infrastructure (FMI) capable of settling transactions in regulated fiat currency monies such as Central Bank Digital Currencies (CBDCs), Stablecoins, tokenised deposits and e-money but also a full range of other financial assets, including securities and money market instruments.
The purpose of the RLN is not to facilitate the Balkanisation of money as advocated by cryptocurrency enthusiasts but to update sovereign fiat currencies to meet the needs of e-commerce, and so preserve the singleness of money on grounds that a migration away from currencies backed by the credit of nation-states will be a retrograde step.
As the name of the RLN suggests, its usage will be restricted to financial institutions that have liabilities to their customers which they are obliged by law and regulation to discharge on request, which is a definition broad enough to include 100 per cent asset backed Stablecoins redeemable at par as well as obvious instances such as deposits.
The authors of the RLN believe that the key to the success of tokenised monies and other financial assets is adoption and that adoption depends not on offering a compelling proprietary design but on securing a consensus among central banks and regulated banks and payment service providers (PSPs) on the necessary infrastructure.
The consensus must be built on agreement that the proposed design of the RLN, including its probable use of blockchain technology, delivers benefits that cannot be achieved with existing infrastructure, which is why the idea is now moving from a positive discovery phase into a second phase of concrete experimentation.
Blockchain is well-suited to support an RLN because it is a Turing-complete (applicable to any use-case) computer that, like any computer, functions as a state machine (a means of storing the status of something at a particular time), enabling it to provide users of a settlement FMI with a clear picture of `who owns what’ at any time.
In technical terms, the RLN will settle tokenised transactions in much the same way as they are settled today – namely, by the bank of the payer “burning” a liability to the payer and the bank of the payee “minting” a liability to the payee, through the transfer of liabilities of the central bank from the bank of the payer to the bank of the payee.
The practical effect of the operation of the RLN is to update the balance sheets of the bank of the payer (reducing its liabilities to the payer), the bank of the payee (increasing its liabilities to the payee) and the central bank (reducing its liabilities to the bank of the payer and increasing its liability to the bank of the payee) with legal settlement finality.
Achieving settlement finality is a crucial goal of the RLN because it is a legal concept and not a technological outcome, which means transactions settled via the RLN will take place in a contractual and regulated environment fully comparable to those delivered by existing settlement FMIs such as RTGSs and CSDs.
The RLN model works as well in cross-border payments as it does in domestic payments, the only material difference being that five bank balance sheets (the bank of the payer and the bank of the payee will each make use of a correspondent bank in the relevant currency) must be updated rather than three.
The RLN vision is aligned with Project Agora, an initiative of the Bank for International Settlements (BIS) in which seven central banks are exploring how a tokenised money FMI operating a common state ledger could be used to fulfil the Group of Seven (G7) mandate to make cross-border payments faster, cheaper, more transparent and more accessible.
The potential benefits of RLN for domestic consumers include a transition from bi-lateral, sequential payments channels to synonymous, multi-lateral payments channels in which all parties to a transaction are paid at the same time, and the ability to assign money to individuals or apps to spend at their discretion to fulfil particular needs or duties.
Settling transactions by updating claims on the balance sheets of banks gives the initial version of RLN a bias against `native’ tokens because claims on bank balance sheets belong to the customers of the banks and cannot be transferred to another bank as if they are bearer instruments, unlike tokenised negotiable instruments such as security tokens.
A second effect of settling transactions by updating balance sheets is to mitigate one of the principal advantages associated with blockchain technology – namely, reducing the need for reconciliation of different accounts of the same transaction – because claims settled via the RLN must still be reflected on the balance sheets of the users.
An implication of settling transactions by updating balance sheets is to preserve those balance sheets as forms of intermediation between payers and payers and buyers and sellers, and the RLN preserves them not merely as a matter of technical convenience but because its designers believe users will prefer to work through trusted intermediaries.
A fourth effect of settling transactions by updating balance sheets is to preserve the value of structured messaging protocols such as SWIFT for allocation of omnibus claims between individual invoices and bank accounts and to facilitate inter-operability between blockchain networks and between blockchain networks and traditional networks.
In the same way that blockchains do no force users to choose `atomic’ settlement, the RLN model permits the netting of payments liabilities (through Automated Clearing Houses, or ACHs) and securities liabilities (through Central Counterparty Clearing houses, or CCPs) ahead of settlement as an option for users.
The preservation of intermediaries and messages reflect the RLN insistence on an economical model in which the FMI delivers a common state of `who owns what’ only, leaving services such as invoice and investment allocations and asset servicing to other service providers rather than, say, `smart contracts’ in the tokens or on the same ledger.
In the current experimentation stage of the RLN, the simultaneous updating of balance sheets to arrive at a common state of `who owns what’ takes place within a single technical environment in which banks operate “nodes,” but future iterations of the model might permit interoperability between individual tokenisation engines or multiple RLNs.
One possibility is that the RLN becomes a virtual Layer One blockchain, akin to those being developed by the Monetary Authority of Singapore (MAS) and Digital Asset (the Canton Network), that provides a foundational platform for transactions but permits banks and non-banks to develop applications to attract customers to the network.
The RLN is likely to operate as a private permissioned network for two practical reasons, of which the first is the regulatory obligation on banks to perform risk management due diligence on any external systems to which they outsource functions, and the second is the regulatory obligation to avoid doing even inadvertent business with criminals.
A private and permissioned network is not synonymous with a closed proprietary network since the RLN is intended to overcome the asset class silos that plague payments and capital markets settlements today by providing an open, multi-asset class, general-purpose venue for settlements akin to the `unified’ ledgers described by the BIS and IMF.
One model for the future governance of the RLN, in terms of membership and rulebook, is CLS, a systemically important public/private sector FMI that settles cross-currency payments between banks in 18 currencies and which is now structured as a mutual entity owned by 79 banks under the regulatory supervision of the Federal Reserve Bank.
No date is set for the RLN to go into production, and none will be set until the current series of experiments has proved the benefits of building a new FMI will outweigh the costs and, more importantly, a consensus has formed among financial institutions that the benefits are great enough to warrant adoption.
Transcript
00:14 Dominic Hobson: Hello, I’m Dominic Hobson, Co-founder of the Future of Finance. My guest today is Tony McLaughlin, who’s responsible for emerging payments and business development at Citi, a capacity in which he works with governments, financial institutions, multinational companies, established technology vendors and FinTechs on useful innovations in payments. Chief among the innovations he has fathered is a new model for resolving the contest between the different forms of digital money: the Regulated Liability Network, or RLN. Now, the RLN has attracted interest not just in the payments industry, but in the securities and fund industries as well, and indeed from regulators, including the Bank for International Settlements (BIS) and the International Monetary Fund (IMF). So not surprisingly, it’s the RLN which is the subject of our conversation today. Tony, thanks very much for taking time to talk to us.
01:07 Tony McLaughlin: It’s a pleasure. Nice to be here.
01:09 Dominic Hobson: I’ve described you as the father of the RLN, but do tell us, I’m sure there’s a longer version here. Do tell us that long version. What is the origin and history of the Regulated Liability Network?
01:21 Tony McLaughlin: Well, I’ve been working in payments for the past 30 years, and for a lot of that time, payments wasn’t the most interesting subject for the broader population. But things certainly spiced up with the development of cryptocurrencies and then accelerated with proposals like Libra and CBDC [Central Bank Digital Currency] and Stablecoins. So everything has kind of warmed up and become much more exciting for a broader audience, and on the subject of payments. And the way that the Regulated Liability Network (RLN) came about was based upon the idea, just based upon the observation, that the debate on digital currency seemed to be evolving in such a way that the choice was either between Central Bank Digital Currencies [CBDCs] or Stablecoins. And to my way of looking at things, CBDCs and Stablecoins are bookends of a spectrum of possibilities. And the RLN was really an attempt to build out the middle, build out a set of options that fall between and perhaps even encompass CBDCs and Stablecoins. So I would like policymakers and other stakeholders in the payment space to understand the broad spectrum of opportunities for building digital currencies and not fixate so much on two specific examples being CBDCs and Stablecoins.
