2023 Financial Markets Conference - Old Challenges in New Clothes: Outfitting Finance, Technology, and Regulation for the Mid-2020s

Policy Session 3: How Might Web3 Change the Financial System?

Web3 is a hot buzzword, but what does it mean for financial services? In this session, Christine Parlour presented a paper on its implications for the financial system, then joined Hilary J. Allen, Stephen Diehl, and Richard Walker for a panel discussion moderated by Marina Moretti.

Transcript

Marina Moretti: Good afternoon, everyone, and welcome back to the meeting room for this third policy session on "How Might Web3 Change the Financial System?" We are going to have an hour and a half in which we are going to have a paper being presented, a session to discuss the topics raised in the presentation, and a final session with Q&A (about a half hour each session), and we have a very nice lineup of speakers for this afternoon. My name is Marina Moretti, and I am a deputy director at the International Monetary Fund. I have the pleasure of being joined by Christine Parlour, the professor and chair in finance and accounting at the Haas School of Business at the University of California, Berkeley.

Unfortunately, Christine got stuck in her transit to Florida, but we're very pleased that she could join us virtually. Next to her, we have Hilary Allen, who is the professor of law and associate dean for Scholarship at the American University College of Law in Washington, DC. Then we have Stephen Diehl, software engineer and blogger at stephendiehl.com. Next we have Richard Walker, who is a partner with Bain & Company. This is a very experienced panel, coming from a different perspective on the issue of web3 and the future of financial services. Let me start by giving the floor to Christine, who will make her presentation, followed by the panel.

Christine Parlour: Thank you very much. Once again, apologies for not being there in person, but it turns out that Houston is an impenetrable barrier that repelled me from going any further. What I want to do is just to set the stage in a very general way and talk about some of the economic benefits that can flow from web3, and what potentially it's going to do to finance and the finance industry. Of all the industries out there, finance has been almost the slowest to automate in the sense that academic research has shown that essentially the marginal cost of intermediation has been almost static for the last 100 years. There's this possibility of essentially automating finance and that will potentially come through web3.

A natural question is what is web3? Unfortunately, this is one of those jargony terms that floats around, that means different things to different people. In general, it usually refers to some combination of blockchains, cryptocurrencies, or tokens. I'm going to use those two words interchangeably, and essentially the idea is to use some combination of these two to provide decentralized products and services.

The completely automated version of web3 is what is known as decentralized finance, and that shares four common characteristics, roughly. Different protocols have different characteristics. Usually these things are open-source, and open access so anyone can use them. They all have native cryptocurrencies or tokens, which is why we have so many of them floating around, and they use a blockchain. Data and programs are stored on a blockchain.

Of course, some different projects will have different combinations of these characteristics. In addition, some are restricted use, what are known as "permissioned" so only some people can use them, but in general they share these characteristics.

One of the driving features behind these protocols is that, essentially, one of the features that they all have is they allow tokenization. What does this mean? We're really familiar with tokens, dollar bills, for example, are tokens. What they do is the transfer of the dollar bill, basically transfers the value inherent in that dollar bill. In web3 and these protocols, the token is not a physical token but it's virtual, and the token is code. It can be used to transfer information, it can be used to transfer value, but it can be used to transfer anything, more generally rights and obligations, access, voting rights...anything that you want.

It's very, very different from thinking about tokens that represent only value. Tokens in this environment can represent lots of different things. The other thing to notice is that there are lots of different types of tokens. In fact, this is an area of active research, and active innovation. With dollar bills, pretty much all countries that have fiat currency have some version of a bill. Different colors, maybe, but basically it's the same thing.

In this decentralized world, there are many different types of tokens. The most commonly used one is something that is known as Ethereum Request for Comments 20, or ERC-20. These are the fungible tokens. There are some that are known as nonfungible tokens, and those are used to keep track of specific objects, or rights, or something specific. The reason why they keep track of something specific is because the metadata basically keeps the transaction history, so it's very, very easy to track the provenance.

There are, however, lots of different types of tokens. For example, one token type, ERC-3643, basically only permits transfer of ownership to people who've registered as accredited investors. The Singaporeans, in particular the Monetary Authority of Singapore, is spending a lot of time investigating different token types and how they can be used. Be aware that even though the token represents ownership and it's very easy to transfer, the types of transfers that can happen, who can own and so forth, is very, very flexible.

What is the benefit of tokenization? If we want to think about reducing economic costs, we have to understand what the costs and the benefits are of any particular technology. In the financial sector, there are very large illiquid assets. Real estate infrastructure, things that sit on banks' balance sheets, for example, are really, really large asset character types that are very difficult to trade. They're difficult to trade primarily because of opacity, because transferring ownership is complicated and the institutional parties who are most likely to set up markets are also most likely to also have an interest in the underlying asset. It's very, very difficult to set up private markets to trade these things.

What a token does is basically, it allows people to express ownership of these illiquid assets or part of them and also to transfer in a way that is actually much safer, in the sense that the person who owns the asset can transfer it. The other thing that's useful about these tokens is that because they're digital it allows one to automate things like payments.

You can have conditional characteristics that are associated with a token that can allow, say, periodic swap payments to occur. They can be used in things like trade. You can have an invoice, you can tokenize it, and then that can be basically used for factoring or anything else that will essentially allow one to unlock value without having to actually go through the physical documentation. That's another potential benefit of tokens.

The final thing about tokens that I want to emphasize is that they allow delivery against payment. If you think about transferring value, token-based payments will essentially eliminate the need for collateral in, say, nostro/vostro accounts that we see in correspondent banking relationships.

If you think about central clearing, which is typically something that we do to reduce risks in some markets, essentially everyone who participates in one of these central clearing systems has credit exposure to the other people who are participating in the system. Delivery against payment for a whole set of asset classes will essentially reduce this kind of counterparty risk, which will reduce the costs of trading these assets.

Tokens are one thing. Blockchains are typically where most tokens are used and are used to transfer value. It's important to recognize that there are many different types of blockchains. There are about 200 in operation, and there are lots of "blockchain as a service" companies, for example. Oracle has a whole suite of products that are blockchain-based.

What do these things do? They store data. They observe and verify transactions. You can see movements between addresses, and they're designed to be self-perpetuating so it will run forever. The other benefit of blockchains, in particular, is they can execute basic code. Essentially what you have is you have a shared resource that different economic agents can mutually use, without actually getting rid of or losing custody of their assets.

Generally, what are the costs and benefits of blockchains? Well, the fact that this is essentially a computation that everyone who is participating in the system can observe, it really means that transactions are verifiable and auditing is easy. Everyone is using the same shared resource. Essentially, what that does is it's like a standardization system. That means that reconciliation is really easy.

If you think about the costs of reconciliation, which depending on the type of transaction can run up to about 7 percent, this represents a reduction in costs. The other thing about blockchains that is useful is it means that economic power is not necessarily concentrated. If people are mutually sharing a platform where they are exchanging value, and everyone can observe this, it means that they don't actually have to transfer that value to an intermediary who stands between buyers and sellers. This can be done basically on a bilateral basis, so this potentially is a way of mitigating or reducing the economic power of intermediaries.

