Thank you for the kind introduction, and it's a pleasure to be with all of you. Indeed, what I want to talk to you about is decentralised finance. In particular what I want to spend the 30 minutes we have together on is first to think about what is DeFi, what are the main building blocks of decentralised finance, and then some of the challenges and maybe even some of the difficulties that decentralised finance might bring about. As I'm sure many of you know, decentralised finance is a promise of a new financial architecture that indeed builds on decentralised record keeping, and therefore rather than going through centralised financial intermediaries, allows participants to access the financial markets in a decentralised and anonymous way. The promise of DeFi, or you can even say the premise that is often proposed is that decentralised finance, different from the traditional financial system, allows the elimination of traditional financial intermediaries, and with it the centralisation of financial intermediation, and the idea of this, that because access to the financial system in the DeFi world is decentralised and open, should therefore compete away unfair rents, and in particular rents that financial institutions might have. This is of course the premise, and I've put for you here, if you remember this was the founder of Celsius who famously said banks are not your friends, I think highly levered hedge funds are not your friends either, which we have all probably understood for some time, right? So, what I want to do today is first show you some of the main building blocks of decentralised finance. Here in particular we'll talk about distributed ledger technology, DLT, which is basically the blockchain. Cryptocurrencies, and in particular stablecoins, which are the main right now guard rail or on-ramp to decentralised finance, and then smart contracts, which are these self-executing computer programmes that are embedded in more modern blockchain, so not so much Bitcoin, but think about Ethereum, Cardano, Solana, where these smart contracts are used to build DeFi apps. This is what I lay out for you, and then let's discuss some of the important economic issues and pressures that come out of decentralised finance. All right, just very briefly, what is distributed ledger technology, DLT. Think of it as a way of doing bookkeeping for the financial system in a decentralised way. Traditionally, how is our financial system built up? It's basically built that people and market participants go through centralised intermediaries such as banks, exchanges, your brokers, etc., to access financial markets and financial solutions. What the intermediaries then do for you is not just the execution of certain contracts or certain services that you need, but they also provide centralised bookkeeping, ledger keeping, if you want. They're also the guard rails into the financial system, because they're normally very fine for us. KYC, Know Your Customer regulation, anti-money laundering, anti-terrorism regulation. What the blockchain or the distributed ledger technology does for us, it in a way turns this dynamic on its head, because why? In a block chain and distributed financial system what we have is now that anyone, any participant can come to the blockchain to self-host their own wallet or to participate in the system, but the verification of what is permissible and legit, transactions that is being recorded on the blockchain is now being done by what is called validators. Because we have to still, even when things are distributed, we have to make sure that the bookkeeping on the system is honest, and is basically not used for double spending, or for taking coins from participants. So, how do we do this in the blockchain technology? Here, it's basically what is at the core of what we call the trustless trust architecture. Trustless trust in the sense that we don't have centralised intermediaries whom we need to trust. The idea in decentralised finance is that the blockchain protocol itself provides the verification, in a way embeds the trust for you. How do they do this? There are two different ways or two different blockchains. There are what is called the proof of work technology, which is think of it as Bitcoin, where computers are solving a hashing game, and in a way fighting against each other for the right of embedding a block, and with it the transactions on the blockchain. In comparison you have the proof of stake protocols where it's not resources that are being brought to the table in order to keep the record keeping honest, but here proof of stake if you think about it is that the largest stakeholders, the ones who have the most coins in the system, are the ones who are becoming the validators in the system. Here the idea, the well-known economic forces behind proof of stake is that those that have the biggest stake in the system have the highest incentives to keep the system honest and keep the system running because they have the highest growing concern. What you can see why I'm pointing this out is, because I will come back to this, is that the very original idea of Satoshi Nakamoto of course was that this validation game, if you want, should stay distributed among many, many small participants on the blockchain, so that it is in a way democratised, and gets away from big block holders or big insiders who are holding a lot of the power. What we have seen is that unfortunately hasn't played out the way this idealised initial version was meant to be. Why? Because what has happened is that even in proof of work, miners have not stayed small and decentralised. In fact what we have seen is that in some of my work, in particular in other people now too, is that miners are becoming very big, and even the Bitcoin blockchain is dominated by some very large and concentrated miners. The reason is already that the incentives of the Bitcoin protocol push for centralisation, and for concentration. Why? Because number one, the hashing game is stochastic, and therefore the larger you are, the more you can self-insure against the risk of not winning the hashing game. On top of it, even in Bitcoin mining and proof of work mining there are big fixed costs. The fixed cost of setting up a rig, the fixed cost of having the human capital to be smart enough to set it up, to negotiate cheap energy prices with a provider. These are all fixed costs that then mean that the returns to being a miner or being a validator are not linear, but they are convex. All this pushes for centralisation. Why is this important? This is really important to keep in mind because it means that at the base layer, people often like to call it layer one, there are already embedded economics of pushing for centralisation