Risk Factors in Stablecoin Design

Transcript

[Music] goodbye hello hello salute Mercy hello everyone welcome we got a lot to talk about I'm gonna promise you three things from this presentation we're going to discuss risk factors in stable coin design we are going to take a sledgehammer to the stablecoin trile Limit using data and uh we're going to pick on Consultants a bit any consultants in the audience oh there's got to be at least one don't worry don't be shy all right so if you've been in cryptocurrency for any amount of time you probably often sleep a bit like this like let's flash back all the way to 2020 and back then there's a ton of money Printing and the

narrative is that cryptocurrency was going to save us from hyperinflation Flash Forward to 2022 and the best performing investment was probably dollars if you bought cryptocurrency you might have seen your portfolio dropped like 95 percent funny enough if you bought stable coins you also might have seen your portfolio drop about 95 if you happen to pick the wrong ones so of course if you were in UST usdn it wasn't a very good year for you they think they're still floating around like five pennies or something like that but not great stable coins in my mind are the killer use case for cryptocurrency if you was anyone at the stable Summit this past weekend there we go there we go I see defy Dave walking in it was a great event um and it was the first stable coin focused

conference with a lot of big brains talking about stable coins and I'm a big fan of stable coins but they're a terrible investment because at best they'll be worth a dollar but there's still a ton of risk factors that mean it might go to zero this is just the nature of the Beast if we're going to be moving liquidity on chain with stable coins we're adding risk factors right there's no way around it it's a derivative it's like when you make a photocopy it's never the original so let's talk about some of the risk factors in stablecoin design the initial stable coin design Choice was of course tether usdt it's got the simplest model you give them a dollar you get a dollar back and you can already probably imagine if you're new to stable coins there's a million things can go wrong with this unfortunately most of the big

centralized stable coins have actually been redeeming dollars for dollars but you can imagine where they just keep the money and most of them in fact do have functions built in in which they can brick their funds you can imagine this could have killed stable coins but thankfully they've mostly been using it you know if you're selling uh you know if you're trying to use stable coins to transact with North Korea to sell nuclear secrets you probably should have your funds breaked I think we'd agree with that if you hacked a protocol like yeah it's probably a good use of bricking funds nonetheless um there's a ton of other risks that come about like tether has been subject to nothing but a Non-Stop torrent of complaints and confusion and concerns what's called Fun in the industry of fear uncertainty doubt they publish attestation reports there's a lot of people that claim the

attestation reports are not sufficient maybe they're lying in them you know we're not here to judge whether tether is lying or being honest just know that that's a risk fact but a lot of people are concerned about and then this also means there's the flip side risk factor which means that you're seeing uh zealous governments zealously trying to prosecute and attack tether so back in the day when all tether was doing essentially allegedly was putting it into the bank and earning yield on it and yields were about zero you could see where you could easily go out of business just trying to Stave off all the legal fees now tether has got a pretty good business now that you know rates are like six seven percent so they're in a really good yield on this but there's been enough questions about tether that it opened up the way for

usdc and other redeemable stable coins so of course usdc was cooking along it looked like it might catch tether and then earlier in March usdc made the mistake of keeping their money in Banks so you know what else were they going to do like they were going to like hide indoor mattress uh but by keeping in Banks uh they chose one that very popular in Silicon Valley Silicon Valley Bank and there was a weekend of Terror where usdc and everything exposed to usdc was at 90 cents there was potential risk is going to depay completely um so it's like what are we supposed to do to be safe in this like stable coins are supposed to be the safe thing to do and there was almost nothing we could do well this is why all the various risks we've seen with stable coins and their

exposure to centralized points of failure have given rise to this idea of algorithmic stable coins and now we know that the consultants in the audience have done a really good job of coming up with ways of further refining this uh you know there's endogenous exogenous we're not going to get into it just because we think it's still too premature to come up with these rigid Frameworks because there's still a ton of innovation but what we mean by this is just on-chain stable coins stable coins where usually you're lending an asset like ethereum and you're getting a stable coin back and this also is a tremendously risky or just trading off one set of risks from centralized stable coins to this other set of risks of decentralized stable coins so you can imagine with decentralized stable coins any of the favorite hacks you've seen on