The thing to notice about blockchains is that they're stunningly inefficient when it comes to computational prowess, if you will, but that's because there's this tradeoff between having things that are observable and that everyone can essentially plug into, and high-speed computing. These are never things that you want to do complex computations on, but for basic adding, subtracting, moving value, and so forth, they're useful.

Capacity is also limited. Part of the difficulty with these systems is to design things that are robust, that can move lots of transactions quickly. Capacity typically is limited on these things. This, to some extent, is analogous to the limits on capacity that we see because of limits to collateral. There are tradeoffs.

Now, one of the things that's come out of decentralized finance that really nobody was expecting was that the use of these tokens and blockchains essentially unleashed a wave of what can only be described as economic innovation. In particular, we've seen design innovation in quite a few of the basic financial systems that we all are used to participating in.

One that's been successful in these terms is what is known as automated market makers, or decentralized exchanges. This is a completely new model of liquidity provision. At the moment in Centralized Limit Order Books, liquidity provision is essentially done by a group of professional liquidity suppliers. HFTs supply most of the liquidity in our markets. What they do is they supply liquidity, but then they also provide prices. Liquidity provision and price discovery are basically joined together.

With these decentralized exchanges, they work in a completely different way. Liquidity demand and supply are separated in these markets. Some people either demand, some people supply, and price discovery is separate from liquidity provision.

Roughly how they work, the details are sort of long and onerous, perhaps, so I'm not going to go through them. Basically, trade occurs through bilateral swap pools. At the moment, most of these bilateral swap pools are tokens on the Ethereum blockchain, but there are projects undergoing, and some successfully launched, that basically allow other types of assets to be traded in pools, in particular, currencies.

I don't want to run out of time, so let me give you another example of design innovation that is new. This is how borrowing and lending works in decentralized finance: because things are not identity-based, everything has to be collateral-based. There's a series of automated protocols that essentially use dynamic collateral management and, essentially, high-frequency floating rates to control credit risk.

How do these things work? Well, basically, the borrowing and lending is essentially done by a platform that bears no credit risk. You should think of the borrowing and lending in these kinds of markets as essentially like the repo market. This is effectively how they're organized, and what they're used for.

What happens in these? Well, borrowers and lenders both face high-frequency floating rates, and the floating rate is set as a function of essentially demand and supply. If more people want a particular asset that is available to be lent, the price drifts up, and if more people supply it the price drifts down.

"High frequency" essentially means that the lenders don't have liquidity risk, in the sense that they can always withdraw their assets. These things are overcollateralized. If it, in fact, is the case that they want to withdraw their assets, they can do it. The price, the floating rate, floats up. The borrowers now have to pay higher rates.

What this does is, it essentially at a very high frequency takes any credit risk away from the intermediary and moves it between the lenders and the borrowers. The enforcement mechanism is through liquidations that are done essentially by profit-maximizing entities, who come in and basically liquidate positions that are under-collateralized.

This just gives you some indication of a series of liquidations that were done on LINK token. These things move in very quickly and the collateral is then sold in markets. The interesting thing about this is that what this new mechanism does is it essentially moves credit risk from being bilateral, to essentially it pushes it onto the market for collateral. The risk is not gone, but basically it's transmitted in a different way.

Another source of innovation that we've seen in this area are what are known as stablecoins. Stablecoins are basically one of the primary pieces of collateral that are used in this market. They originally came about because it was really, really difficult to onramp or offramp into the crypto sphere, and so stablecoins were born.

There are a couple of different types of stablecoins. Fiat-collateralized ones act a lot like money market funds. Essentially, they're supposed to be 100 percent backed by liquid assets. The price fluctuates a little bit from where it's supposed to be pegged. This is because people are buying and trading it in marketplaces in this environment. That's where the price fluctuation comes from. The ultimate source of value is the fact that it can be redeemed. Some of it is more difficult to redeem than others.

There are also what are known as crypto-collateralized ones. These take risky assets, and essentially use the senior tranche as a means of payment. These are very much like a CDO structure. There are also algorithmic stablecoins. These are essentially like private central banks. These, of course, are backed by the belief that the project is going to last forever and the spectacular blowups have been basically these algorithmic stablecoins.

What do these stablecoins do? Why are they potentially of interest? Stablecoins can be very useful for reducing cross-border payments. There's a lot of anecdotal evidence that cross-border trade flows are being settled in stablecoins. If you think about the cost of going through the correspondent banking system, the collateral required in nostro/vostro accounts, this is completely eliminated if you can just pay with stablecoins.

The other benefit of some of these tokens and systems is essentially integrating delivery against payment. This reduces cost of collateral in central clearing mechanisms, and so this potentially reduces the cost of participating in these financial markets.

Finally, it's important to recognize that having a blockchain where everyone can see assets, and the transfer of assets. This really reduces opacity. There are a lot of markets for assets that are typically large, that are not in public-lit venues. There's a lot of information asymmetry floating around these things, so making the trading and the ownership and the transfer public and verifiable ... this is potentially going to be really important in reducing the costs of these kinds of assets. Thank you.

Moretti: Thank you very much, Christine. We're going to move to the panel, and we're going to organize the panel around three key questions that we want to explore, based on Christine's presentation. The first is digging a little bit deeper into web3: what is it and how we might transition into it. The second block of discussion will be about the benefits and use cases of web3 and decentralized finance. Finally, a final set of Q&As regarding actual risks, after which we will open to the Q&A. I see already some questions coming in. Please do come in with your questions.

Let me start on what is web3 and start with Stephen. Christine has mentioned that web3 means different things to different people, and it is often described as a new phase in the web's evolution. As a technology expert, can you help us understand a little bit more what the key features are of this concept?

Stephen Diehl: Certainly. It's important to say that web3 is not a particularly well-defined term. It's largely a term that's arisen out of marketing, and it's effectively a rebranding of crypto. Web3 presents as this narrative about a new financial system that's going to be built on blockchain technology. Unfortunately, a lot of those beliefs are really not predicated upon actual facts derived from the capacities of the technology.

In the spirit of this conference, actually, I claim that crypto is largely an "emperor has no clothes" situation. All the things that people claim crypto can do are either better done by existing solutions, or they are effectively a new form of speculation, like nth degree speculation on nothingness, or they are a gambling product. Despite 14 years of trying to experiment with this kind of technology, we have seen shockingly few actual real use cases, and the ones that we do see are either strictly inferior to things that already exist, or they just don't work out.

Crypto could disappear tomorrow, and it would be irrelevant to markets. These technologies are effectively creating a sort of digital casino in the sky, in which people can gamble on purely narrative-driven, fundamentalist assets back and forth between each other, in a kind of zero-sum closed loop.

This is a completely noneconomic system. It's a system that presents with all these grand aspirations, but when you dig down into the details about what the technology actually does and what it intends to do, those two just don't line up.