03:07 Dominic Hobson: In terms of populating that space between CBDCs and Stablecoins, it would obviously have been helpful to involve banks, and indeed central banks, and of course the non-banks, among them the Stablecoin issuers and the various payment service providers [PSPs] which have arisen in the last maybe 15 years. How easy was it to involve each of those three groups in the development of the idea of the RLN?
03:35 Tony McLaughlin: So I think everyone has a kind of common interest in modernising payment systems. As you mentioned, the central banks, the commercial banks, the regulated non-banks, and even the innovators who are currently outside of the regulatory perimeter, all want to innovate and increment. And there’s this technology which has been developed, which is the blockchain and tokenisation-type technologies, where everyone’s scratching their heads how to apply that to the innovation of payments. And there have been many attempted implementations of blockchain and tokenisation to upgrade payments. But what might have been lacking is a kind of common vision that brought together the interests of all of those parties in a common project. Because when it comes to payments networks, adoption is the whole of the law. And in order to get adoption, you can try to do it a couple of different ways. You can build a mouse trap and then ask the world to use your mousetrap. And that approach is very difficult to achieve scale because some people might not like your mousetrap. The alternative way around is to try to build consensus around a concept and a vision. And then if consensus is achieved, then building the mouse trap around the consensus is easy. That’s the approach that we have taken in RLN, is try to put forward the set of hypotheses that everyone can agree is useful to inquire into, even the sceptics. So the neat thing about RLN is that if you’re a blockchain sceptic or a blockchain adherent, you’re welcome to join us in the exploration of these core hypotheses, which, to be concrete about it, is an interoperable network of regulated money. And regulated money includes central bank money, commercial bank money, e-money, and in the future, regulated Stablecoins. That’s what RLN is. So that common vision is something that we can talk about in a public/private collaboration. And either there’s a consensus to build something around that concept, or there’s not.
06:12 Dominic Hobson: Let’s talk a little bit about how that interoperating network of regulated money works in the technical sense. The starting concept is pretty easy to grasp. The liabilities of regulated financial institutions get tokenised onto a shared ledger – deposits being the obvious example of a liability. You tokenise those onto a ledger, you can then use them to issue and settle claims. It’s not very different from how commercial bank money works today. Why did you reach the conclusion that tokenisation was a crucial step here? That might seem an obvious question, but I think it’s worth exploring.
06:49 Tony McLaughlin: It’s an obvious question. The answer may not be that obvious. So one of the core recurring questions that has plagued the application of blockchain and regulated financial services over and over again is what is the technology uniquely good for? There have been many experiments where banks and others have taken a given use-case and they’ve asked themselves, `Does this use case work on blockchain?’ And the answer has always been, `Yes.’ And that’s not surprising because a blockchain is just a computer system, and blockchains like Ethereum and their derivatives are Turing-complete. So it’s no surprise that you can run your use-case on a different type of computer. But just running a use-case on a blockchain doesn’t penetrate into what a blockchain might be uniquely good for. So, what is a blockchain? A blockchain is a `who owns what’ machine. And it’s a very good `who owns what’ machine, because who owns what is established through public key cryptography. So it’s quite definitive in terms of determining who owns what. And a blockchain is also a good state machine, meaning that the participants in a financial transaction have a good view on the state of that financial transaction. So the reason why tokenisation engines and blockchains and shared ledgers are interesting in the context of RLN is because in RLN we are trying to solve for settlement. And in a settlement machine you need to know who owns what in a very precise manner. Now, the technology is only one part of the story, because the `who owns what’ machine, the state machine, the tokenisation engine, is a technical construction, but settlement finality is a legal construct. So this shared ledger, this tokenisation engine, this `who owns what’ machine actually needs to live within what’s known as an FMI, a financial market infrastructure, because it’s that financial market infrastructure wrapper that gives the events within the state machine legal significance. So, Dominic, the interesting thing about tokenisation is, `Can we leverage the unambiguous maintenance of state of financial transactions within an FMI in order to improve financial settlements?’ That’s the rationale. And the reason why that goes slightly beyond use-case thinking is we know that the use-cases will work. What we’re trying to figure out is whether the application of this state machine to financial transactions has a Delta benefit over and above the way that we conduct financial settlements in today’s world.
10:43 Dominic Hobson: I’m glad you described blockchain as a computer system. It’s often misdescribed as a database, and that your description of it as a state machine certainly resonates in that computer metaphor, if you like. Well, it’s not a metaphor, it’s an actuality in this case. Anyway, tell me, you’ve described it’s a `who owns what’ machine and the counterpart of who owns what is, of course, `who owes what.’ So can you explain how the liabilities are actually assumed and extinguished inside the RLN, including ultimately, of course, in settlement in central bank money? How’s this going to work in practice?
11:21 Tony McLaughlin: Yeah, in practice in RLN, you imagine there being some kind of shared ledger operated by an FMI [Financial Market Infrastructure]. So the fact that it’s operated by an FMI, which is a regulated entity with a rule book, essentially means that anything happening in the shared ledger has legal significance. So to the extent that tokens move within that shared ledger, they have legal significance. Now, in this shared ledger, there are, let’s say, partitions of that shared ledger which are owned and operated by each of the participants. So a good way of thinking about these partitions is it’s a little bit like an embassy in a foreign country. You know, the US [United States] embassy in the UK [United Kingdom] in London isn’t really US territory, but by treaty, it’s kind of as good as US territory. And that’s what a partition in RLN is. It’s not really Citibank’s balance sheet, but it’s by treaty, by the rulebook of the FMI. It’s as near as you can possibly get to being Citi’s balance sheet. So each institution has a partition, each institution has an embassy within the RLN shared ledger. And the concept is that I can take claims from my balance sheet, I can take assets and liabilities from my balance sheet and represent them in that shared ledger. And after all, Dominic, that’s what tokenisation is. Tokenisation is just representation. Tokenisation happens when you go to theatre and you put in your jacket and get a ticket. That’s tokenisation of your jacket, just representation of one thing for another. So in RLN, in this partition, in this embassy that each institution has on the shared ledger, you can represent a deposit or another asset or liability that you have on your balance sheet in the shared ledger. And then when it’s in the shared ledger, it becomes essentially interoperable or composable or transferable to other participants in the shared ledger. So I can … If that is … If that much is clear, I can then go on to illustrate the basic operation within the RLN. How do we achieve payments? How do we achieve DvP [Delivery versus Payment]? Would it make sense to go on to describe that?
14:24 Dominic Hobson: I hope you’re about to say that the liabilities get extinguished by the transfer of assets through the ledger.