the front page of wrecked news that's a risk for stable coins uh that are on chain so we're talking about oracles uh very often oracles are subject to manipulation in some cases like mochi's usdm they're just hard coding the Oracle price to one which is allowing for certain Arbitrage attacks there's governance attacks that can come from flash loans where you like accumulate all the governing power like happened to Beanstalk there's just simple smart contract risks maybe you have a bug in the code like most people even if your coder can read through all the code in either scan and still can't quite audit it so there's a ton that can go wrong but there was one interesting Trend that came from the collapse of UST which was the largest stable coin uh collapse that I've seen uh this was in 2022 triggered this horrible bear market we've been in

for some time and this was at the time of its collapse people were looking at it and saying all right looks like this like about 80 backed by Luna and around that time it became severely severely in poor fashion to have an under collateralized stable coin nearly every stable coin decided to ditch this idea that it could be under collateralized move to an over collateralized model now as we're going to see that's not necessarily like necessary but this is like a trend that's happened so enter the stablecoin trilemma because if at stable coin Summit I think nearly every presentation touched on this at least once it's become very popular as people have looked at these kind of trade-offs to say well I think we see something like a trilemma and I suspect it's Consultants because if you're a

consultant like anyone at McKinsey can come up with a 2d Gantt chart but if you have a full-fledged trilemma even if it's backed by slimsy data you can probably get promoted so with the concept of the stablecoin triloma if you look at the usdc's the redeemable stable coins of the world they're completely not decentralized but you get the other two prongs you get Capital efficiency and you get a Peg and then they looked at the on-chain stuff and said well it's not fully Capital efficient what this means is that with redeemable stable coins it's usually you trade a dollar for a dollar on chain you want to have like a little bit more collateral backing it so they tend to be over collateralized Capital efficiency is the inverse of the collateralization ratio so if you're two to one back you're only 50 Capital efficient if you're forward

one backed you're 25 Capital efficient Etc so any of the lending based coins that are over collateralized they lose in capital efficiency but they get decentralization and they get pegged and of course dokwan's uh UST didn't do so good with the peg but claim the other two so we were really interested in taking a look at this because we were like really interested in looking at how we could break this down um we found one paper that was pretty interesting they presented a very different uh stablecoin trilemma with some different uh factors and we definitely recommend people read it but we know it's after lunch so we're not gonna like get into this much math don't worry um just needless to say like there are ways attacking this analytically and rigorously using data I'm a huge geek so I was pretty interested in if I there

was similar stuff for this aforementioned stablecoin trilemma so I went looking for data on the stablecoin trilemma and here's what I found there hasn't been a lot of rigorous attacks on is this actually valid so for the remainder of the presentation we're just going to dive in I'm going to present a very condensed version of all the research I've done on the stablecoin trilemma it's all available on my blog which you can find at the end of the presentation but let's dive into these three prongs of the stable coin Trail Emma most people argue that the peg is probably the most important because if your stable coin is worth zero it's lost its Peg it's useless there's actually a couple ways of attacking this question of the peg uh the first is just the binary is it one or is it zero and this is of course what everyone cares about

but when I looked at the data there wasn't a ton of good keys out of it the full explanation is on the blog but basically like if your coins gone to zero because like someone had a secret key where they could steal all the money it doesn't really matter whether it was Capital efficient or not it's because someone stole the money right if um there was a design design bug in the protocol it doesn't matter if it was decentralized or not it was just a bug and maybe a better implementation it would have been successful so this binary one to zero wasn't very informative when I was looking at the data behind the peg but what was kind of more informative was looking at just how closely if it is maintaining Peg does it keep to the peg so I specifically picked a window to put everything on an even

playing field since the launch of curve USD a couple of months ago up until slides were due like a week ago so unfortunately I didn't get going there which was just released this weekend but it gave us enough data on a kind of an even Time Field like not so good for TUSD because TUSD has been undergoing a lot of fun but we missed like the usdc dpeg weekend and you can see that if you just plot the prices based on coin gecko like you can kind of tell visually which one's keeping the Peg and which one's not there's a few ways to attack it from here you can use standard deviation is what I used there's also ways of using like mean average error a two one dollar which is useful for like liquidy USD which tends to float above a dollar and some people say that's not so bad or you