Moretti: Thank you, Stephen. Let me move to Hilary and follow up on one of the points that Christine made, which was: one of the attractive aspects of web3 and decentralized finance is that the economic power is not necessarily concentrated, but you have mentioned in the past that this is indeed one of the drivers of web3. That is, the distrust of the established firms, and not just Wall Street, but big tech. The question is, isn't it true that tech giants and venture capitalists are already eyeing web3 as an opportunity to profit? What does that mean for the future of web3?

Hilary J. Allen: Thank you. We know intuitively that web3 won't be economically decentralized because if intermediaries couldn't profit in the world of web3, why would VCs be pursuing it aggressively? They need to generate returns, and they plan to do so by funding the new tier of web3 intermediaries.

That's the short answer to your question, but the bigger part of this question is this concept of decentralization, which is so key to the narrative in this space. The key point I want to make here is that you the audience should not forget what you know. Silicon Valley likes to dazzle with its technological solutions, but it is infamous for not researching the history of the domains in which its technology will be deployed. You know more about economics than the developers of blockchains often do.

We've learned from Christine's paper about the technological decentralization of a blockchain, but the technological decentralization does nothing to guarantee economic decentralization. In other words, a system can have lots of nodes, but if someone controls a lot of the nodes then they control the system. Concentration of economic power in the crypto sphere is often worse than we see in traditional financial markets.

You know why economic power concentrates. As things scale up, the need for centralized governance, particularly to deal with unexpected events, becomes even more inescapable, and those who can profit from control will take it. We've had, in fact, the technology for decentralized organizations for hundreds of years. A general partnership could theoretically work without centralized economic control, but that simply doesn't work at scale.

The paper discusses transparency as a possible solution to these control problems, and I disagree. I don't think transparency solves these problems. If the founders own the majority of the governance tokens for a DIAT, for example, then even if I have the ability to scrutinize the project, I have to trust them because they're the ones that have the economic control.

I would say that the transparency is also overstated in this space. It's not just information that's needed for transparency. You need the ability to process that information, and opacity can be born of complexity. Here we have financial complexity being overlaid with technological complexity. Being able to untangle the code as well as the financial aspects of this is a big ask for everyone, and I don't think that this is a transparent system.

Moretti: Coming to you, Richard. We're talking now about a possible future with web3, so the question is, what would the transition look like? Are there going to be interim stages of adoption of technology like web3, and what would they look like? And what do you think would be the key obstacles to further adoption?

Richard Walker: Great. Thank you, President Bostic and staff, for inviting me to participate in the conference today, and this panel specifically. Web3 and blockchain are general purpose technologies. Web3 technologies have emerged and scaled through crypto but will extend into new uses and new use cases going forward, based on transformation of all aspects of finance.

To understand where we are now, a parallel is the internet in early days emerged and scaled through gambling and vice-based activities. While those activities persist today, the state of the technology then was weak. The belief that it could be something meaningful was low. Today, the internet has extended as a general-purpose technology to provide an amazing infrastructure for finance and commerce, and web3 is positioned to do the same.

Web3 companies have attracted $95 billion in investments between 2017 and 2022, across 4,000 companies. The functional areas covered by these companies include blockchain and platforms, core infrastructure, developer tools, financial market infrastructure, and user applications. The innovation over that five-year period was significant and provides the foundation for what's happening today in the space.

Now, God created the world in six days, but didn't have a legacy environment to support. Web2 has been in development for 25 years, and we're now in a transition to a new technology stack, governance model, and process architecture that makes up web3. However, we're likely to live in a web2 environment for some time, and the change will be incremental and gradual as we shift to a web3 tech stack.

For banking and financial markets, the disruptive thesis of web3 disintermediating parties in financial transactions is quickly fading as regulated entities engage with web3 and innovate around new processes, new workflows, and redefine the value they bring to financial intermediation. We're working on a report now about the future financial market infrastructure and have the following findings from interviewing and surveying over 50 market participants, including global banks, broker-dealers, exchanges, asset managers, and web3 natives.

The first point is that the future-state infrastructure of financial services, validated by industry executives, will likely be delivered on shared, interoperable, blockchain networks used by regulated incumbents. The second finding is that, while this feels inevitable to the industry, there's still much to do as current-state functionality standards, interoperability, and regulatory clarity is not yet there for widespread applicability and use.

However, financial services institutions aren't waiting, and distinct blockchain business networks are being delivered today, organized around use cases in jurisdictions with legal and regulatory clarity. The journey to convergence of private and public networks, ledgers, and infrastructure stand to extend the value of those private chain business networks that are being developed today.

What we conclude in the report is that the near-term engagement by financial market participants is not about the right public or private blockchain. It's certainly not about crypto. It's about the right business network ecosystem and use case to gain experience and develop assets that will lead to accelerated adoption as the market matures and doing so in a regulatory-permissible way.

Case in point: iPhone 1. Not super functional, but you had to buy iPhone 1 and continue on that journey to get to iPhone 14. If you hear that the technology doesn't scale, or other issues with the technologies not being very good, that's the way innovation cycles work. Market participants need to engage and use the technology so that it may mature rapidly and achieve its full potential in reshaping financial markets.

Moretti: Thank you, Richard. Let me follow up to this question and stay with you on the next as I open up the next block of questions, which is benefits and use cases. In your work, you are supporting major financial institutions understanding and leveraging technology for their own competitive advantage. From that vantage point, can you give us a practical example of use cases for financial services? Also, can you walk us through what would be the practical benefits for the average consumers from web3?

Walker: Sure, Marina. There are myriad use cases that we're seeing delivered in the market today in production, again largely on private networks, but there are a few examples of use cases delivered with public networks, such as EIB bond issuance on the Ethereum network. That example was a digital twin, if you will. It was a traditional bond issuance, but they did a registry on the Ethereum network for the bond. They've gone through three iterations of issuance as part of a program to understand, evolve, mature the technology, on their way to scaled use of it.

Use cases that we're seeing include cross-border payments. Cross-border payments occur today using SWIFT [Society for Worldwide Interbank Financial Telecommunication]. SWIFT provides the messaging, the messaging tells you what to do when it gets to the jurisdictional network for the settlement, but the settlement doesn't occur on that network. It still is very expensive and takes some time. As Christine noted in her paper, cross-border payments for nostro/vostro accounts do tie up collateral, which has an impact on liquidity and does carry with it counterparty credit risk. That can be eliminated through some of the use cases we're seeing in cross-border payments.

Collateral management: a company called HQLAX, out of Switzerland, is a great example on a private network. They're using the Corda network to lock collateral in place through a tokenization mechanism, and then using that token as a guarantee against the collateral so you don't have to go through the cost expense and cycle time associated with moving collateral from one custodian to another. We're seeing intraday repo trading as a use case. Previously not possible because of the myriad steps associated with posting collateral, holding that collateral, purchasing Treasuries, holding those Treasuries, liquidating Treasuries, returning that collateral to the Treasury department, or a group at an institution who's executing the repo trades. We expect that repos will likely manifest in three 8-hour windows per day, to post lazy cash to earn a return, making much more efficient use of capital management by corporate Treasuries.