14:30 Tony McLaughlin: So let’s talk about the most basic operation in RLN, which is a payment. So, in common parlance, when we think about making payments, people think about sending something. But in actual traditional finance, what’s really happening is at the beginning of a payment, my bank owes me £100. And at the end of a payment, your bank owes you £100. So what operations have we conducted? We have extinguished a liability at my bank. We’ve created a new liability at your bank. And your bank will only be happy to accept that new liability if they receive a matching asset. And that matching asset will be central bank money, which is another kind of liability. It’s a liability of the central bank towards financial institutions. So here’s how a payment is made in RLN. At the beginning of the transaction there’s a token in my bank worth £100. That token means that my bank owes Tony £100. So I instruct a payment from me to you and the RLN network will extinguish or burn that token – the Tony token – and the RLN network will mint or create a new token in your bank’s partition. So there’ll be a new token created, a Dominic token, which means Dominic’s bank owes him £100. And in the central bank partition there is essentially a wholesale CBDC [Central Bank Digital Currency]. And so reserves will move, £100 of reserves will move from my bank to your bank. So that’s all happening within the same computer system, within the same settlement environment, within the same legal environment within the FMI[Financial Market Infrastructure]. And when those token movements take place, they have legal finality of settlement within the FMI. So what’s happened? My bank’s balance sheet has gone down £100 of liabilities towards Tony and down £100 of assets, which is the reserves we were holding in the central bank. And your bank’s balance sheet has gone up by £100 of liabilities and have received a matching asset in the form of wholesale CBDC within the partition at the central bank. So that’s the basic operation within RLN. Essentially, what have we done? We’ve updated three balance sheets. We’ve updated the balance sheet, my bank’s balance sheet, your bank’s balance sheet and the central bank balance sheet. And that’s what RLN is. RLN is a machine for updating multiple balance sheets in sync[hronisation] with legal certainty. That’s what RLN is. It’s a balance sheet updating machine.
18:00 Dominic Hobson: That’s very clear, Tony. And I’m fascinated by how well a computer engineer’s or software engineer’s view of the world, your use of the terms `state’ and `partition,’ how well that maps onto how payments actually work in practice in the existing system. So thank you. It’s very clear explanation. Just a question before we leave the technical side of it, who’s going to be doing the tokenising? Do you envisage the RLN supplying a tokenisation capability or functionality? Will you have a tokenisation engine inside the network?
18:35 Tony McLaughlin: At the current stage of development, which is an experimentation phase, or let’s say proof of concepts and pilots, we’re testing the hypothesis that the existence of such a state machine, such an FMI [Financial Market Infrastructure], would beneficial. And some aspects at this stage are not central to the hypothesis. So maybe in the future there isn’t one RLN. Maybe the tokenisation is done within individual institutions and organised in a slightly different way, but for the purposes of testing the hypothesis, what we’re typically doing is building a sandbox, a single technical environment, albeit one that can be distributed with the participants running nodes and the tokenisation is taking place, the minting and the burning and the transfers are taking place within a single technical environment, a single sandbox. We will get on to questions at some point about, `Well, what if an individual bank has its own tokenisation engine, and how does interoperability work?’ So all of these things will come up, but they’re not necessarily germane to the core hypothesis or the core thesis, which is to answer the question whether a common state machine would be beneficial, a common way of updating multiple balance sheets at the same time would beneficial because, as you know, achieving a common state across multiple different computer systems isn’t trivial. Achieving programmability, if everyone’s running different ledgers, is not trivial to achieve. So we kind of conveniently don’t tackle those issues head on while we’re still at the stage of trying to inquire into the concept of having a single state machine.
20:39 Dominic Hobson: Have you thought at all about how that interoperability problem might be solved? As you point out, if we’ve got different institutions using different tokenisation engines, creating their own protocols, and as you indicated at the outset, some people might not even do this on a blockchain. You don’t have to use blockchain technology. So is there anything to be said beyond what you’ve just said at this point, that interoperability is not yet something you’re addressing? Or is it something you’ve started to worry about already?
21:07 Tony McLaughlin: I think there is interesting work out there on interoperability. There’s public information about the work that SWIFT, for example, have done to bridge different CBDCs [Central Bank Digital Currencies] and other blockchain environments. So I think these bridges will come to exist. To my mind, these bridges will have to be run by trusted institutions. I’m not sure that these bridges are only technical constructs. And indeed, it would be nice to think that at some stage we might be able to move towards some kind of what I call a virtual layer one concept, which is whether people are running on different technologies, they can act together as if they’re the same layer one. Now, that’s not a trivial problem to solve, requires a lot of collaboration, and it’s not clear what the path would be to reaching that virtual layer one outcome, but indeed, in the RLN environment, we don’t think there will be one ledger to rule them all.
22:23 Dominic Hobson: Do you think it requires a public initiative of some sort to get to that virtual layer one you’re describing, because it’s a very powerful idea? If you have this common layer, and then everybody, entrepreneurs, existing companies, startups, FinTechs, can start to develop, if you like, applications on top of that. So it’s an open system. Does that, do you think that needs to be publicly owned and controlled in some way, because it requires all that collaboration you just talked about?
22:54 Tony McLaughlin: You know, I would point again to something which is in the public domain, which is the global layer one initiative by the MAS [Monetary Authority of Singapore], which is hinting towards, you know, moving in that direction. And again, public information about developments such as the Canton Network, by Digital Asset, and other initiatives by providers of private permissioned ledgers to demonstrate interoperability. If we’re going to experience positive network effects and get hundreds of institutions operating on a state machine or a virtual layer one, then we need to solve the interoperability, and we need to start the conversation going about how we actually encourage people to run nodes on these distributed infrastructures. Because, frankly, if there’s no incentive for participants to run nodes, if people are just API-ing [Application Programme Interface] into hosted nodes, you might question the idea of using DLT [Distributed Ledger Technology] in the first place. So I think people want to have the assurance that, if they’re going to run nodes, they’ll have sufficient other people to talk to, and that they won’t have to run a node for a bunch of individual silos.
24:25 Dominic Hobson: Yeah, you need the network effects. You described the tokenisation process by analogy with handing your coat in at the theatre and getting a ticket so you can retrieve it when the performance is over. Do you have a position on the debate between that type of tokenisation, where in effect the token is just a representation on the ledger of an asset that exists somewhere else and all really the holder is actually holding is, as in the case of a theatre coat ticket, a digital title to an asset which is not actually on the ledger at all. So that’s one type of token. The there are these native tokens, so the token exists only on the ledger itself. I guess CBDCs [Central Bank Digital Currencies] would, in most models, have been example of that. So have you had to develop a position on that asset-backed versus native token debate?
25:18 Tony McLaughlin: Yeah, that’s a fascinating subject because clearly the tokenisation story starts with the original cryptocurrencies like Bitcoin, which are clearly, you know, they’re theatre … They are cloakroom tickets that don’t point to any coat, so they’re not a reference to something external, they’re intrinsic in the network and don’t point to something outside of the network. And that’s, I feel like the paradigm that people have had. And these things are essentially bearer instruments. So that’s the original kind of crypto vision of tokenisation. When you’re trying to apply tokenisation to the regulated financial system, what you find is that most of the, if not all of the, financial instruments that we deal with in traditional finance are actually a form of claim on someone’s balance sheet. So again, the simplest claim is a deposit. Your bank account with your bank is a liability on the balance sheet of that bank. And almost every other financial transaction you can think of is the movement of a claim from one balance sheet to another balance sheet. So when we are tokenising in regulated finance, what we’re typically doing is creating a representation of something which is on someone’s balance sheet. And they’re also typically not bearer instruments. They’re typically claims that exist in the context of a relationship between a customer and the customer’s financial institution. So this is one thing very important in the RLN construct. In that payment that I talked about before, where there’s a token that’s burned on in my bank’s partition and minted in your bank’s partition, you might ask yourself, `Well, why not just send the token? Why not just send the token from my bank to your bank?’ And the reason for that is because that token means my bank owes me £100. That’s the relationship between my bank and me. That’s a liability on their balance sheet towards me. That liability can’t be sent to you because you don’t have that relationship with … My bank doesn’t know who you are, your bank knows who you are. So in RLN, what we’re not doing is creating bearer instruments. We’re essentially updating balance sheets. Now, are there places where bearer instruments could be useful in the regulated space? Well, potentially some aspects of the bearer nature of tokens could be exploited, could be used. For example, in the UK we have very usefully updated the law in relation to trade documents. And so the tokenisation of instruments like bills of exchange is extremely interesting. And a bill of exchange, being a negotiable instrument, you could see that as working in the following way, which is, if I sell you a bill of exchange, you will get paid on maturity as the holder of that token. I still would rather see that happen in the context of a banking relationship, of a KYC [Know Your Client] relationship, rather than being a pure bearer instrument. The trouble with bearer instruments is they’re quite amenable to financial crime. You know, there’s a salient lesson in the risk of bearer instruments in the movie Die Hard. If you remember, at the end of the movie Die Hard, it was revealed that the guy wasn’t a terrorist at all. He was just a criminal who was trying to penetrate the safe to get his hands on a bunch of bearer bonds. And we’ve been trying to get rid of bearer instruments from financial services for decades. So I’m not very much bought into the idea that we should move back to a world of bearer instruments. I’m much more of the opinion that there’s a lot of runway to tokenise existing liabilities and assets on balance sheets. And so the tokens refer to balance sheets, if that makes sense.