can use mean average lower mean average error which is how often does it just go below that dollar Peg but for various reasons like I stuck with standard deviation because if you just buy something a dollar ten on the open market it goes down to a dollar you still like lost 90 of your money um and this notice these are all backwards looking if you're interested in forward-looking can you predict a one to zero d-peg there's some interesting research that's been done by xenophon Labs using um Shannon entropy in curved trading pools to predict future depegs so this might actually be the most important research that's being done in the space which is the forward-looking how do you actually predict the d-peg we're just looking at wallets pegged how closely is it staying at Peg so using this we were able to come up with a list of results

um you know some of the ones you might expect the law the redeemable ones are very good at holding one to one to Peg whereas some of the more decentralized ones like have the looser Peg and now that we've got the data we can start doing things like plotting at Android analytically rigorously um so throughout we're going to just make note of which of these factors actually contribute to a big market cap here in the bottom right you can see usdc and usdt did pretty good we don't know if that's the tail wagging the dog did they become big because they're good at holding their Peg or are they big enough that helps them hold their Peg that's a question for Plato a couple quick notes while we're on the subject of Peg if you're on chain there's actually no way you can calculate the

price of a dollar right dollars only exist off chain everything on chain is a derivative of dollar so the question of how do you calculate what a dollar is worth on chain is not trivial curve usd's approach is to take four different coins uh stable coins see what their price is roughly on chain and then essentially weighted based on how close it is to pegs so like TUSD during this period was going a little wild but curve USD still kept it tighter Peg than the TUSD and some of the other stable coins and then finally like you might say that Peg's the most important factor but there's plenty of cases where like you don't necessarily care about holding a strict Pig fpi frax price index is a stable coin that's pegged to inflation so it's good if it's going up right but you might argue okay it's going up but

it's still pegged to Dollar Plus inflation but there are examples of stable coins that are sacrificing Peg Aladdin Dow's f of x protocol is a pretty interesting stable coin-ish coin that's launching they're essentially taking ethereum and cleaving it into F eth which is the trying to take ethereum but about 10 of its volatility and then putting the rest of the volatility into a high volatility X eth which functions as a leveraged lung uh this um their white paper basically lays out the case that the f eth is not strictly a stable coin but it's stable enough that you could do things like plan for employee expenses plan for rent but still be exposed to some of the upside if the market does go up and then of course LUSD as mentioned

before it was more focused on kind of keeping between 90 cents and a dollar ten it's mostly stayed above a dollar because uh basically with like the it's as a lending based stable coin when people are like rushing to repay it there's a huge demand for LUSD that kind of pushes the the peg above a dollar so most people who hold LUSD don't mind but LUSD was basically saying we're going to sacrifice the Peg and we're going to focus on the second Peg the second prong of the stable coin Trail limit decentralization this was the one that was maybe toughest to quantify because I couldn't even find a definition that everyone agreed on it's one of those things that people are like I can kind of recognize it when I see it so I just used that I used my Twitter account to get a statistically

significant sample of votes from people saying on a scale of centralized decentralized how do you consider this coin to be and the results were pretty good my audience like tends to be biased towards curve so I put curve USD above LUSD it should probably be flipped but at least they're in the same reasonable tier and of course like the usdcs and usdds of the world or were extremely centralized given this data now we can for example see how does Peg correlate against decentralization in the bottom left corner you can see that all the redeemable stable coins they are not very good on sen on decentralization but they are you know pretty good at uh holding their Peg on the right hand side you see more of the redeemable stable coins they did 50 50 right 50 of them like their pegs

pretty wild but 50 of them are about as stable as the redeemable stable coins and then of course the stable coins that are living off chain have access to the most liquidity get the most market cap the usdts usdcs finally we're going to look at the third prong which is capital efficiency and this is where we suspect we shouldn't see most of the trade-offs because most of these uh stable coins are trying to basically like become as over collateralized as possible as a way of like staying more stable uh fortunately most of the stable coins did at least provide numbers on their collateral backing so we were able to you know use these numbers now who knows if they're accurate like in some cases like they've been called into question but this also kind of mostly maps to what you would

expect you know you see the redeemable stable coins don't have a huge collateral backing they're much more efficient and then you see like susd is really focused on uh they have like five to one collateralization ratio at the moment LUSD is 2.5 to 1. so this is the cleanest correlation chart we've seen amongst any of these prongs of the stablecoin trilemma down on the bottom right you see all the redeemable stable coins that generally have high Capital efficiency they also keep the tightest Peg and in the upper left you see the susds now usds the highly overcloudwise ones have the loosest Peg and this you might be asking like isn't this going the wrong direction because weren't they becoming over collateralized to protect their Peg but it seems like the data's showing the