Syndicated lending: to originate a syndicated loan takes about six weeks because of the back-and-forth between organizations, the documentation, the cost of verification. There is a consortium of banks that have a network today to reimagine the process, the cycle time, the collection of information, verification of information, and the ability to provide capital to organizations that need it. If you're an organization that's going to market for funding through a syndicated loan, six weeks could be a make-or-break kind of cycle time.

Seeing other examples, such as home equity on the Provenance network and mortgage origination later this year. Bain & Company, we have a company we own called Umbridge, which has developed a mortgage tokenization platform for Lennar homes. They will be tokenizing the mortgages of all new homes starting in the fourth quarter.

The express purpose of this is to accelerate the mortgage cycle time so that you can provide that mortgage more quickly and with greater assurance. Because it's against new homes, it effectively becomes a digitally native asset, because of the ability to capture, from the genesis of that physical asset, the home, all information carried throughout the life cycle of the home.

You would ask, how might this improve the experience of consumers? I would offer that consumers and institutions both will benefit from this due to increased trust, transparency and immediacy from the bond example and the closed cycle from a T+7 to a T+0, to mortgage origination, as well as the better use of assets held in households.

Christine mentioned tokenization. In the future, I could see everybody in this room having a tokenized version of cars, homes, things of value, which you could use for posting collateral just as institutions do. You could even use intraday repo types of transactions for your own capital management. There's the examples of the use cases that can be delivered with this technology, and the benefits that it could realize are great.

The last point on this, before I turn it over to my colleagues, is you might say all of that was possible before. The level of transparency, the removal of opacity, to Christine's point, the connectedness and the removal of friction, because you're not using a centralized database within an institution's four walls and exchanging data between institutions which need to map to institutionally unique, bespoke data structures, which often requires reconciliation and a protracted cycle time.

Sure, you could do that, but we haven't done it to date because of the cost and the operational risk and the diminishing return as soon as you have to invest that level using existing technology. The things we're talking about are only going to happen with new tech on a web3 stack.

Moretti: Thank you, Richard. Let me move back to the skeptics table, if I may say, and move to Stephen from a technological perspective for your views on the various benefits that Christine has covered in her paper, as well as just laid out by Richard.

Diehl: There's a bit of a disconnect between the systems that are proposed in Christine's paper, which largely describe these public networks, and what Richard is talking about, which are describing these private networks which are running inside of either consortiums or back offices inside of centralized institutions. What Richard described, these so-called private blockchains, are not a new idea. They're actually older than me, which may not actually be a bold claim. They're about 35 years old.

Private blockchains have been around for a very long time. People have been experimenting with this technology for considerable amounts. They are, for all intents and purposes, glorified databases. Automatic reconciliation, transparency, and auditability were solved problems 35 years ago.

This new generation of this technology is neither particularly novel, or in practice actually terribly useful. Everything that we've heard described today could be done on a centralized database more efficiently, more transparently, and with better reconciliation than the sort of ad hoc systems that people have been experimenting with over the last 10 years.

Private blockchains have a really, really terrible track record in industry. They don't work very well. For 14 years, people have been building on these ad hoc solutions they bought from a lot of management consultants, and there are shockingly few success stories related to these things. In fact, there's a lot of spectacular failures, the most notable being the Australian Stock Exchange which was a $100 million digital transformation project to transform the clearing system behind the Australian Stock Exchange. It was a large piece of institutional graft, basically, that resulted in a lot of failure.

That's really the pattern that we see with these solutions. They are really a solution in search of a problem. With crypto, unfortunately, "crypto" is the answer for these people for every single question. Because at the end of the day, the crypto ecosystem is about one thing, which is pushing tokens on the public. All of this sort of sophistry about digitization and streamlining are really just a kind of cover for a lot of venture capitalists to really push what are effectively unregistered security offerings on the public.

I'll echo what Chairman Gensler said yesterday. For the crypto industry, noncompliance is their business model. All of the sophistry around these applications and financial services are really in the service of selling unregistered securities, and most of these private blockchain projects are for reputation laundering and legitimization of this basically mass noncompliant offering of securities to the public.

That has serious ramifications for both safety and soundness if these systems are integrated into our collective banking system. I'm speaking to you today as just a software engineer, but I'm speaking to the bank regulators and I can't overstate this: if these systems are let to integrate with our banks, and if these unregistered security offerings are allowed to be held by banks, there are very, very serious ramifications that can happen as a result, because these assets are shares of nothingness. The value of a Bitcoin is a purely faith-based assumption. These things have no fundamentals. They trade purely based on sentiment.

This is a really dangerous proposition, because yes, it is technically possible to create an unbounded amount of financial assets out of thin air. We've tried this before, and it doesn't end well. All of this crypto, propositions about integrating these things into this new digital economy, is really in service of mass securities fraud.

That's my deep concern about what this technology is actually doing in reality. The aspirations may often be good, but the aspirations could be done with much simpler technology that's not built on noncompliant financial offerings. Everything that you could do with the blockchain, you could do more simply with a simple database.

Moretti: Thank you. Let me stay on that side of the table. Christine, I'll give you the heads up that when we're finished with the round of questions. I will task you with trying to bring those views together in your reactions. Hilary, the benefits of web3 and crypto in general are often described in aspirational terms, its value line and its potential. Is this real, aspirational and achievable, or just inaccurate?

Allen: The claims made by promoters of web3 are fundamentally unachievable. I've just explained why it's decentralization or democratization would eventuate, because technological decentralization does nothing to ensure economic decentralization. Another claim we hear, and we've heard today, is that this technology will make things more efficient. Christine acknowledged that a blockchain technology is in and of itself inherently inefficient, and I just want to stress that the consensus mechanisms used to verify transactions on a public blockchain: they must be inherently inefficient and wasteful. Otherwise, it's too easy for a bad actor to take over.

Blockchains can't process large volumes of transactions, and the delays and transaction fees can be significant at peak times. Inefficiency and waste are a feature, not a bug, and they cannot be fixed. If the technology itself is so inefficient, how can we generate any efficiencies? Honestly, a lot of this is hype and bluster.

To respond to some of the things that Christine identified in her paper, she argued that efficiency gains can arise because there's no clearing or reconciliation step. We know that that means that there's no netting. Without netting, the volume of transactions that need to be processed will far exceed what centralized financial infrastructure needs to process. That volume of transactions must be processed by infrastructure that is, as we've already established, inherently inefficient.

The discussion of instantaneous settlement also reminds me of some of the conversations that arose around FTX's proposal, in the before times, when FTX was still a...well, you know. FTX had a proposal before the CFTC [Commodity Futures Trading Commission] to clear margin products for retail participants with a nonintermediated model, and one of the benefits that FTX touted to the CFTC was this instantaneous settlement. There was a roundtable at the CFTC to discuss this, and some of the established players commented that in fact, they had been technologically capable of doing instantaneous settlement for a really long time, but they hadn't done it because it was a bad idea because instant settlement doesn't give any grace or flexibility for mistakes or liquidity crunches.

