Scaling DeFi for global Finance

Thank you so much for having me here today. I'll be talking about scaling DeFi for global finance. My name is Gordon Liao. I serve as the chief economist and head of research at Circle. But I'm a longtime Uniswap friend.

First, I'll start by talking about the evolution of Uniswap, especially on v3 from an empirical research perspective, before jumping into where I see the future could be in scaling for global finance. So this is early research I had done while when I was at labs in 2022 with Dan Robinson. We looked at Uniswap's market depth and it was to our surprise that these three market depths at that time already surpassed Central Exchange's. And this is not just for long tail tokens, where obviously Uniswap dominates, but this is for blue chip tokens, even for USDC ETH type of trades. So we measured the market depth based on the liquidity profile and actually found that These market that are quite competitive to central exchanges Additionally over time we also looked at you know, not just the trading of Of crypto native tokens, but also the trading of stablecoins.

Stablecoins against stablecoins, USDC against ERC, which represent the foreign exchange market, is actually emerging and starting to become a reality on-chain, on Uniswap. So this is another paper I wrote in 2023 with some of the colleagues from Labs, and we looked at the cost of transaction, cost of doing foreign exchange transaction, inclusive of the cost of on-ramping or off-ramping, which are known to be the more expensive part of it. And we actually found that for moderate or small-sized transactions, in this graph we're showing up to $500 of transfers. Using Uniswap on layer twos, you're already getting a very competitive cost relative to traditionally going through banks. And you guys are not seeing the slides.

That's probably why people are looking around. Let's see. I might need some help. One second. All right, now we can see the slides.

So this is the slide I was showing earlier, but you guys didn't see it. Basically, the market depth on v3 has already been competitive against central exchanges, and this is from 2022, and now I think it's probably similar, if not even more so. And if we look at the development of on-chain foreign exchange market is already starting to emerge as a competitive source of liquidity for small value transfers, but although I would say that this is still a long way to go because the volume for stablecoin to stablecoin type of FX swap transactions are still quite low today. Now fast forward to today, what has emerged is with Uniswap X, you have the rise of order flow auctions, which actually greatly improved the cost for swappers, mainly because you now have not just the AMM pools, but also the ability for internalization of order flows by market makers, by solvers, and this is work by some of the lab colleagues. But order flow options also pose its own challenge.

Specifically, it poses a challenge of both competing and reducing AMM liquidity provision. So this is actually from a recent paper that I think we might have not yet released, but we can show you some part of the result which is internalization of order flow trades actually widens the spread in v3 and also it reduces TVL for liquidity pools. So this is actually becoming, even though it's helping swappers, but liquidity providers, LPs, are kind of being computed away to some extent. Moreover, competition in the DEX space has become quite intense, not only for the swap volume, but also for liquidity. What we find through this research is that DEX liquidity competition is pretty much a zero-sum game in that when other DEXes pick up more volume, it's not just picking up the volume, but it's also picking up the associated liquidity provision associated with it.

It's not necessarily expanding the pie in terms of bringing more liquidity providers. So this leads me to ask the question, are we evolving towards the end of passive LPs, the type of LP that initially started with Uniswap v2 in which someone could just deposit tokens and forget about it? From the macro data, it does seem like that it seems to be the trend that TVL has been largely stagnant, even though volume has picked back up. So, in fact, if you look at the TVL to monthly volume type of ratio, it essentially has reduced from the 1 to 5 ratio in 2022 to now 1 to 10. So that's signifying much more concentrated liquidity and much more active LP process But perhaps at a expense of more passive LPs So going back to you know what I think Texas and specifically Uniswap are Best for on one side you want to build the best trading experience, building the best trading experience would allow expansion of this space to eventually overtake tried-to-buy trading venues such as Nasdaq.

But at the same time, I think we shouldn't be lost on this idea that what's so amazing about Uniswap and DEXs generally is that it also provides a LP experience. And it is only through improving the LP experience you can eventually expand and perhaps also take over part of the traditional finance in the asset management class, so becoming more like the Black Rocks of the world. So this leads me to ask, how do we actually scale for institutional adoption in other ways? Obviously, there are many features that are still missing today. Things like privacy, permissionings, to some extent for security type of tokens, identity solutions, those are still the very much needed pieces.

But I think at the end of the day, it's really liquidity that is at the core. Without liquidity for the tokens, you can't onboard real-world asset tokens. Without liquidity for even just more liquid staking tokens, you can't provide the same type of economic security for proof-of-stake chains. So drawing on some of the research conclusions from the more recent paper, we basically looked at what are the channels to increase liquidity. Well, one channel is through more concentration.

The other channel is through more TBL increase. More concentration of liquidity that you get from lower gas fees, higher returns, and lower volatility, those are things that naturally help liquidity providers to become more active and provide more concentrated positions. But if you want to increase the TVL, which is what I was mentioning about serving this function as asset manager, that is a gain in AOM type of gain. So what would increase the TVL is not just a better transactional experience, but also there's a missing factor there of a set of tools that needs to be built. Less internalization by liquidity, less internalization of liquidity by order flow actions is one way to increase TVL, but we don't really want that because that would be kind of taking back the gains from the swappers.