opposite and then on the question of efficiency versus decentralization we got um again it's the design choices the way I like to look at it in the bottom right you see that most of the collateralized redeemable stable coins are of course Very Capital efficient but usdd decided to be over collateralized they said all right we're going to take two dollars for every dollar and now they can take that and put and earn pretty good yield that might be a good business model similarly amongst the decentralized stable coins like you do see some of them are very Capital inefficient and some of them push very close to Capital efficiency that's worth noting that frax was under collateralized about like like five percent or so until it became out of fashion and it functioned and held its Peg very well so I do consider it to

be mostly a design choice and then of course the ones that take a dollar for a dollar get the highest market cap all right let's put it all together we've got now three dimensions and we're dealing with two dimensional slides how can we actually like visualize what's going on here how can we like present the stable coin try dilemma in a way that like a dumb person like me can understand so we're going to try and do a three-dimensional chart here uh so what's going on here is we've taken all the data we've collected and we've basically ranked in a scale where zero is scoring the absolute best and one is the worst in other words if it's closest to where the three lines intersected the origin that's a absolutely dream stable coin it's keeping Peg it's decentralized it's got efficiency if we're positing a trilemma

we would expect that we would see the coins fall into one of these extremes so for example a cbdc would be completely centralized but perfect Peg and perfect Capital efficiency similarly at the top it's losing out on Peg and at the right it's losing out on Capital efficiency makes sense right now for positing that there's going to be some sort of trilemma we can sort of Imagine like some sort of mesh some sort of boundary Beyond which none of these stable coins can pass so in other words it's going to be entirely possible that you could have a stable coin that is worse on some of these factors like this red dot here it's worse on efficiency decentralization npeg so that's theoretically possible but you're not going to be able to get one that's close

to the origin and you can debate that what the mesh should look like the boundary maybe it's curved I did here's a flat screen the important thing here is that like they're for there to be a tri limit you can't actually get something too close to the origin so here's what it looks like when we actually look at our experimental data we'll start with usdc which is pretty close to a cbdc it's very Capital efficient it's very good at keeping Peg other than that weekend in March but it's not so good on decentralization indeed all redeemable stable coins kind of cluster here around this corner so similarly we can take a look and we can very easily find examples of other stable coins that would clearly fall outside of this mesh but we can also find plenty of examples of stable coins that fall very much

inside this mesh these are stable coins that very close to the origin die I think is a very good example and that it didn't give up too much in the way of Peg relative to the redeemable stable coins it's better than some of them it didn't really give up that much in the way of capital efficiency but the audience that was following me said that they're much more decentralized similarly curve USD it's about 80 Capital efficient it could be about 90 if everyone was doing the Max D gen leverage but it's much stronger on Peg than even some redeemable stable coins and it's um very good in decentralization so I think that what this says to me is not that the stable coin trial element necessarily needs to be thrown out but it's a little bit reductive another way of looking at it is you could do spider graphs for

everything we just did here we're reversing it so one is where it ranks really good usdc is really good at keeping Peg really good efficiency but not so good at decentralization if this was a true stablecoin trial element you would probably assume that you can't draw on much bigger triangle than this but we can like as mentioned frax is really good at keeping Peg it was really good at decentralization compared to some of these and in even for time's been more Capital efficient than some of the Alternatives there's a lot of cases where you can actually get some pretty good improvements on this so I would posit that the stable coin triloma is basically just like a mental filter that we're putting on ourselves where it used to be the case that you said it's

impossible for a human to run a five minute mile and then someone did and it was impossible to run a four minute 30 mile and then someone did if you're accepting the stable coin trilemma you're basically just limiting the what you're saying you're going to be capable of doing with your stable coin Your Design all of these involve risk and reward trade-offs right we said from the outset that there is no free lunch if you're centralized there's one set of risks if you're decentralized there's a different set of risks so none of these are completely erasing risks but we can still look for improvements on all these Dimensions one really good exam sample LUSD they have a stock at the stablecoin summit they're also speaking here at etcc about liquidy V2 in which they're also rejecting the stable coin trilemma they said we figured out a way to