This becomes particularly critical with blockchains, where it's extremely challenging on a public blockchain, at least, to reverse mistaken or fraudulent transactions. That irreversibility is a limitation that you really need to reckon with if you're planning to build a financial system on this stuff. Another potential source of efficiency that we've heard a lot about relates to cross-border payments, and I would submit here that a significant part of any efficiency gains here are regulatory arbitrage.

Remittances, cross-border payments, they have to go through lots of KYC [Know Your Customer] checks. If those are skipped, then you can potentially improve cost and speed, although Christine does acknowledge in the paper that the on- and off-ramps for crypto typically require exchanges and bank accounts, and then these add the time and expense back into any kind of remittance or anything like that.

The real question we have to ask ourselves here is whether we want to pursue a technology where the bulk of its efficiency gains come from skirting existing regulation. Ultimately, a lot of the problems that Richard identified are not technology problems, so you can't solve them with technology. It's critical that we don't take a techno-solutionist's perspective to this. We have to identify what the problem is first and see if the technology is actually the problem before thinking that the technology will fix it.

Moretti: Thank you, Hilary. Let me stay with you as we move to the third block of the discussions, and we have already touched upon it so far, about risks. As an expert on the financial stability aspects of new financial technology, your work has explored the threats that different fintech innovations pose for our financial system. In this respect, can you cast some light on the possible risks emanating for financial services for migration to web3? As a follow up, you have written about the risks that are common between shadow banking and crypto. If you could, also let us know what your thoughts are in this area.

Allen: How much time have I got? [laughs] I wrote a paper, "De-Fi: Shadow Banking 2.0?" so I clearly have a lot of thoughts on this. In that paper, I focused on how the destabilizing forces of leverage, rigidity, and runs in traditional finance are replicated and sometimes exacerbated in blockchain-based finance. I suspect that this crowd knows runs and leverage pretty well, so I won't spend much time on them, but I will point out that the asset tokenization discussed in Christine's paper is designed to create more and more financialized assets that can be borrowed against.

We learned in 2008 that that does not always end so well. Even in 2008, though, there ultimately had to be a house somewhere for the asset to exist. In the blockchain-based finance, tokens can be created out of thin air and then borrowed against. This has been key to many of the crypto failures that we've seen over the last year or so, FTX, Terra Luna, et cetera.

I want to come back to rigidity, though, because this is maybe not so intuitive. The complex bankruptcy remote legal structures that were used to create mortgage-backed securities caused a lot of problems back in 2008 because they were inflexible to changing circumstances and that prevented needed amendments and created uncertainty about valuations. Blockchain-based finance is highly dependent on smart contracts, which Christine's paper describes as vending machines.

This is a common way of describing smart contracts, but the analogy should immediately give us pause. How often do vending machines screw up? They screw up fast, and then there's no recourse when they do. Do we really want our financial transactions to work like that? To make the limitations of smart contracts a little more concrete, in my book, Driverless Finance, I worked through what would have happened if the credit default swaps issued by AIG had taken the form of self-executing smart contracts. The short version is that it would not have ended well. Without opportunities for the parties to renegotiate, and for federal authorities to intervene, the smart contracts could have precipitated AIG's insolvency as early as 2007, to my mind.

Smart contracts cannot be programmed to verify every condition you want to check, to cater for every eventuality. It is not possible to envision all future states of the world in advance, so a smart contract can never cater for all possible eventualities. We need to keep that in mind. There will always be need for changes, for amendments, et cetera.

When you have all of this done on a blockchain with a complicated governance structure, the changes in amendments can be hard to implement in a timely fashion, creating rigidity that makes the whole system more fragile. It's also worth noting, as an aside, that you don't need a blockchain for a smart contract to run. A smart contract is just a computer program, and you can use computer programs to automate checks in a centralized system without the irreversibility problems that come from operating them on a blockchain.

Before I finish up, I do want to focus on some of the novel operational risks associated with blockchain-based finance. Christine's paper acknowledges that each blockchain is designed to be economically self-sufficient and to operate in perpetuity, but some fail, so let's talk about why they can fail. Blockchains are open-source software, and this is a kind of software that is not "set and forget." It's going to require monitoring, and it's going to require maintenance.

Coming back to our economic incentives point, who is going to invest time and effort into maintaining blockchain software's resilience? The Bitcoin blockchain, for example, currently depends on four or five people, depending on who you ask, to maintain its code. That leaves open a lot of questions. What power do these core software developers have? How are they chosen? How are they compensated? Are they compensated, and if so, by whom? Who determines when a software update is needed? Who ensures that it's developed, and who ensures that miners and validators on the blockchain will accept it? Can developers, miners, validators be counted on to get the infrastructure up and running in a timely fashion after an outage?

Do we really want our financial system to run on this kind of infrastructure? Let's compare this to the principles for financial market infrastructure, the PFMIs [Principles for Financial Market Infrastructures], which require providers of regulated financial infrastructure to, among other things, identify operational risks and develop systems, policies, and procedures, and controls to mitigate the impact of such risks, to engage in business continuity management planning, to design their systems to have adequate scalable capacity, and to ensure a higher degree of security and operational reliability. All of that would be missing from a system built on public blockchains, and we need to be very careful of those.

Moretti: Let me move to Stephen. You have in the past, and earlier today, been quite vocal in making the point that the technology is not as good as some have been portraying. What is in your mind? You've written a blog on the case against crypto, so could you elaborate? Specifically, what is the case against it, regarding some of the specific use cases mentioned in Christine's paper?

Diehl: I wrote an entire book, actually, called Popping the Crypto Bubble, that outlines basically my entire thesis that the technology is not fit for purpose, for many of the same reasons that Hilary outlined. Central to everything that Hilary has said is that in all of these public blockchain technologies, you have to trust somebody. There is no way around it.

Finance and economics are always based on trust because everything in capitalism is based on trust. Markets are based on trust, and unfortunately, with these new systems, basically, it's transferred the trust over from regulated financial institutions to a small pool of software developers that can encode logic on these unstoppable, unaudited, opaque, corruptible, and inefficient systems that are brittle, subject to errors, and subject to corruption.

Smart contracts are the product of human beings. They're the product of software engineers. I'll tell you, as a software engineer, you should not trust software engineers. We make mistakes just like everybody else, and when you make these mistakes in a way so that they cannot be reversed, there's no human in the loop, you set yourself up for a set of circumstances in which disasters can happen rapidly, and with no controls.

This is a bad idea. We don't want a more brittle financial system in which we don't have the same kind of discretionary processes by which we can reverse transactions, we can unwind problematic features. The entire foundation on which crypto and blockchain is based, is based on a set of ideological assumptions in which we should not trust people, we should trust code. That's a fundamental fallacy.