30:32 Dominic Hobson: It does make sense. Yes, it does make perfect sense. A couple of other things about blockchain. People make choices about which blockchain protocol to use. Most of the tokenisations which we monitor here at Future of Finance tend to be onto Ethereum. They choose Ethereum for obvious reasons, I guess. The other thing which is debated a lot is whether your network needs to be private or permissioned, whether it can be public and permissioned – a contradiction which turns out to be workable, in fact. I wonder to what extent you at RLN have had to think about those two things. Do you have to choose a protocol? I think the answer to that is probably no, from what you said earlier. And secondly, does this network, because, after all, as you say, you’re talking about updating balance sheets here, does it have to be a closed network, or can it be open to all and sundry, or some combination of open and closed?
31:31 Tony McLaughlin: Yeah, great question. As much as possible when we’re testing the hypothesis, we try to remain technologically neutral, because as you, I think, indicated before, in order to keep track of state, you don’t necessarily need a blockchain. Now it might be that a blockchain is a very good way of keeping track of state, but in a sense it’s arbitrary. You could have a normal database as some kind of authoritative data store, or you could maintain state on an Excel spreadsheet or on a paper ledger. So the concept is technologically neutral. But then when you get into the actual choice and you make the hypothesis that, well, actually blockchains are pretty decent state machines, or you would want to keep track of state through public key cryptography, that might be a reasonable assumption, and you would want to have some kind of immutable record of the transactions – that’s a good assumption. Let’s say that we’re going to use a blockchain or shared ledger or DLT [Distributed Ledger Technology], take your pick of your preferred terminology. And then the next big dividing line is, `Do you use public or private permissioned?’ Now what is the consideration about that choice? Well, the consideration is that whenever a regulated institution is using an external computer system, it has to consider that as being an outsourcing. Now that’s exactly the same as if I choose an outsourced traditional tech vendor or if one was to choose a Cloud provider. So, you know, think about this, Dominic, that if a bank or a regulated firm is considering a move to Cloud, it has to do something called third-party risk management. And third-party risk management is a process to establish whether that is a responsible outsourcing. So it’s exactly the same hurdle if you want to use an external computer system. I mean, Cloud is just outsourced compute and storage, and blockchain is outsourced ledger. So it’s the same hurdle: third-party risk management. And then you ask yourself at this stage in the development of blockchains, which path is most likely to be able to pass third-party risk management? Well, public is very difficult because, for a start, there’s no third-party to risk manage. So that’s a major impediment. But there are other impediments in the public space. For example, regulated firms would be extraordinarily conservative about participating in a network that might result in the payment of gas fees to sanctioned entities, which is a consideration on one of the largest public blockchains. So the question is just a practical one, Dominic. It’s not an ideological one. It’s what kind of network can pass third-party risk management. And in the current state of the world, it’s much easier to see how a private permissioned blockchain could pass that hurdle. So that’s the path that we’ve been walking in terms of RLN experimentation, because it’s not clear in this stage of development that the public blockchains can pass third-party risk management. Not clear that we can establish that they are responsible outsourcings.
35:55 Dominic Hobson: Earlier, you described the RLN as a new financial market infrastructure, a new FMI. If it is a new FMI, it has obviously to be owned by somebody and governed by somebody. Have you thought about who should own it, who should operate it, who should write the rule book? And to your last point, about private permissioned, decide who can actually access it? Are those questions being answered already?
36:21 Tony McLaughlin: There are questions being addressed, for sure. And, you know, building new FMIs [Financial Market Infrastructures] is not simple, especially something which is novel like RLN. But there are several precedents. I like to think back to the formation of continuous linked settlement [CLS], which is an FMI that was created around about 2000 or before in order to solve FX settlement risk, which operates in multiple currencies and has a college of supervisors and is set up as a bank. And so that is one way in which a new FMI can come into being. And such an FMI can come into being if there is a consensus between the public sector and the private sector, that it is required and it solves a problem and that it’s needed. So RLN is more about consensus building than it is about testing technology. It’s more about asking the question of, does the community, does the regulated community, the public and the private regulated community agree that there is a gap that needs to be filled? And if that agreement is reached, then you have to then agree what’s the location of the legal entity, what’s the shareholding of the legal entity, what’s the rulebook, etc. etc. All of that stuff becomes much easier once consensus has been reached that the need exists to build the utility.
38:16 Dominic Hobson: So, in the same way that Herstatt Risk created a need for CLS, so central banks and private banks worked together to create the entity we now know as CLS, which I guess combines public and private in its governance and its operations.
38:32 Tony McLaughlin: Yeah, but Dominic, remember that Herstatt happened in 1974. And so it took quite some time for the consensus to come together.
38:44 Dominic Hobson: Okay, we must wait 25 years for this to come to fruition then? You said earlier that adoption is the whole of the law. It was the phrase you used, and that’s clear. Even the best ideas are going to need market participants to pick them up and use them to generate those network effects, which you also mentioned. The good news is, of course, as I mentioned in my opening comments, that both the Bank for International Settlements [BIS] and the International Monetary Fund [IMF], with its XC platform for cross-border payments, they both pick this idea up. And endorsements like that are clearly very useful when you’re trying to attract banks, because banks and central banks do listen to the BIS and the IMF. I just wonder if that’s enough. Does the RLN need a … You’ve talked a lot about the need to build consensus, but … Is that the adoption strategy?
39:40 Tony McLaughlin: So I wouldn’t characterise BIS and IMF work as being an endorsement of RLN. I think what’s really happening is there’s a common interest to explore the tokenisation thesis in the regulated community. And I think, as we’ve touched upon, the tokenisation thesis in the regulated community is different from the tokenisation thesis in the crypto community. We’re after different things. In the regulated community, we are not trying to build a trust-less system, we’re not trying to build censorship resistance, we’re not trying to build non-sovereign currencies, commodity currencies. We’re not interested in Tokenomics, we’re not interested in Proof of Work. So all of that stuff which is intrinsic to the tokenisation thesis in the original crypto vision is not of interest to us in the regulated space. We’re interested in a subset of the potential benefits of blockchain DLT [Distributed Ledger Technology] shared ledger. And that subset is the potential to have a new kind of FMI [Financial Market Infrastructure] or new kind of settlement venue, which is always-on, multi-asset state machine, programmable, a way of updating multiple balance sheets unambiguously at the same time with legal finality of settlement. Those are the things that we’re interested in. And individual banks, BIS, IMF and RLN communities and others are investigating that hypothesis. Which, again, I just want to emphasise the tokenisation thesis in the regulated space is a subset. It’s not a subset, it’s just different from the tokenisation thesis in the crypto space. So I think there’s a very interesting point where, in my view, we’ve moved beyond mono-line or proprietary visions. If you think about a CBDC [Central Bank Digital Currency], for example, a CBDC is another silo. It’s a mono-line implementation which only contains one form of regulated money. And so I think what we’re seeing is, in places like Brazil and in Korea, and in the unified ledger, is the idea of having platforms that contain multiple forms of regulated money. Not just single forms of regulated money. You know, one, one observation, one, let’s say, tentative conclusion, that I’ve come to in this journey is using DLT to recreate an existing silo, to me, doesn’t have much appeal. What I’m interested in about DLT is multi-asset settlement venues. Because what multi-asset settlement venues would get us beyond is … What does everyone complain about in traditional banking, Dominic? Silos. Silos are special purpose infrastructures for given legal instruments. Everyone complains about silos. So what’s fascinating about the shared ledger technology, and tokenisation is moving beyond the silos, because at the end of the day, Dominic, every financial instrument is a claim. So why can’t we have a machine where parties can come together and transfer claims? You know, deposit is one type of claim and a security is a different type of claim, and a letter of credit is a different type of claim. Why can’t we have a general-purpose venue for the transfers of claims? That’s what’s interesting about DLT.