improve on all of these fronts and we'd encourage all of you to do the same but that being said like we know that Consultants don't get paid if you don't have like a nice trilemma so we're going to propose let's just put this one aside for a second we're going to propose a new trilemma and this one is just for the Consultants we're going to call it the big three consulting firms trilever which is that if you've worked at one of the consulting firms will have you in crypto but if you've worked at two of them you're probably like living in Powerpoints and you should stay out of crypto thank you everybody this has been risk factors in stable coin design I think we have a time for questions we have time for questions hi I have a couple questions so uh kind

of like unrelated the first is when are you integrating llama in curve Sim Winery what sorry integrating llama in uh curve Sim oh that's a good question I don't know the answer to that um I will check on that for you do you have a way of following up foreign it's like at 1.5 percent right I think it's about that yeah and what do you think is going to be like its impact on the overall lending Market because right now the rates for like usdc and all that stuff are like three percent and the DSR is at three percent and something yeah that's really like being minted at like yeah it's a really good question I guess I wouldn't be able to offer a good prediction on that to be honest

um I'm still trying to figure out which stable coins gonna hit two dollars first hmm um yeah thanks that's a really interesting talk um I wonder there's like there's one extra element here to add to stablecoin adoption and and growth rate one of them is like the network effect so even if we come up with like a better stable coin here we're like okay disabled is great it's ten percent better than tethering all of

these Dimensions but there's like a hundred billion tethers out there that you can use and uh do whatever you want right away so I think to overcome that your the design of the stable community has to be much better or have like a really um sort of aggressive go to market strategy so I was wondering in the research were there any sort of things that you sold work that would like actually have a chance of taking down uh you know essentially uh 80 market cap uh coin down to so you know like how do you compete in that market right where the network effects are so hard yeah it's a really good question and I do agree that uh one of the problems with this idea of a trilemma is that there's a ton of other factors that people aren't considering I think the utility

of a stablecoin how many chains does it live on is it easy to buy real world assets using the stable coin or not um I think there's a lot of things you could add into it um so I didn't get a chance to quantify all that but I actually think that that would probably be like a much more interesting thing to look at foreign so maybe I'm just not up to date too much but um I was wondering why does curve USD have much higher Capital efficiency than LUSD for example so that's just a design decision and you also users can set essentially the when they take out a loan using curve USD they can choose kind of how up to the edge they'd like to get the best you can do is like 90 or so LTV which allows for

like 9x leverage and that's like an issue with lending based stable coins you can't actually get one to one and that's a lot where a lot of the concept of this trade-off came from because if you could do one-to-one they'll offer like basically infinite leverage right so it doesn't actually work to get like right up to the edge and most people don't feel comfortable at 90 so the typical average that people have taken out so far is like 80 percent hi can we hear your definition of decentralization of like oh yeah it's my definition of decentralization so this is one of those things I recognize when I see it I've heard it be uh like the the closest to a definition of

decentralization I've seen is the like number of unique wallets that it would take to execute a majority change on a protocol if it's actually on chain and then the on off chain ones wouldn't count because that would just be subject to off chain but by like stable coin like uh when you asked your in Twitter like uh which which stable coins were more decentralized like did you get any other questions like what do you mean by this question or people just leaving questions or comments regarding the first yeah thankfully no one pushed back on it so the text of the question was simply how decentralized do you consider blank and you know it's open to interpretation but you can see where like there's some interesting votes

where like um you know usdc is as centralized as it gets but still people were were like coming up with like shades of like how decentralized is usdc versus usdt like in my like book it's an easy vote but now we can at least like kind of fractionally say that people think tether is slightly more decentralized for what it's worth than usdc all right what would you consider more decentralized a stable coin unseen stable coin that is available on every L2 or

something like LUSD that has let's say it is much more difficult to implement or change the protocol level but that would only be available right so I think I need a bit more data to answer the hypothetical course um because for all I know that one that's across multiple chains maybe that's got one signer that can execute everything right in which case would not be at all decentralized um in my opinion like liquidy is the gold standard for decentralization like they're so decentralized they don't even operate their own front end they allow other people to operate front ends and take whatever fee they would like so like that way there's no like single place that agents could come and gun down if they wanted to like take down the

website so liquidy in my opinion is what you should look to as the gold star of decentralization all right thank you everybody [Applause] [Music]