Trends in liquidity provisioning behavior across DEXs

Transcript

[Music] foreign [Music] hey guys uh nice to be here today um yeah so my name is Sid Shaker I'm going to talk a little bit about uh the trends that we've been seeing in terms of uh Dex liquidity provisioning uh so obviously decentralized exchanges are such a big part of our industry today a lot of crypto is driven by you know trading and a lot of this trading on chain hybrids on decentralized exchanges um and the uh the lifeblood of these exchanges is liquidity providers right so um really wanted to dig into the data a

little bit you know what kinds of liquidity provisioning behaviors are we seeing across dexas across these different ecosystems and you know share some of these findings with you guys today yeah first a bit about me um my name is Sid Shaker I currently lead the blockchain research team at coinbase um our our main focus is understanding what happens in each of these different chains uh whenever uh we coinbase onboards a new chain and also really diving into the data to understand what kind of users uh what kind of behaviors uh how people actually use these different chains and how people use these applications um on on each of these chains on each of these layer twos and part of this is to

drive an understanding of how we can actually evolve crypto and move it forward and you know we're actually actively hiring for our team at coinbase right now so if you want to chat feel free to scan this QR code or just come talk to me after but yeah some of these findings and some of this research is from our team and from the data that we've been Gathering so yeah just starting off uh just a high level overview in terms of the Dex landscape overall um you know you'll be seeing a pretty big transition from you know the early uh 2020 to where we are today in terms of the segmentation of dexas in the ecosystem uh obviously what started out as a pretty dominant run by a uniswap uh we've seen a pretty uh good segmentation

in terms of different varieties different flavors of dexes on different chains as well some interesting Innovations coming out recently of course as well with Maverick with Trader Joe pancake Swap and so on and so forth but even today if you see the orange bit at the at the right hand side in 2023 unit swap V3 is still uh kind of the dominant decks in terms of overall volume and and diving into you know pairs across all of these dexes the most dominant pair Still Remains each usdc which is the most traded pair in all of D5 and all of uh uh you know in any on-chain uh decks and uh you know as you can see in the chart over here dwarfs every other pair in terms of

transactions and all also number of unique addresses transacting on the pair by far and this is a pretty interesting Trend we've seen partially this is due to a lot of dapps using ethos DC as a kind of routing pair to move liquidity in and out of to perform their functionality but also in general it's attracted the most liquidity and the most sticky liquidity for some reason which is an interesting phenomena that we've been seeing and even when we you know look into pairs that have at least one stable coin involved with them the defining trend of this past cycle has been stablecoin Supply growth with usdc and usdt market caps increasing tremendously we still see that the usdc pair is by far and away uh the biggest

volume pair when compared to even stable stable coin pairs in isolation we had the stable Summit a few few days earlier we were just talking about this trend which is it's pretty remarkable uh the stickiness and the kind of durability of ethusdc as a pair and uh you know when you kind of look into LPS overall and dive into what kind of LPS are actually providing liquidity in these pairs um what's interesting to see is actually etsdc has a pretty diverse demographic of LPS the largest LP today on ethos DC only holds around 12 of the lp market share and then as we go down the list to a stable pair usdc usdt you know we see

that 75 percent of the lp is held by three LPS which is pretty remarkable and and then with uni eat it's a volatile uh you know altcoin pair uh we we see you know again a pretty big uh Domination by two two large LPS so you know on the one hand it's it's a positive in terms of you know in terms of decentralization of the lp base for ethusdc but it's also something this is a stat that's not often looked at when people decide to provide liquidity in pairs like what is the composition of an LP pool and how large are these LPS actually so I wanted to take a look into these distribution profiles and then pretty interestingly even like uh looking at a per day stats this is very interesting which is uh

just how many people actually provide and remove liquidity on any given day on dexas um through many of the past few months uh on Main at ethereum we haven't seen more than 100 addresses either provide or remove liquidity as you can see in this chart over here on a DEX so it's a pretty small population overall you know we have you know over several thousand people just in this conference Alone um actually maybe just a show of hands how many of you guys have provided liquidity on a decks wow so it's this whole almost this entire room is provided liquidity but the on-chain data shows something pretty interesting so um this is on Main net of course on layer two's uh you know we're seeing