44:06 Dominic Hobson: Yes, I can see why it’s attracted such interest in the securities and funds and money market industries as well as payments. But just to stay with adoption a minute, if we are moving towards this world where we abolish silos, we update balance sheets, it’s difficult to disentangle cause and effect here with network effects, but network effects will certainly help this network to grow. Does that mean that not only do you just have to have all asset classes and all types of financial institution on this standardised venue for settling liabilities? Does it mean that the institutions which do commit have to commit totally their balance sheet? In other words, all banks have to submit all liabilities to get the benefits of the RLN? Or can you tokenise some of your liabilities and keep others back in the old world? Do you need to be totally committed to this as a single financial institution for it to work? Or can you take part on a suck-it-and-see basis?
45:11 Tony McLaughlin: No, I don’t see that `all-in’ being necessary at all. I mean, if you’re a farmer in the Middle Ages, you take some of your pigs and some of your wheat to the market and you transact some of your stock at the marketplace. Is that a good analogy? I don’t know. But the question about RLN is not about tokenising every aspect of your assets and liabilities. It’s just, `Would it be interesting if there was a place where you could do so?’ Why is it that we’ve got special purpose rails for the transfer of subsets of the assets and liabilities? You know, when we talk about the transfer of deposits, deposit liabilities, and central bank reserves, we call that a payment system. But there are other infrastructures for the settlement of securities. And we act like a security settlement system and a payment system are different species. But if we abstract the fact that they’re just the transfer of claims, it becomes the same. And this is the history of computing, Dominic, as you know. We had a calculator with four buttons, plus, minus, divide and multiply. And then we moved to general-purpose computers. It would be ridiculous to have a computer that could only do YouTube, and another computer that could only do Spotify, and another computer that could only do Netflix. No, we have general-purpose computers because we’ve reached a level of abstraction where we understand that these are all the same, that they require the same processing. So RLN is simply asking this question, `Can we have a venue where banks can represent assets and liabilities in that venue and exchange them with other participants?’ And in that venue, those movements of tokens, those operations on the tokens, have legal significance. So it’s really asking this question whether we can make that breakthrough, that conceptual breakthrough towards multi-asset settlement environments, and the recognition that every financial instrument and every financial transaction boils down to balance sheet updates across multiple firms.
48:07 Dominic Hobson: That’s why what you’re describing is an infrastructure, not a commercial closed network. It’s a kind of standardised layer for all these different asset classes and different types of financial institution. So I’m wondering, what would put people off committing to it? Is it a fear that they would lose control, that if you put your liabilities into the RLN, you have to do it the RLN way, and you run the risk of assets and customers defecting to your competitors? Has that sort of issue come up, that loss of control come up as an issue at all in your discussions?
48:49 Tony McLaughlin: Well, Dominic, the core thing is, what’s the Delta benefit of this venue existing? And here’s a couple of difficulties that will arise. They’re conceptual. So in financial institutions, people work in the payments business, or the trade business, or the securities business, and they’re very deeply into their use-case. So the perspective is from an individual use-case perspective, and as we said before, if you put a use-case onto a blockchain, it works. The question is, does it work any differently to how the existing traditional environment works? So that’s one difficulty, which is, when we’re looking at it from a siloed perspective, we’re always trying to find the Delta benefit in something that already works. That’s one issue. The second issue, frankly, is we’ve realised that having this central venue or having this new venue, it actually doesn’t get rid of reconciliation, because whatever happens in that venue has to be reflected back at the mother ship. Because the mother ship, meaning the individual institution, is a balance sheet and that balance sheet needs to have a consolidated vision of the assets and liabilities. So anything happening in the venue has to be updated back at the mother ship. So it’s not as if you can just, have … You mentioned before these native tokens running around which are not updated on anyone’s balance sheet. No, those liabilities and assets have to be updated on the mother ship. So the mother ship has to synchronise with this venue to reconcile with this venue. So I would say that the core exam[ination] question is, `Does the existence of this new type of venue have a meaningful Delta benefit? Are there things that you can do on it that you can’t do with existing rails?’ And that’s the purpose of all of the experimentation which is going on.
51:31 Dominic Hobson: On that reconciliation point – the need to reconcile the state of the RLN with the state of the balance sheets of the individual banks – does that argue for an expectation that any user of the RLN would be running their own node in the network, if you like? They wouldn’t simply leave it to the operating company to run those nodes for them? They’d want to be densely involved themselves to facilitate that reconciliation work?
51:56 Tony McLaughlin: Dominic, one may ask, if there’s a DLT [Distributed Ledger Technology] where people are just APIng [Application Program Interface] into hosted nodes, what’s the point of running it on a DLT in the first place?
52:08 Dominic Hobson: Yeah, you said that. Yeah.
52:11 Tony McLaughlin: And so, you know, another thing, you know, one question is, `Is there a missing layer of the financial system, which is, should there be this industry state machine? [That] is one question. And then the second question is, if you believe that should exist, I mean, I think an industry state machine is very interesting. If you think about it, taking a few steps back, we have industry messaging utilities like SWIFT, so we’ve abstracted out. So secure structured messaging is a layer in the financial system that everyone uses, but we don’t have an industry state machine. Would it be interesting to have an industry state machine? I think potentially yes. Then the next question is how do you implement it? And some people might say, `Well, you don’t need a blockchain at all, you just need a big authoritative data store, centralised, and everyone can API into that authoritative datastore.’ And some other reasonable people might argue that actually, it would better to implement that on a blockchain where all the participants are running nodes, because that might be more resilient. So those technical implementation arguments can be had. But we would be in a much more informed place if we had that discussion having agreed on the benefits of there being an industry state machine.
53:51 Dominic Hobson: That industry state machine, if I’ve understood you correctly, would dispense with the present need we have, as you say, for exchanging data in these structured message formats like SWIFT. I mean, apart from the standardisation of the machine itself, the state machine itself, is there any need for standardisation here in terms of how data is represented or does that just kind of go away because you’ve got this single model capable of being used by any financial institution in any asset class? It’s a single way of doing things because you’ve reduced all those issuances of liabilities and issuances of assets to a single process. You standardised the process itself and you don’t need to build this abstract layer on top to cover up the fact you haven’t done that.