At the heart of crypto and public blockchains are a set of assumptions baked into the technology: that we should have censorship-resistant networks. That's a bad idea. That if you lose your keys, you lose your assets. That's a bad idea. That we should not have a human in the loop, when things go wrong, for customer service or for worst case scenarios. That's a bad idea. That we should have immutable ledgers in which data cannot be removed, especially if it's like publicly identifiable information. That's a bad idea.

These are all baked into the systems because the people who designed Bitcoin and Ethereum had certain ideological principles about not trusting existing financial institutions. Uunfortunately, that's just not how the real world works. We have to build systems designed to accommodate the needs of actual real people and real humans. Crypto is inherently a very antihuman technology, because of its brittle ideological assumptions that are baked into the design.

When you try to solve these things by making them private or introducing governance structures, you basically end up recreating exactly the same kind of structures we've had since joint stock companies in London in the 1700s. You're just reinventing all of the things we already know how to do, on top of a brittle, unstable foundation that doesn't even get money right. Crypto assets can never function as money. They're never going to function as a medium of exchange, a store of value, or a unit of account.

There's no way to do this without basically reinventing all of the existing structures we've built for the last 300 years of doing capitalism. Unfortunately, in the process of doing that, they're also speed-running basically all of the financial disasters of the last 200 years in a very, very short time frame. For what? To basically just reinvent payments? We already can do payments. Stablecoins, for instance, exist because crypto exchanges are largely operating in a noncompliant manner, outside of the traditional financial system, outside of our regulatory controls.

They can't get banking access for a very good reason: because the banks won't touch them, until they need this sort of dollar-derivative product by which people can basically buy casino tokens, move them to offshore tax havens, and then gamble on crypto assets. That's really what stablecoins are being used for. They are being used for gambling and regulatory arbitrage. The entire structure that is being proposed here is inherently antihuman. It's reinventing things that already exist, in a less efficient and basically unsafe way, for exactly the same reason that Hilary mentioned.

Moretti: Thank you. Let me move back on this side, to Richard, to give you an opportunity to push back with the counterarguments.

Walker: Thanks, Marina. I'm not pushing back. The current market technology, what the banks and financial market participants use today, is at end of life. The biggest risk is status quo, hearing arguments for "nothing to see here, move along. What we have is sufficient." I work with these banks. I've worked with these market participants. They've done everything they can to performance-engineer those platforms to support the level of market volume that they can. Large banks spend $10 billion a year or more on technology. My gosh, that can't be the most efficient way to deliver a technology infrastructure.

What you've heard up here over the last few minutes has been about decentralization in crypto, and my clients are in a post-crypto period. This has nothing to do with crypto. This is not about decentralization. As I said in earlier comments, the view is that there's a role for intermediaries in this future architecture based on a web3 technology stack. It's very attractive to the market participants, because of the end of life of current technology.

Because the current technology is what's opaque, corruptible, high cost, rigid, and very fragile. We've seen that in consent orders to the largest banks around resilience, when a bank has a wires outage for a couple of days and the OCC issues a consent order to fix that. That bank has to spend a billion dollars to do it on current tech. Gosh, current tech's awesome, seems to really be working well, let's not change a thing. The highest risk is status quo.

With regard to "your keys, your crypto." In the crypto world, which I'm not talking about but if I was to comment on it, code is law. In the application of a web3 technology stack by regulated institutions, it's law as law. When you tokenize a real-world asset, a sovereign currency, real estate, a bond, law as law. It's less about custodying the private keys as it is about servicing that asset like you would a traditional asset.

Many of the arguments that you may have heard, or the perspectives, I would say—actually, there's no arguing but the perspectives—there's a chasm between what's really going on with the banks who are investing in building tokenization platforms, private networks, with the expectation that those private networks will be connected through public networks to create a network of networks. Once the functional capabilities, regulatory clarity, scalability, and other factors are in place, and there's every confidence it will be, because web2 is a 25-year cycle. It took 25 years to get to a level of maturity that we enjoy today, and we have no expectation that web3 will happen any faster than that. In fact, it may take longer, particularly as you bridge and live in this web2.5 world for a period of time.

The space is still developing, but the orientation is post-crypto. The platform we believe that is going to monetize the next generation, or the next level of investment in the web3 infrastructure, is the tokenization of private assets. There is $330 trillion in private real estate in the world, there's $90 trillion in private infrastructure in the world, there's $160 trillion of private debt in the world, there's $55 trillion in private company equity. It's over $600 trillion in assets that are not in the financial system. They're highly illiquid, not very serviceable, that are ready to come into the financial system to be regulated or to be serviced by regulated financial institutions.

What will that do? That will materially expand the financial system, because when you bring those assets in and they're used as collateral for access to liquidity, you have to expand the monetary system. It's backed by real-world assets.

We've published a report around the interest in private investments. A couple of data points: wealth is held today 50-50 by institutions and individuals. However, investment in private assets today is 85 percent held by institutions. Individuals would like access to invest in these private assets for diversification, and also because private investments have returned five times what the public markets have returned over the last 20 years.

Based on our data, there's $8-12 trillion in AUM positioned by private, high net worth, ultra-high net worth, and affluent individuals to move into private investments of tokenized private assets. That $8-12 trillion will be the mechanism by which the next level of development of this technology ecosystem is monetized, the deficiencies are addressed, new capabilities are introduced, and the evolution of the tech stack and the re-expression of the financial market infrastructure will be realized.

Moretti: Thank you. Let me ask everyone a final quick question before going back to Christine for her reaction to all these points that have been made. I just would like to give you a minute and a half each to answer the following question: If you were to recommend to policymakers—we have a room full of them today—one action to mitigate the risk to consumer and financial stability emanating from web3, and/or one action to maximize its potential benefits, what would that action be? Let me start with Richard, and then Stephen and Hilary, and then I'll move back to Christine.

Walker: Thanks, Marina. As part of the report again, the interviews and surveys with 50 market participants, largest in the world, around the future financial market infrastructure, and how public ledger fits into that, we did ask in the survey: what is holding you back from moving forward? Overwhelmingly, the number one response was regulatory clarity. Number two is interoperability, and then scalability, but number one by a good measure was regulatory clarity.

There are two things that that are needed. Regulatory clarity is one, and that requires engagement by folks in this room and outside of this room, for public-private partnership with these institutions to help to create the structures, the guardrails, the clarity necessary for innovation to continue. It's happening internationally. It's happening in other jurisdictions. Generally, the largest banks in the US feel that the US is falling behind, based on what they're seeing advance in other jurisdictions around the engagement of this ecosystem. Regulatory clarity is one.

The second thing would really be cash ledger, on chain at scale. What's holding back the continued innovation—while private blockchains may have been invented 35 years ago, not all old things are bad, and it takes some time for things to develop. For instance, the light bulb was invented and it took 40 years before there were street lamps with light bulbs. We're right on schedule with 35 years ago private blockchains being invented and we're about five years away from scale adoption. That's good news.

The thing that's holding back the transformation of the securities trading, and of the DvP [delivery versus payment], and really rethinking how you trade, clear and settle, similar to what's going on with the DLT regime in the EU to prove that trade, clear, and settle is a single atomic transaction, is a currency on chain to support that. There are a number of experiments in play, but a tokenized deposit model would be a welcome gift to the growth of the ecosystem.