more and more provision provisioning going on especially on Arbor trim today but then we hope that grows even more with the launch of base coinbase is Layer Two but this pretty interesting stats for mainnet ethereum actually and then when we actually look into the volume of liquidity provided by these hundred or so LPS um it's pretty significant you know and you refer from a few hundred million dollars to a few billion dollars a day goes in and out of these pools uh and as was mentioned earlier in the talk earlier today a lot of this liquidity is just in time liquidity liquidity that just comes in just when a swap is happening and then comes back out and is executed by pretty sophisticated Bots that are clearly running automated strategies and monitoring the price

across many different venues um and so it's a pretty staggering number in terms of how few the LPS are but also how large the size of the lp uh is being provided and removed is uh and this is kind of the state of the ecosystem today and when we think about why this may be happening what are some of the reasons um this is a great piece of research but done by ambient Finance in terms of LP profitability or LP pnl this is a pretty debated hotly debated metric uh uh and you know there's many ways of measuring this but just in terms of comparing to uh centralized exchanges there's been an ongoing discussion in terms of toxic flow toxic order flow where uh when for a given pair if price Discovery happens

on a centralized exchange the LPS on a decentralized exchange who are providing passive liquidity essentially get arbitraged uh on their liquidity preposition where say for example on a binance or a coinbase and as I may move first and then a Trader who wants to execute Arbitrage may see a price difference between an LP pool on uniswap and that on a centralized exchange and then execute that trade as Arbitrage and the lpu is providing passive liquidity um essentially gets a worse price uh because they're not capturing that best price on the centralized exchange so this is actually kind of a fundamental issue with providing on-chain only liquidity in the sense that if price Discovery happens on other venues

um you know the fees and the the returns gained by LPS actually has to surpass that in terms of uh the loss that they suffer from Arbitrage Traders and what's called toxic flow today and it's a pretty deep topic there's several papers been written on this today but it's also something that uh you have to look into the data a bit more to dig deep because there is no clear way of telling who's toxic and who's not who's arbitraging who's not one way to say this in terms of one interesting Trend here is is smaller swaps are generally better for the LPS in terms of the pools uh there as you can see in the chart over here compared to larger swaps where large dislocations happen on on liquidity

pools we often see this p l chart as you can see here uh increase with with as trade size increases and some of this is due to just the way that LP provisioning is structured today on unisoft V3 which is the largest exchange as we saw earlier fee tiers people can only provide liquidity on certain fee tiers for certain pools and as you can see here obviously the industry as a whole has evolved quite a bit to become generally a bit more efficient and with features generally trending towards the more cheaper features the lighter sections at the bottom of this chart are cheaper ranges for people to trade on but even with volatility you see here at

certain periods of volatility we've seen liquidity shrink to pretty small amounts especially in in certain uh price ranges and that results in more trades being executed on more expensive features this leads to all sorts of domino effects in terms of arbitragers in terms of LP profitability and then also in terms of just uh a normal on-chain user who's trying to do a swap getting a good price on an asset so a lot of Domino effects just in terms of how LPS can express their fee to your preferences for different assets so this is one of the potentially one of the reasons why this kind of dislocation happens and this is an interesting look in fact when you know when you think is is all hope lost for Passive LPS uh

maybe maybe not you know we saw pretty large positions these are two examples just two examples out of many of profitable LPS these are two examples of LPS who've been LP on the eth usdc pair for a whole year since July of 2022 the exact ranges they use they entered the pool when it was at you know Seventeen hundred dollars in July last year the top LP had you know 7.5 million dollars of LP and today they've made 3.5 million dollars of fee p l which is a pretty significant you know almost 50 percent uh gain um that would not have happened uh had they been in and out of the pool in terms of LP they just stayed in the pool