54:42 Tony McLaughlin: I don’t see the state machine overlapping or getting rid of the need for messaging. I mean, the intuition that I often use to give an idea of how we might evolve is that if I was to pose the question, `What’s the best way of organising a dinner party with ten people? Is it through email or through WhatsApp or a messaging app?’ Most people would agree that doing it through a messaging app is the best way to proceed. And why … The question is why is that? And one of the reasons why it’s easier to organise a dinner party through a messaging app is because everyone has a current view of the status of the arrangements. I know you’re bringing a bottle of red. I know that I’m bringing the dessert. I know someone else is bringing, you know, a bottle of white. We can have an up-to-date understanding, shared understanding, of the status of the arrangements. In today’s world, we’ve just got the messaging part, we haven’t got the state part. So what RLN could be conceived of as being is this new infrastructure. We’ve got the messaging and the state combined. And anything happening in that venue because it’s within an FMI [Financial Market Infrastructure], has legal significance. So I think we still need the messaging piece if we’re going to use blockchain for this state machine is a good `who owns what’ machine, but it’s not the machine for moving a thousand lines of invoice data. It’s a bean-counting machine. It’s `I own this and you own that.’ But in a payment, in a commercial payment, I might be settling 1,000 invoices with one payment. And so I’m sending a thousand lines of invoice data. That thousand lines of invoice data is not going to be in the state machine. So it’s the fusion of the messaging with the state machine which makes all the difference. Potentially. It’s a thesis.
57:07 Dominic Hobson: The state machine doesn’t do everything.
57.11: Tony McLaughlin: Correct.
57.12: Dominic Hobson: What you end up with is like a network of networks, and those networks might do many different things, including your example there of the allocation of a single invoice across a thousand accounts. And you get something similar in the allocation of securities after a settlement has taken place. And that work will not be done by the state machine, it will be done by the users, and possibly by networks of users which are kind of doing that work already, aren’t they? Which – you may want to say something about that – but it brings me to another question, which is netting. One of the criticisms often made of blockchains in general is that they, because they have this atomic settlement model, dispense with the benefits of netting, by which I mean the netting of payments liabilities by the automated clearing houses before they go into the central bank RTGSs [Real Time Gross Settlement system]. I mean the clearing of securities transactions before settlement through central counterparty clearing houses [CCPs]. From what you’ve just said, I would assume that those benefits of netting can be preserved because they remain netting networks and you can use them if you like, but the ultimate exchange of updating of balance sheets takes place inside the state machine. Everything else happens outside it. Have I understood that correctly?
58:33 Tony McLaughlin: Yeah, I see no reason why you can’t build the logic for netting within an RLN environment. And netting is indispensable. Netting is indispensable. We should not, you know, throw the baby out with the bathwater. And in RLN you can decide the points at which you do the settlement in central bank money, and that can be on a deferred net settlement basis. I see no problem with doing that.
59:07 Dominic Hobson: Okay, so you could also choose `atomic’ settlement. There’s nothing fundamentalist about the RLN needing to switch to atomic settlement. That will just be a choice, right?
59:17 Tony McLaughlin: Exactly. Again, we need to be very careful in this whole endeavour not to carry over stuff from crypto into regulated financial services. So here’s something, you know, an amusing, you know, thing to think about, which is, `Is there any such thing as a negative Bitcoin?’ There is not, but there is a negative balance you can have in your bank account; you can have an overdraft in your bank account. How do you represent a negative balance in a blockchain? We need to preserve the degrees of freedom that we’ve got in regulated financial services and not live within the confines of what has been created for the crypto ecosystem.
01:00:13 Dominic Hobson: And equally we have to carry into this state machine – I’m using this phrase, I rather like it – concepts from the traditional markets, including the legal concept of settlement finality, which you referred to earlier. So a transaction can’t be reversed even if the counterparty falls over. Now that’s a fully realised legal concept used by our existing settlement systems, the RTGSs [Real Time Gross Settlement systems], but also in the securities markets as well in CSDs [Central Securities Depositories], all of them performing to these internationally agreed standards set by the CPMI [the Committee on Payments and Market Infrastructures], IOSCO [International Organisation of Securities Commissions], the BIS [Bank for International Settlements] committee. Talk to me about settlement finality inside the RLN. Is it a technical issue or technical obstacle you need to overcome to build that in?
01:01:04 Tony McLaughlin: I’m glad you mentioned it. I mean, I have to laugh when people in the blockchain community say that a blockchain achieves atomic settlement, because settlement is not a technological construct in the regulated space. Settlement is a legal construct and there needs to be an event. I mean there needs to be some way of demonstrating an event which constitutes settlement, but that settlement event is given legal meaning by it taking place within a contractual, in a contractual and/or regulated environment. So again, blockchains might be a good way of evidencing settlement events have taken place, but they only actually take place within a legal construct. That’s why, again, when people in the blockchain community say that code is law, I say it’s not. The law is the law and the law comes first. And the technology has a significance in instantiating legal constructs or supporting legal constructs. So finality of settlement is what we’re trying to solve for in RLN. We’ve got bank balance sheets being represented in a multi-asset ledger and assets and liabilities moving between participants. When they move within the venue, they need to have legal significance. Now that legal significance can be evidenced by the technology, but it’s subservient. The technology is subservient to the law.
01:03:01 Dominic Hobson: Something else blockchain evangelists talk about, obviously, is disintermediation. We’ve talked about this state machine as being programmable, a single programmable platform, if you like. Now, if you talk to a blockchain person that means, `Oh well, we can disintermediate everybody and everything that these various intermediaries do can be executed by computer code rather than people working in banks and investment banks and asset managers.’ What’s your answer to the … If somebody said to you, `Does the RLN imply high levels of disintermediation?,` what would your answer be?’
01:03:38 Tony McLaughlin: I would say, again, it’s a carryover. It’s people confusing the crypto tokenisation thesis with the regulated financial services tokenisation thesis. These are two different things. The crypto world is trying to build a trust-less censorship-resistant world without central banks, world without regulated financial institutions. That’s the crypto, the original crypto project. And when you take that idea into the regulated space, it doesn’t apply because we live in a world of regulated balance sheets and those balance sheets serve a purpose. Those balance sheets are performing an intermediation role. Banks and insurance companies and these central bank balance sheets perform vital intermediation roles. So we have to understand that the original crypto intention was to build a trust-less world that doesn’t have intermediaries. But that vision does not apply to the upgrading of the traditional financial services, which is based on trust and based upon the value-add that balance sheets provide to the world.
01:05:15 Dominic Hobson: And it’s updating those balance sheets which, to go back to one of your earlier observations about why what the RLN is not going to be moving around – or have running around – are these data objects, your classic tokenised asset, if you’re approaching this from the classic blockchain perspective. You’re actually doing something different, aren’t you? You’re delivering settlement through this state machine, extinguishing liabilities with assets. You’re not moving data about those assets or liabilities around at all, are you? That data component remains outside the state machine, doesn’t it? I hope I’ve understood that correctly.
01:05:56 Tony McLaughlin: Well, there’s data that has to be connected to settlements because it’s germane to the settlement. Again, if I make you a payment for a million dollars, which is settling 1,000 invoices, you don’t just need the million dollars, you need the data about the thousand invoices to update your ERP [Enterprise Resource Planning] system. So there’s always data which has to accompany the settlement. And that’s the yin and the yang that we try to deliver through RLN. It’s the messaging plus the settlement coming together and the movement, the updating of balance sheets, because many financial transactions require multiple balance sheets to be updated. So wouldn’t it be nice if there were a place where those multiple balance sheets could be updated in an orchestrated manner with legal finality of settlement. That’s the RLN thesis.
01:06:59 Dominic Hobson: But it’s data about the settlement only, isn’t it? This bank owes that. It’s a `who owes what’ machine. To take your earlier example about the single payment, which actually represents 1,000 different payments to 1,000 different accounts, that takes place somewhere else. So the data focus inside the RLN is very narrow.