Moretti: Thank you, Richard. Stephen?

Diehl: My takeaway for policymakers and regulators is the term "regulatory clarity." Whoever in the crypto industry invented that term should win an Oscar for acting, because there's been regulatory clarity for 70 years. It's called the Securities Act. Every single crypto asset is a security, and not liking the framework is not the same thing as it not existing. Chairman Gensler made this extremely clear yesterday. This entire industry is entirely built on noncompliance and a lack of enforcement actions that have been brought prematurely, and there needs to be a lot more enforcement actions to bring the space into compliance.

Unfortunately, what's happened is that we've seen a series of meltdowns from the result of not having proper investor controls on these products. They're being sold directly to the retail public, and they've caused immense harm. I'm sitting here as probably one of the youngest people in this room, and, like, I'm a millennial. I graduated in 2008. The shocks that everybody in my generation felt, I feel that in my bones.

This, combined with the pandemic and the global financial crisis, is something that's a very real reality to many people in my generation, who are never going to have the same opportunities that everybody in this room has had. When I see the expansion of crypto assets, what I see is a new form of predatory investment that's being sold to the public as a way of farming what little resources people have, out of an increasingly growing precariat class. Crypto is basically promising them economic salvation, but in reality it's really just an over-levered Ponzi scheme.

I'm really concerned that if this stuff is given any kind of legitimization, allowed to grow beyond its current bounds, or to become integrated into the financial system, that we're going to see another financial crisis that makes the global financial crisis look quite tame. And all for what? What is crypto actually adding to the world, except for more gambling? These assets have no intrinsic value. The technology could always be replaced by a much simpler solution. Crypto is not the answer, and regulators need to get ahead of this so we don't have another financial crisis. My entire motivation behind crypto skepticism is caring about the little investor who is being sold a bucket of lies and hopes and dreams that just are not tethered to reality.

Moretti: Hilary, please go ahead.

Allen: If the question is what I would prescribe for web3 in the way the term is usually used, which is this idea of decentralization, I've said many times: consider a ban. Short of considering a ban, keep banking and crypto as separate as possible and aggressively enforce existing investor protection rules on the books to protect investors.

Richard's articulated a different version of web3 that I've not really heard articulated in this way before, and his version seems to be based primarily on tokenization. When I hear that the desire is regulatory clarity to allow for tokenization, that sure makes me think that tokenization is a regulatory arbitrage play, and a way of turning things that can't, under our existing laws, be treated as financial assets or traded in a particular way, into something that can be by doing an end run around the laws.

I guess my exhortation for policymakers looking at this sort of more institutional form of blockchain technology is, please make sure that the tokenization that is being explored isn't just a function of regulatory arbitrage.

I just want to conclude by saying that Richard gave us a false choice. Richard gave us a choice between the status quo and blockchain, and I am not a defender of the status quo. I've written a lot on operational risks in banks. There absolutely needs to be tech improvements, but if we're choosing from an entire universe of options to completely rebuild the technology on which our financial system is based, blockchain is the last technology we would start with.

Moretti: Thank you, Hilary. Let me give the last word to Christine. We have a lot of questions, not much time. Christine, please go ahead.

Parlour: Okay, just very briefly, as I mentioned at the beginning, web3 means different things to different people. The skeptics seem to be focusing very much on Bitcoin, and maybe Ethereum and some of the projects that are developed on those blockchains. Richard is very much focused on the back end, the financial market infrastructure. They're two very separate things. One shouldn't compare Dogecoin and all the sort of silliness around that, with realistically how one might want to change existing financial market architecture and some of the existing legacy problems that we have.

It's also useful to recognize that you want to have solutions to things that don't work well. Some things work reasonably well, and some things work really badly. Large, private assets, Richard focused on these. If they haven't been traded and if people don't quite know how to handle them, then presumably there has to be a solution just to make it easier to bring those into the financial market and to make them more liquid, and essentially to increase their value. That is something that needs a solution, and we should look at those sorts of examples, as opposed to things that work relatively well.

I could go through much more detail, but I suspect that you don't really want me to monopolize time. I just want to also focus on the technology associated with blockchain. Yes, it's absolutely true that Bitcoin is energy inefficient. It is also true that Bitcoin is controlled by a relatively small group of miners. That's fine, but a lot of the other blockchains are much more ecologically friendly, there are larger groups of people. You don't necessarily have to have a proof of work consensus protocol driving a blockchain. The permissioned ones don't do that, and they work perfectly well. Those arguments about the technology are not quite accurate.

I'd also like to point out that you can have a regulatory intervention in blockchains. It's not the case that once you have a smart contract, everyone throws up their hands and runs away screaming. As we know, assets from, say, North Korea, are blocked. People don't move them out of those wallets. As we know, Circle has frozen stablecoin wallets that have been involved in bad activities, so it's possible to have a regulatory footprint in some of these blockchains. That should be recognized.

The final point that I want to make is that the US has historically, because of deep and liquid capital markets, been the economic center or financial driver of the world. Given that we've had three distinct jurisdictions come out with frameworks to regulate tokenized assets and integrate them into their financial systems, in particular, UA-Dubai, MiCA UK, and I guess also the Monetary Authority of Singapore, they are moving ahead full steam, and probably in the US we should be very careful about other historical money centers being centers of innovation.

Moretti: Thank you, Christine. These are all good points that lay our groundwork ahead. Let me pick up some of the questions here. One of the questions goes back to the point of immediacy, smart contracts, that were just mentioned. This is a question from Valentina. I don't know where she's sitting. She said, "I guess that it will be valid also during stress time, as amplifying price spirals and volatility in financial markets," which is the point that was made. Shouldn't we be concerned about more financial instability at the door? Hilary, do you want to pick this up? Others can come in if they wish.

Allen: Sure. If you've got a smart contract running on a permissionless blockchain, it's going to be very hard to halt, very hard to undo, and I am very nervous about using it to automate transactions, particularly margin calls and things like that, because there will be systemic impacts on asset prices, et cetera. Now, Christine has just said, on a permissioned blockchain, where it's just essentially a system, and there is someone in control of it, et cetera, then it can be stopped, and that's right. I don't see how that's different from what we currently have. We automate all kinds of transactions.

What is different, and what is usually talked about in this context, is a smart contract operating on a permissionless blockchain. Automating transactions happens in lots of other systems, and that can have the consequences that you're worried about if you don't have the necessary checks in place. If you have the necessary checks in place, then that's less of an issue, although I can get into automation biases and how people don't watch the switch as much as they should.

Walker: I would use smart contracts to implement circuit breakers. It's a way to implement a governance model and a rule set, and it seems to be well-suited for something like that.

Diehl: Smart contracts running on permissioned chains are glorified databases. They're reiterations of things that we already do. That's not interesting. On permissionless chains, they are unhaltable code that has software bugs baked into it by default. Unless you can write perfect code—which, spoiler alert: as a software engineer, we can't—these are a really bad idea. You want to have a human in the loop to handle edge cases, to handle externalities, to handle aberrant conditions around these assets.