of the system. We have seen this happening even right at the blockchain level. This is one of the myths I want to dispel, because while blockchain technology can have many useful applications, we don't want to run into it under a wrong premise. That's something, the first thing I want you to keep in mind. Now, I've put you some slides, you can look at some of the details in more detail, but this is what I wanted to highlight about the distributed ledger technology. The second important ingredient into decentralised finance is the type of tokens, in particular stablecoins, that are being used to trade in and out of DeFi apps. If you think about it, you can't use dollar or euros or pounds easily to trade on decentralised financial applications. Instead what people are mainly using are stablecoins. Stablecoins, similarly we need to understand how are they set up. Here we have seen there are two big distinctions in what - the type of stablecoins that we have seen. There are what is called traditional asset-backed stablecoins. Think of Tether, Circle USD, more recently Binance's stablecoin. They are basically similar to what we used to call a narrow bank. For every dollar of stablecoin issued, there is a dollar or a pound held in treasury or other liquid assets at, you say in the case of Circle, at a regulated financial institution. You see already, here the trust in the stablecoin is borrowed or inherited from the traditional financial system. From the regulatory ecosystem that provides for the bookkeeping of this quote-unquote narrow bank. In some sense, even the majority of DeFi that exists right now, very much because most of it is using stablecoins like Tether, like Circle USD, is still dependent on the regulated financial system as well. Because what we've also seen, there have been attempts of having crypto-backed stablecoins. There is DAI, which is an over-collateralised stablecoin that has indeed been relatively stable but it's very small, and then more infamously there was TerraUSD or Terra Luna, which was backed by Luna as a cryptocurrency, but because there were no true assets that were backing it, I'm sure you've all seen how it imploded over a period of days. This is basically, I just want to point out that if you are a complete crypto enthusiast, you still however live in a very hybrid world where a lot of the infrastructure, even for the DeFi ecosystem, is very much built on traditional regulatory framework and the protections of the traditional system because people are trading through these stablecoins. With that, let us now come to the last and core part of what builds the DeFi infrastructure, and that's what's often called or what is called smart contracts. What are smart contracts? Think of this as pieces of code that are embedded on the blockchain, in particular modern blockchains, again like Ethereum, Cardano, Solana, those ones, that are self-executing code. Once you agreed on a smart contract in a blockchain, the software will soon say as certain price levels are hit or certain conditions for the smart contracts are achieved, they will self-execute. This is in particular important to understand given that most of the - or actually at the moment all the DeFi apps live in permissionless and pseudonymous blockchains, meaning, blockchains, permissionless means that validators are anonymous and anyone could potentially become a validator, so therefore it means once a stablecoin has executed - sorry, a smart contract has executed, you can't stop it. Even if both parties to the smart contracts were to realise that they have made a mistake in how they set up the contract, they can't stop it, because it will self-execute. For any situation where one party realised that there's a mistake in the code or a mistake even in what they agreed upon, they cannot unilaterally stop it, and typically they unilaterally don't have access to the regulatory environment because in a permissionless and pseudonymous world you wouldn't even know whom to serve the notice to. This is one of the risks that I will come back to. Let me just say, with these smart contracts that are then self-executing on the blockchain, some of the most ubiquitous applications in DeFi that we have seen are around decentralised exchanges and DEX, borrowing and lending protocols and yield farming. Given that our time is limited I will just go very briefly to how you should think about these decentralised exchanges as an example, and then show you some of the complications or difficulties that arise I this kind of setup. What is a decentralised exchange? At its core it relies on what is called an automated market maker. The whole idea here if you think about it is that we want to, in decentralised finance we want to get away from decentralised intermediaries, including decentralised - sorry, centralised exchanges, because if you think about what happens on Coinbase, on Gemini, on Binance, any of our favourite exchanges, what a participant does is they send their coins to the exchange and they forfeit the ownership of their coin. Because the coins will go into the wallet, either the cold wallet or the hot wallet of the exchange, which ultimately means you are trusting again a centralised intermediary. The idea is in a DEX is to set up a trading system that does not rely on having to centralise all the coins or all the liquidity within one wallet or one intermediary. How does this system do this? Quite ingeniously, what happens is that we have this type of market structure which is called a liquidity pool, and a swap, as in a DEX is basically set up of two liquidity pools where you can trade or exchange a token like Ethereum against a token like Bitcoin, or any pair that you would like to trade. Now, the way this works here though is that these two liquidity pools are set up, and the ratio in which tokens are held within the two liquidity pools are defining the exchange rate between the two tokens., so the exchange rate say between Ethereum and Bitcoin. These liquidity pools then support two main operations. First, they provide liquidity itself, and then they allow the swap between the two tokens. In order to incentivise or entice participants to provide liquidity in the system, what happens is that the depositor of a token or of the token to the pools receive claims to the shares of the two pools in proportion to the two tokens. These are then called the so-called liquidity pool tokens. What happens is that the providers of liquidity receive fees through the token that they hold for the underlying, say Ethereum or Bitcoin token, that they are providing to the pool. What is different in the economics of these type of pools from a centralised market maker is that the provision of liquidity or the