01:07:24 Tony McLaughlin: Well, that is .. I mean, RLN would deliver both. Yeah, RLN would deliver both. I mean, as a package, it might not all be built within the shared ledger. But here’s another lesson, which is … I think there have been many examples which have demonstrated that you don’t want to take a DLT [Distributed Ledger Technology] and just build everything into the DLT. You don’t want, for example, all the business … You don’t necessarily want to build a computer system where all of the business logic is written in smart contracts in the ledger. So I think we’re moving towards a situation where we’re trying to really narrow down what is the DLT or the shared ledger or the blockchain really good for. And I characterise that as really being the `who owns what’ machine or the state machine. But that doesn’t mean that every bit of business logic needs to be written as smart contracts in the ledger. And it doesn’t mean that all data flows have to be transacted across the ledger. So why do we believe that everything in an application needs to be built on the DLT? There’s no need for us to have that all-or-nothing approach.
01:08:53 Dominic Hobson: Another criticism you’re going to face from blockchain evangelists – probably already are – is that the RLN model preserves this `singleness’ of fiat currency so that the consumer will not get access through the state machine to non-sovereign forms of money like Stablecoins. Talk through what your response to them would be because, as I understand it, the RLN can support CBDCs [Central Bank Digital Currencies], or at least wholesale CBDCs. It can also support different forms of commercial bank money, including Stablecoins, including tokenised deposits, and presumably also e-money. So actually the RLN is a more capacious model for different types of digital money than maybe some of those blockchain critics are indicating. Perhaps you could just talk us through those layers – the CBDCs, Stablecoins, tokenised deposits, e-money – how they all fit into the RLN machine.
01:09:57 Tony McLaughlin: So the clue is in the name: Regulated Liability Network. Why is it called Regulated Liability Network? It is because if you have an account at the Bank of England or a central bank, if you have an account at a commercial bank, or if you have an account at an e-money provider, then they owe you the money back. In other words, they have a liability on their balance sheet. It’s a special type of liability. It’s promised to pay the customer on demand, at par value in national currency units. So that’s a promise to pay, a promise to pay you back. And the regulation that surrounds the formation of the central bank, or the formation of a commercial bank, or the formation of an e-money company, the purpose of that regulation is to make sure you get the money back. That’s why it’s a regulated liability. It’s not just a `best efforts’ liability; it’s a regulated liability. So there’s space in RLN for any regulated liability which is offered under licence by the nation-state.
01:11:15 Dominic Hobson: So a regulated Stablecoin, that is, one issued by a bank, or an e-money backed 100 per cent by cash deposits at a regulated bank, they would both fit into this model?
01:11:27 Tony McLaughlin: So were there to be a regulatory regime and issuers issuing Stablecoins under that regime that were regulated liabilities, then potentially they would be eligible for inclusion into RLN, and then they would be all interoperable with those other types of money which are in our RLN. But again, Dominic, drawing a sharp line between the thesis promoted by the original crypto vision and the modus operandi of the regulated space, we are unambiguously and unashamedly supporting the development and the improvement of sovereign currencies. I mean, I’ll tell you, when I first read the Bitcoin white paper, which is a very significant amount of time ago, it really shook my understanding of what money was, because for the first time I saw a thesis there which was that a form of money could be created which didn’t emanate from the nation-state. I had to ask myself, `Does the nation-state have a monopoly over money? Is money the prerogative of the nation-state?’ And the answer that I came to personally and professionally is, `Well, of course it does, because that prerogative emerges from the social contract.’ There are certain monopolies that we grant to the nation-state in the writing of laws, in the monopoly over force, in terms of the military and the police, and the monopoly over many aspects, several aspects of our lives that we cede to the nation-state. And one of the most important is the prerogative over money. So RLN’s idea is to make sure that sovereign money meets the emerging needs of the digital economy. Because if users of money migrate away from sovereign money to unregulated money, that’s not a good outcome for our societies. So RLN is a defence of sovereign money, and it’s an idea to upgrade sovereign money in the face of a rising tide of unregulated alternatives.
01:14:21 Dominic Hobson: If I put it to you plainly would Tether, for example, be admissible to RLN? My instinct is not.
01:14:29 Tony McLaughlin: Well, for an issuer to be eligible for an RLN – again, an RLN doesn’t exist today, but in its conceptual form, it would only admit participants who are regulated in a jurisdiction, so have a licence in a jurisdiction to issue the instruments that they’re licensed to issue. And those instruments, the kind of Stablecoins that we envisage within RLN, are much like e-money, meaning you give a dollar and you get a dollar back. It’s not an instrument which is trading on an exchange where the price can vary from a dollar. If you give a dollar to an e-money institution, you expect to get a dollar back. So it’s that kind of instrument, that regulated liability, which is redeemable at par value on demand in national currency units, which is conceivable within RLN.
01:15:43 Dominic Hobson: So even if the likes of Tether were depositing money with regulated banks, that wouldn’t be sufficient, because they’re not necessarily guaranteeing to give you a dollar back when you ask for it.
01:15:55 Tony McLaughlin: We’ll see how Stablecoin regulation develops in different countries, but there wouldn’t be – without talking about individual issuers – the hurdle would simply be that there is a well-formed regulatory regime and an issuer in compliance with that regime. And the token would be unambiguously a one-to-one redeemable at par value instrument.
01:16:25 Dominic Hobson: I suspect you’d know it when you see it. A couple of final questions for you, Tony. I’m conscious I must let you go. But the first is about a use-case which is emerging for RLN, and this is cross-border payments. This work has been going on, at the detailed level done by the CPMI [Committee on Payments and Market Infrastructures] to try and make cross-border payments faster, cheaper, more transparent, more accessible. And that’s driven at the very highest level from the G7 [Group of Seven] and United Nations. This has emerged with the BIS [Bank for International Settlements] taking the lead here. I was very interested to see a speech by Augustin Carstens, the general manager of the BIS, talking about shared programmable platforms, a unified ledger, a very clear reference to the RLN model as a way of making cross-border payments faster, cheaper, more transparent, more accessible. In line with that mandate from the global regulatory authorities, we’ve seen the BIS now launch its Project Agora. It’s going to explore – these seven central banks – how to make tokenised money, how tokenised money could make cross-border payments better. So it’s clearly not a kind of Utopian project here. It’s actually become a real use case, hasn’t it?
01:17:45 Tony McLaughlin: Yeah, it’s a … That recently announced project is exploring the concept of there being some FMI [Financial Market Infrastructure] presumably, running a kind of shared ledger. And in that shared ledger there are essentially wholesale CBDCs [Central Bank Digital Currencies] from multiple jurisdictions and commercial bank money tokens from multiple jurisdictions. And if you think about … Let me just try to make it concrete and give you a real example. So. in today’s world, imagine a UK manufacturing company that’s buying an industrial robot from Japan, and imagine that robot costs US$20 million. So you will have the importer on the British side will have their UK bank, and most likely that robot is priced in [US] dollars. So the UK bank will have their US correspondent bank. So that’s the UK side of the transaction. And then in Japan, there is the Japanese exporter, who have got their Japanese bank, and that Japanese bank have got their US correspondent bank. And then in the middle, you have the Fed [Federal Reserve Bank], because that’s where the final settlement is going to take place. You have five balance sheets that need to be updated. You have the UK bank and their dollar correspondent, and the Fed and the Japanese bank and the Japanese bank’s dollar correspondent. So, in today’s world, those balance sheet updates across five different balance sheets are effected using messaging. There is no common state machine, there’s no FMI sitting above it, which establishes legal finality of settlement end-to-end. So imagine you have those five balance sheets represented on a shared ledger, and all five balance sheets are updated in sync[hronisation] within a rule book within an FMI that operates 24/7, and that transaction can take place almost instantaneously. So that’s the kind of vision, or the kind of hypothesis, that will be tested in Project Agora. And let’s see what the results are. Let’s see whether that kind of infrastructure solves the pain points in cross-border payments or not. You know, one thing, Dominic, I would say that the great thing about these types of experiments, again, I said that there’s room for the sceptic and room for the adherent. In entering these experiments, the best attitude to have is that you don’t mind whether you find … You don’t mind whether you find evidence that supports the thesis or falsifies the thesis. Falsifying the thesis is as valuable as supporting the thesis. That’s the best way of going into these experiments.