The thing that's never talked about in all of these Ethereum-based solutions is that all of these smart contracts depend on the purchasing of these tokens. It's a "pay to play"-like slot machine in which you can execute programmable logic. All of those tokens are unregistered securities offerings, so this entire premise of public, permissionless blockchains is predicated on regulatory noncompliance.

Moretti: Let me pick up this question, which is an interesting one: Bringing this discussion into the overall theme of the conference, would SVB have managed to survive if they could have accessed liquidity in T+1, either in the public markets or from the Fed discount window, and would this be a source of financial stability? I've heard this question being formulated in a slightly different but related way, which is: What would the bank run look like in a world of CBDC? I'll just open it up to anyone who would like to answer.

Walker: I'd like to offer a couple of perspectives around the regulated liability network. It's a pilot that was executed by the Federal Reserve with a group of banks, based on a design pattern from a city that published a white paper. That facility would have provided an immediate and always-on liquidity facility for SVB, because the design of that and what was piloted is every bank has a partition within a shared environment. The Fed has visibility into liabilities held at each bank. Always. It's not like it closes the window. Reserves come into that environment through a "Fed coin"-type of architecture. Reserves which can then be tokenized to create liabilities. That facility would have been consistently available to SVB. It is not a stablecoin, it's not a bare asset, it's a new architecture for a liquidity window facility from the Fed. We should see the report this week or next week. Out of that pilot that was done by the Fed, led by the New York Innovation Center.

Allen: I don't mean to be facetious, but I'm not sure if I understand the question. Is it just that it could have been bailed out quicker? Is that the concern ...

Moretti: That looks like the question. that was the first part. My add-on to that question was, would the run be even quicker? Which is already a concern that we have heard in the past couple of days, that the technology enabled fast runs.

Allen: Yes. That's an interesting question that I sort of have engaged with a little bit in the context of SVB, which is I'm not sure that it was the speed of the money leaving Silicon Valley Bank. To my understanding, there were people still queued for wires at the time that Silicon Valley Bank was put into receivership, so it's not that everybody could get their money out exactly at the same time.

There's some interesting literature. Stephen Kelly has just come out with a post about whether the speed of the run has been overstated. I do think there might be something to the speed of the rumors spreading through social media channels, and that's something I want to look into further.

If we're talking about the response, I don't think the issue was that the federal authorities didn't intervene, or were technologically unable to intervene, fast enough. It's more a question of politics: did they want to intervene? What was their decision-making? It wasn't that they didn't have the ability to push the button fast enough.

Moretti: We have a lot of questions. They're all interesting, but I'm going to be picking a couple ... maybe this one.

Walker: Let's do a rapid-fire round: one question, one person, one minute.

Moretti: All right. I'll start with you; you made the suggestion. Blockchain technology is not new. Why are there only 200 blockchains in existence? This is from Phoebe.

Walker: I received a similar question about tokenization from a large European bank. If it's possible, why haven't we already seen it? It's because of the primitives that need to be put in place for the full potential to be realized by the environment. But why are we only seeing 200? In the survey that we did, we asked banks: today, what do you use? In the future, in 10 years, what will you use: public, private or hybrid? And, how many do you expect to exist? The majority of market participants believe that there'll be 15 or fewer blockchains. You don't need 200, because it results in fragmented liquidity which can be solved through a network of networks and their operability. There'll be fewer blockchains in the future. There may be some smaller, special, fit-for-purpose blockchains. But the question around why we only see 200, I would answer: because that's enough.

Moretti: This one is for Hilary. It's going to be tough to have a brief answer, but: what should be the key elements of stablecoin regulation?

Allen: Ban them; that's quick. [laughter] Short of banning, honestly the key is to keep them away from banking. They will always be runnable, just like money market mutual funds are always runnable, but the difference is we have money market mutual funds integrated into our financial system so they matter because they're an important source for certain commercial paper markets, et cetera.

Stablecoins currently aren't providing capital. I guess they're providing capital to the US government because they're investing heavily in Treasuries, but there are other options there. They're not performing an important capital formation function, so I see no reason to do anything other than to hang them out to dry. Don't have stablecoin reserves in a bank, don't let banks issue stablecoins. That's the key way to go.

Moretti: Let me just add that there is a lot of work on this at the international level. The Basel Committee of Banking Supervision issued a standard in December precisely to define the rules for banks' exposure to crypto assets, generally including stablecoins, as well as a lot of work at the Financial Stability Board on precise regulation of crypto and stablecoins.

Allen: Just to add one more sentence: stop calling them "payment stablecoins." They are not used for payments.

Moretti: Fair enough. Last question for Stephen, from Mark: If current blockchain technology is not the path to DeFi, is such tech even possible?

Diehl: Why do we want DeFi? I mean, a lot of these things are predicated upon the assumption that this new, novel thing should be built, and it is inevitable. From everything I've seen about DeFi, it's recreating existing financial structures in a way that's inferior to what already exists.

Decentralized finance is predicated on the assumption that we could remove humans and human discretion from financial activities, and that's a false assumption. Decentralized finance, for the most part today, is a set of unregulated lending activities, unregulated securities offerings, pump-and-dump schemes, illegal security offerings. None of these things are particularly desirable solutions, and not everything that can be built should be built. A world built on a decentralized financial system will always result in the same kind of concentration, and the same kind of economic incentives and information asymmetries as our current system.

Technology is not a panacea. It cannot be used to solve human problems, because technology is just a reflection of our values. It's a reflection of exactly the same economic conditions. Left to our own devices, software engineers are capable of doing horrible things, just like the traditional financial system is. One takeaway from this entire thing today is that there is a false dichotomy between the existing financial system, which is deeply flawed for reasons that Hilary discussed, but crypto is not the solution to that. Web3 offers nothing of benefit to the future of finance.

Moretti: Let me use the last 12 seconds for the final question for Christine, which is the right way of concluding this session. What evidence should we look for, going forward, to learn which side of this debate is correct? That's a question from Todd.

Parlour: Every other industry is automated, and it would be presumably beneficial to have some degree of automation coming into finance. We can look at the adoption of some of these technologies in the back-end infrastructure in Europe and in Asia, and we can look at essentially the cost of basic financial services that are retail-facing. We can look at the cost of, say, remittances, these ridiculous taxes that people in low-income countries pay. We can look at the cost of cross-border payments for trade. These will also be benefits.

Essentially, we can just look at the cost of intermediation and see if it goes down. This would be a very reasonable way of seeing how these new technologies being adopted are going to improve financial services.

Moretti: Thank you, Christine. With that, we close this session. Thank you again, to the Federal Reserve of Atlanta, for hosting this panel. I hope you took out of it a lot more information, and perhaps not a lot of clarity on the way forward on web3, but it was a very stimulating exchange of views. And thank you, Christine, for holding the fort there in Texas.