benefits of providing liquidity are neutralised in this system. What it means is that you get a stable price or fee for providing liquidity. You cannot strategically withhold liquidity in order to change the price here. Which some people say is a big benefit of having these type of setup of liquidity pools. However, it also - and we haven't seen this play out in terms of real stress yet, is it also changes the incentives of how liquidity is provided to the pool. Then ultimately to then wrap this up, how does a swap contract then happen? For example, like the example if Uniswap, which is one of the largest DeFi app, is basically swapping x for y, the exchange rate depends on a particular implementation of this automated market maker protocol, and it's determined, pre-determined, deterministically, what is called the bonding curve. You know at every point in time what your swap contract provides you. Then swapping x for y, or Ethereum for Bitcoin, increases the relative shares of the tokens in the two pools, and then through this bonding curve relationship lowers the price of one token versus the other. Think about the following. Even in this world that is now decentralised because people can come to it free whenever they choose, free in the sense that there's no intermediary that stops them from participating or that checks eligibility or what have you, if you look at how this market looks right now, the swap market in DeFi is extremely concentrated. You see that Uniswap holds almost 3/4ths of all the liquidity in the system, and most of the transaction in the system, and then the second biggest here is the Curve Protocol, and then all the others are much smaller. Why do I point this out? Because you can see even in a completely decentralised DeFi app, we still have concentration of where the transactions go. The reason is now not regulatory constraint or the fact that somehow regulation creates market power, it's that inherently in financial markets, because liquidity creates economies of scale and network externalities, the same economic pressures that are at play in traditional financial systems are also at play in this new financial architecture. This is something that I think people really need to keep in mind, that within financial markets, the economics of financial markets has in many applications in-built pressures for centralisation and for market power, because of massive economies of scale and scope, and because of network externalities. At least in the traditional world, how do we get around this is then through smart regulation. Trying to level the playing field for new players to come in, or to reduce the power of existing large exchanges to abuse their market power. In this type of system where the visibility to any regulator is extremely limited, given that the same economic pressures are at work, but regulation at least right now is absent from this market, this creates real problems for how we keep that system to be working for the benefits of their participants. Finally, let me skip the other applications, let me point out two more things I wanted to say about smart contracts, beyond the built-in centralisation we already talked about. The other thing that I believe we really need to solve within the world of DeFi to make it sustainable and functioning for its participants is that we have to think about the fact that in a world where contracts are self-executing, it means that contracts ex-ante have to be complete. Economic theory has a lot to say about contract incompleteness and completeness. Oliver Hart just famously won the Nobel prize last year for his research on incomplete contracts, but the work or the economics of complete contracts basically means that participants who are more informed, who are more experienced will be the ones who ex-ante have more information and therefore are able to write better contracts than participants that are maybe financially less sophisticated or more fragile. The world of self-executing smart contracts is also a world where institutions and participants who are in the know and who are experienced have vast advantages over those who don't. Given that, again, remember, it is very difficult to use the legal system exposed to remedy any issues that ex-ante were not imputed into these contracts, it really creates massive imbalances. That's why I think, I put the Dilbert thing here, but thinking about how we provide consumer financial protection in a world where the axis exposed to the legal system is quite limited. It's a concern that we shouldn't take lightly. Similar related to this, if you think about the trustless trust architecture, people often point out because everything is on the blockchain it's transparent, it's visible, there shouldn't be any hold-ups, and there should be no abuse. Some people go as far as saying this is why that system should be systemically not so risky because you can see what's being written on the blockchain. That of course, we've seen many examples, and Terra is only one of them, is only true in theory, because a system that becomes very complex and very interdependent, even if everything is on the blockchain, we have seen that even very smart people don't understand all the interdependencies, ex-ante, and are ex-post when things are self-executing and they didn't understand some of the interdependency it can create a lot of fragility and spill-over effects in the system. I just want to point out and I will close with this, is that it's also not completely trustless, because most people, even institutions, even all of us, we know a lot of finance, we might not know all the last sophisticated coding languages that are being used on blockchains, we have to then trust that the people who do the coding of the apps that are set up are not building in loopholes, mistakes, or worse, deliberate loopholes that can be exploited. It doesn't mean that just because again stuff is written on the blockchain that there is no trust needed, it's now different intermediaries whom you have to trust. If you think about it the coders who are doing coding for many financial institutions, they don't abide by fiduciary responsibility, so we haven't set up that type of regulatory ecosystem, and anyone who recently followed what happened to Wintermute just two days ago, potentially there's an issue in how their software was set up, and that provided the loophole for the hack. These are some of the things I want you to take away from this talk, that basically the fact that DeFi apps live on a system or in an ecosystem that is free access and even might be transparently visible on the blockchain doesn't in itself solve the traditional problems that we have in traditional financial systems. Thank you very much.