01:21:38 Dominic Hobson: You’re being a bit self-effacing there, I think. Do you, do you not feel that with Agora, this idea you had and which you’ve helped develop, is actually finally coming to life?
01:21:49 Tony McLaughlin: I am happy to see the community moving beyond this idea of having mono-line or recreating existing silos using DLT [Distributed Ledger Technology]. I’m happy to see a public/private collaboration emerging, because one of the unfortunate things about pure CBDC projects is it is indeed a defence of the sovereign currency, but it’s only one slither of the sovereign currency. The sovereign currencies are public/private collaborations. So I’m absolutely heartened to see the emergence of public/private collaboration in actual projects. I think it’s much better, by the way, than consultations. I think consultations between the public and private sector have their place, but they can be augmented by actual experiments and projects where you bring together the public and the private sector around the same table. So I’m very heartened to see that kind of public/private collaboration develop, because sovereign currencies must be upgraded to service the needs of the digital economy. If we don’t upgrade the sovereign currency system, then people and businesses will migrate towards unregulated forms of money, and that’s bad for nation-states. So that’s the reason why Project Agora and other public/private collaborations are so valuable.
01:23:48 Dominic Hobson: Upgrading is, as you say, good for nation-states, but what’s going to be good for domestic customers about this? Now, in the case of cross-border payments, it’s pretty obvious: the benefits will be faster, cheaper, more transparent, more accessible cross-border payment services. But if you think about the benefits of RLN in a purely domestic context, what is the man and woman in the street, if you like, going to get out of this? Are they also going to have faster, cheaper, more accessible payments or are they going to get other things as well.
01:24:21 Tony McLaughlin: Quite possibly, this RLN infrastructure, from a domestic perspective, will be quite a powerful platform for innovation. There was a project called Project Rosalind by the BIS [Bank for International Settlements] and the Bank of England, which demonstrated innovative use-cases built upon a ledger that only contains central bank money. How much more powerful would the use-cases be if the ledger not only contains central bank money, but also other forms of regulated liabilities, like commercial bank tokenised deposits, potentially e-money tokens? So let me again try to make it tangible. The rails that we have today kind of force particular interactions. Let me give a normal, an everyday interaction to illustrate this. So let’s say that you are ordering some food from a food delivery platform, and how does the money work? So you’re ordering a takeaway from a food delivery platform. You’re doing for many people. So let’s just say it’s £100. Okay, so you pay £100 to the platform. The platform doesn’t actually need to receive the £100, but in today’s world, it’s the only way of doing it. You send £100 to the platform and the platform then keeps, let’s say £15s, and then it gives £5 to the driver, to the delivery person, and it gives the rest of the money to the restaurant. These are separate transactions. One payment from you to the platform, one payment from the platform to the delivery driver, one payment from the platform to the restaurant. Now, as a consequence of the platform receiving all the money, they also have to have a payment institution licence, because they’re handling third-party funds and platforms who are doing food delivery don’t want to have a payment institution licence. So what if there was a network where all the parties are in the common shared ledger and you are debited the£100, the platform is credited their commission, their £15, the rider is credited the £5, and the restaurant is credited the £80 without the money passing through the platform. So my point there is to say that the fact that our rails today are kind of bi-directional. Send and receive money between two parties means that multi-party transactions have to happen in a certain way. You know, what about in the future, that when you buy something in a shop, part of the money goes to the merchant, but the VAT goes directly to the government. So these interactions aren’t possible in today’s world because our rails are bi-directional. If we had this orchestration platform in the middle for value transfer, we might be able to build interactions that are not possible today, even from a domestic perspective.
01:28:20 Dominic Hobson: And if I reduced the benefits of what you’re describing to two words, it would be less risk – actually more than one word, it would be like two words – less risk and more efficient.
01:28:32 Tony McLaughlin: Yeah, less risk, more efficient, but also more verbs. Meaning that payment systems, again, payment systems today are based upon the verbs `send’ and `receive.’ But there are other verbs that you could apply to value transfer systems that are not currently supported. So I’ll give you another example that I like, which is, think of the verb to `assign.’ So in today’s world, if I want to give my son some money to play on Fortnite, I have to send him some money. But he’s only 12, he doesn’t have a bank account, so why can’t I assign him some of my money to play on his video game? Or if I want to provide some money to the lady that looks after my father to buy groceries, I either have to pay for it myself or send money to her. I can’t assign money to her to buy the groceries. Or maybe in the future there will be AI [Artificial Intelligence] agents. If we want AI agents to make payments for us, wouldn’t it be useful to be able to assign a pot of money to the AI agent to conduct business on our behalf? So that ‘assign’ function isn’t a feature of payment systems today. And I come back to the analogy of the calculator. Our payment systems today are like calculators with four buttons. What if our value transfer systems in the future were extensible with respect to function? In other words, you could add new verbs into the value transfer system. Do we want value transfer systems with four buttons, or do we want them to be extensible with respect to function? I hope I’m not being too abstract or speaking hieroglyphics. You know what I mean?
01:30:53 Dominic Hobson: No, you’re being very clear. You’re describing how payments could evolve, actually into something of a more interesting service industry. It’s not just sending and receiving money, but actually it’s providing service as well.
01:31:05 Tony McLaughlin: And then imagine that there are other assets on the same network. Imagine that there are bills of exchange and promissory notes, and equities, and government debt and corporate debt, and the sky’s the limit. As long as something fits on a bank’s balance sheet, it could be instantiated within the network.
01:31:33 Dominic Hobson: You can turn your financial assets into money or into payments as well.
01:31:37 Tony McLaughlin: Well, it comes down to the fact that the financial world is really the transfer of claims between one balance sheet and another.
01:31:51 Dominic Hobson: Tony, I have one final question for you, which is just about where the whole RLN project has got to. As I understand it, the discovery phase was completed with EY and UK Finance. An experimentation phase has started. Now, there’s obviously this work going on at the BIS [Bank for International Settlements] via Project Agora. So when can we expect the RLN to actually go into production?
01:32:20 Tony McLaughlin: In 26 years.
01.32.21 Dominic Hobson: Like CLS?
01:32:22 Tony McLaughlin: Yeah, like CLS. We are in the experimentation phase, the hypothesis testing phase, and it’s very … many people ask me this question, `When does it go live? When does it go live?’ That kind of pushes me into being a proponent or someone who’s trying to sell this concept, which I’m not. What I’m interested in doing is finding evidence for or against the hypothesis, which is the conceptual hypothesis around RLN, and also critically trying to find evidence as to whether there is or there is not a consensus to build something new. Which, obviously, that second part flows from the first part. If there’s strong evidence in support of the hypothesis, then it’s likely that a consensus can be formed to actually build it. So there are two steps, Dominic. Number one, gather the evidence around the hypothesis. Is it supported or is it falsified? Number two, is there a consensus to actually build something? Because FMIs [Financial Market Infrastructures] are very expensive to build. They’re not cheap. And let’s revisit at the end of these experimentation phases to see whether or not the evidence supports or not and whether there’s a consensus to move forward. But, if you forgive me, I won’t be pushed into the camp that the only answer is RLN. If it transpires that RLN leads to some new hypothesis or new synthesis, it will have more than served adequate purpose.
01:34:22 Dominic Hobson: Tony McLaughlin, thank you very much for taking so much time to explain to the members of Future of Finance the thinking behind the Regulated Liability Network. We’re very grateful for your time. Thank you.
01:34:34 Tony McLaughlin: It’s a great pleasure. Thank you.