with those same ranges throughout the whole year and uh you know that's one case for Passive liquidity uh potentially outperforming if the asset itself has some volatility but remains somewhat stable so there's more research into this area as well um in terms of optimal liquidity provisioning but sometimes passive and longer time preferences uh can uh you know win um yeah so and then this is like this is the big Trend that we're obviously trying to focus on now which is layer twos right um coinbase is launching base there are Layer Two uh pretty soon and uh it's open for developers now on on mainnet and uh you know more and more liquidity is moving to Layer Two is now arbitrum uh other layer twos are gaining

a lot of Dex volume and this is really the metric to look at in terms of on-chain data you know a lot of people look at tvl as a big metric uh but actually tvl is a little bit of a flawed metric because when liquidity moves across different layers two layer twos um you know one side of the liquidity is held on a bridge and then another set of liquidities minted on the other side and so sometimes a lot of times tvl is kind of double counted when in reality the real metric to look at is how much liquidity is available to trade on a decks you know that's active liquidity that's stuff that you can actually utilize on that given chain so this is a big metric that we're trying to monitor in terms of just how where the money is flowing and and where users are flowing

in terms of having their liquidity available to trade um and so these are you know high level this is just an overview of the lp landscape today and you know know some key takeaways you know for for the average person it's pretty hard to to LP in general right now um I think this trend uh will only continue to grow um with more complexity more features more options more chains um so we need we need kind of like simple ways for people to to just provide liquidity uh like we used to have in in the in the unit swap V2 days and it's pretty hard to do that but at the same time there's the Counterpoint which is we need more ways of expressing preferences right we need more features

uh we need more uh um ways to change up the features when volatility occurs or different conditions occur right so I think it's a balance here I think we need to have in general more metrics available similar to what we saw here in terms of uh the composition of LP's on given pools but also simpler interfaces that can combine The Best of Both Worlds potentially by using vaults or other mechanisms to uh you know to improve this landscape so yeah that's uh that's it for me and I'll open any questions [Applause] um you had that graph that was symmetric

on the add and remove is that exactly symmetric or is that just a signifier that all liquidity was just in time liquidity over this period yeah quite a lot of it was just in Tyler it's not exactly symmetric there's some that you know overall liquidity has grown but this is also cumulative across several different pairs thank you so much for the presentation I was just wondering because there's some interesting research going on recently on on new metrics measuring performance of liquidity Provisions um do you have a way of assessing what you have just said for active versus passive portfolio management strategies so pointing for example at lever the metric of you know active rebalancing of

portfolios and so on and so on so is there a way of of of assessing this given the data that you have yeah um so there there's I think it's an emerging kind of science um of assessing it uh one way it's a bit tricky in terms of uh forecasting because uh a you have to kind of forecast what the volatility of a given pair is going to be uh and then also the the chain or the network it's in right with gas fees when you're getting in and out of positions and rebalancing the gas fees are a big component with more liquidity moving to layer twos this becomes less of an issue um but uh yeah the other aspect I didn't actually mention here a big portion of this also with variability is Mev uh when large trades are executed how much of the portion of the p l is going to

Mev Bots uh that are uh existing on these so there's a lot of factors in here um but happy to share a few papers a few research that's in this direction thanks for the presentation that's quite interesting so um I guess my question is around we see that basically critical provisioning is becoming complicated um there are many factors said and like some people are going to be more profitable than others so there is no skill involved it's not like back in the day where there was one option for the critic version um so do you think we're going to see the emergence of the are we going to recreate the asset manager from trapfi are we going to sort of have people or hedge funds and sold who are more skilled than others and hence they will be sort of creating products that people

are invented rather than automating like we have tried to do with the first gen protocols yeah I think so um I think the difference here is uh the strategies are on chain so they're public and they're auditable and you can see what what's happening uh right which is a big step up which is what we're all here for crypto is transparent right so I think there have been some early efforts with this as well with automated vaults and strategies manager strategies but there's also trade-offs there because if the strategies are open and on chain folks can front run them folks can see exactly what they're doing so it's a constant race to profitability so potentially there's a middle ground here where some of this competition happens off chain maybe a strategy runs off chain and then it's executed on chain

but definitely seeing that's that's probably the trend forward you know we have all this complexity on the back end but then the regular user is left out right if they can't participate in this ecosystem so we need that that middle ground all right thanks so much [Applause] [Music] foreign [Music]

thank you