First question that came through, you talked a little bit about proof of work and proof of stake, how do you think the future of either of those systems links with compatibility with ESG?

Oh yes, that's a big question and an important one. I think what we have seen is that most of the efforts now are to build proof of stake blockchains, because inherently they're much more resource efficient than proof of work blockchains. Even, famously, Ethereum just did the merge to being proof of stake. I think this is probably where the future of lots of blockchain technologies will be. Especially we're sitting here in Europe, MiCA regulation pushes even more for adoption of proof of stake over proof of work. I can tell you there are obviously crypto enthusiasts who really do not like proof of stake, exactly as I was saying, because built into proof of stake is concentration. It builds on the idea that the biggest stakeholders are the ones who have the growing concern, but then they're also the ones who might have the market power.

It almost feels like we need to change the term 'decentralised finance' from…

In that word, probably yes. I personally have written some papers with my co-author Igor Makarov where we are trying to suggest a hybrid system where rather than having a completely permission system, we have a regulatory environment where yes we might have proof of stake, but we allow validators to not only do the validation of the transactions but also to work with the regulators to validate the identity of people so that we create an environment or an ecosystem that allows for some verification of KYC, AML, and other regulatory goals.

Interesting. While we have got questions on here, I do just want to give the audience the opportunity, since that was so fun, to raise their hands. Again, you can think about it in a sec. I suspect when you had the slide up that talked about automated market makers there are a few [?statutory 0:31:12.1] people in the room that were probably just slightly nervously sitting on the edge of their seats. Do you foresee a role in which - to this point on concentration, do you foresee a role in which traditional market makers are part of the digital ecosystem?

I think they are already part of the digital ecosystem. Oh, you mean of the DeFi…?

The DeFi, yes.

I think what I have seen is that there are a lot of intermediaries already are participating in DeFi, are participating on DEX as well. The issue for many traditional financial institutions right now is that because the regulatory environment especially in the US is still very much in flux and uncertain, I think participants who have more regulatory worries or are more sensitive about regulatory risk are staying outside more, and the ones who have more appetite for regulatory risk are participating more. That's something that I don't think is ever good for a financial system. You want to have a level playing field in terms of expectation about regulation so that the most competent players can come into the fray.

Okay, so I'm going to give the audience one last chance. Okay? What I will ask in that case, I'm going to throw a question from our first tour, where Pippa Malmgren mentioned central bank digital currency. That's somewhere where it's currently very uneven. I don't know whether you have any thoughts on how likely they are to develop, and where and why?

Yes, so that's a big point…

It's a talk on its own! You've got two minutes!

I have two minutes, exactly I will be short! I think central bank digital currency in the US, I think the fed has had a very careful stance on it. Recently, Darrell Duffie, together with Xavier Vives and a couple of co-authors brought out a whitepaper around CBDCs, where their suggestion is that the US should build out, and the fed in particular should build out the capability and the technology, but doesn't necessarily issue its own stablecoin. Just to be prepared and be at the edge - the cutting-edge of the technology. I think that's a good idea. I would say, if you think about it, the issues that have to be solved when we issue stablecoins is that it gives an opportunity to individuals to even more quickly run on commercial banks. We don't want a situation where people can immediately at the head of any rumour, even when it's not well-founded, pull their deposits out of banks and run into the central bank. Because if you think 2008, we had to bail out, or the US had to bail out several banks that was not a fun situation to have. You maybe do this once in a century, you don't want to do this all the time, because then the whole financial system becomes very politicised. Suddenly, whether a bank, or even whether there is a perception that a bank has more access to bailouts from the central banks, will tilt power in the market dramatically towards those banks. Because those banks are then perceived to be safer, therefore can get deposits much more cheaply, and therefore build market power. That would be really a disaster. This is why I think we need to be extremely careful in thinking about how we, if we were to implement a CBDC, how to design it well.

That's brilliant, thank you very much. That's a great answer. We're not going to get a digital sterling any time soon. Let's see, anyway. Thank you very much.