Instead, I think we actually need to build active tools for the liquidity providers. So how do we do that? Well, one thing is to recognize the similarities between the liquid staking tokens and the real world asset tokens in which you want to expand. I think by examining what's happening in the liquid staking token space, you can actually learn what is becoming an issue for the real world assets as well. That's because liquidity is the constraint that transform these stake assets or yield bearing real world asset tokens into something of utility.

So for the stake assets like stake ETH, which is kind of sitting dormant there, the only way that it becomes useful other than the yield is to have fungibility and liquidity with the underlying ETH. Then you can use it as a gas token. For RAA tokens, it's actually liquidity also plays an important role. It makes the, essentially the yield bearing RAA tokens and transform it into the underlying settlement utility, which is provided by the stablecoins. So having this fungibility with stablecoin is quite core to the expansion of real assets.

Now so far we've seen that the poor liquidity of liquid staking tokens actually limited its adoption beyond what it currently is. Part of that is because without liquidity, you can't adopt liquid staking tokens as a collateral asset. Similarly, for real-world asset tokens, without liquidity, you can't adopt real-world asset tokens as collateral assets or pledges as collaterals on exchanges. So that's also limiting RWA's in a way. So the similarities between the two actually allows us to think about how do we design for both type of tokens, liquid staking tokens on one hand, and RWA tokens on the other hand.

And at the core of designing for these, solving for the liquidity for these sort of packed asset tokens, I think there's a fundamental trade-off between fungibility and resilience. If you want the most fungibility, means that you can have fixed price swapping, fixed price liquidity. And if you want to have the most fungible type of liquidity, you basically just embed the underlying liquid token into the reserve of the last liquid token. So an example of that would be, for instance, DAI or Sky incorporating USDC into their reserve. That provides you with fixed price pegs, but I would say that's not the most resilient because once you use up the reserve, then all of a sudden you end up in a pretty bad situation without any sort of liquidity.

So on the other spectrum, you have more resilient liquidity that's provided by essentially price discovery mechanisms, such as the bounding curve that for third by essentially price discovery mechanisms, such as a bonding curve that facilitate price discovery. And the usage of unlevered LPs also help in that it doesn't create too much concentration. In the middle ground, you have more concentrated liquidity as we currently see with some of the Uniswap pools, but it also suffers from the issue of if the price deviates too much out of range, then the tokens end up de-packing. So there are unique issues with both RWA type tokens and liquid staking tokens, in that if you think about these sort of tokens, they all are coupon or dividend-bearing. And in fact, if you think about all sorts of real world assets, almost all assets have some sort of dividend or coupon, which is actually quite hard to incorporate today.

Either you go through a rebasing mechanism in which you have fixed price, but variable quantity. So that's like StakeEth or the rebasing versions of some of the RAA tokens, where you had to accumulate the prices, then it's no longer the coupon-paying version of it, rather it's the accumulating version of those tokens. The problem is, AMMs can't really change the price without arbitrage, so this leads to MEV leakage through these sort of accumulating of coupon processes. So all this to say is you end up with more, worse liquidity in these tokens, the Hendo adoption. So again, I like to study liquid staking tokens because it's a way of understanding the constraints that real world asset tokens also face.

And the most widely traded liquid st token is the stake ETH, or in Unistar pool, it's the wrapped stake ETH. What you actually see is that there's quite a bit of arbitrating volume that happens every day around 12 p.m. UTC time. Why is that?

Because that's when the coupon payment happens. So imagine if you actually have a bond that's paying coupon every day at the same time, you also would observe similar type of processes. Now this is problematic because essentially the profit, the returns to these coupons are almost entirely captured by the ARBs. So you can see that the cumulative return of the wrapped steak ETH also jumps around 12 PM UTC time every day The other problem is as the price of these Accumulating coupon type of tokens increases you naturally have the LPS are actually being left behind further and further back Because the LPS do not necessarily rebalance that frequently. So this leads to a potential opening for a design space that I think v4 hooks would bring.

So one of the hooks I played around with actually helped designing through the hook incubator program that Uniswap Foundation started, and along with Atrium, is building a forward-looking dynamic fee hook that helps to capture this value of predictable price changes due to coupon or dividend repayments, as well as the predictability from these coupon rebasings to help rebalance the LP positions. And the value propositions are both for the LPs in reducing the divergence loss by providing this dynamic fee to capture these coupons, but also deepening the liquidity of swappers and the issuers alike. I see I'm slightly out of time, which is great for my last slide. So going back to this general idea of scaling DeFi for global finance, I think of DeFi as decentralized liquidity, not just automating processes of centralized intermediaries. I also think of DeFi as decentralizing balance sheet.

So that's why I have a picture here of a unicorn in a community farm, because I think of the decentralizing liquidity and balance sheet is about community farming rather than factory farming in the more centralized space of market makers. As well as leveraging finance to some degree, that if you're able to have our passive LPs to participate in these pools you can capture much greater set of these more dormant assets are just lying around that allows more decentralization liquidity but also the leveraging of finance and to achieve that defib vision I think we need to build for better LP experiences to improve the liquidity resilience, but also to provide the asset management functionalities