LP-collateralized Lending
ETHCC 6 · July 2023 · 18:03
Anthony Hayot presents a risk-management framework for lending against liquidity-provider tokens. The speaker focuses on borrowing the underlying assets and assessing the collateral's behavior.
Speakers
Topics
From ETHCC 6, hosted by Ethereum France.
Transcript
foreign [Music] CPU at secured I did most of my career in the traditional finance and Industry doing trading designing instructed products for BNP for almost 10 years and I'm now co-founding secured which is a liquidity backlending protocol and we'll see in this presentation why this is the safest way to land and borrow digital assets but to start let's make a quick recap on how lending protocol works as you may know let's say that you own some digital assets we'll call token a and you want to generate some passive
income on these tokens so you can deposit them in a landing pool and they will be accessible for borrowers to borrow and pay you interest in exchange if I am a borrower I will come to this lending protocol and I will have to lock some asset we'll call token b as collateral which has a value higher than the loan that I want to make and this is a security to make sure that I'm not going to walk away with your money in order to maintain this security we're going to look at the collateral factor which is a ratio of the collateral value over the loan value and we'll make sure that this ratio stays above a certain level that will call a liquidation
threshold if at a certain point in time this collateral Factor goes below the liquidation Factor then you have a liquidation process that is triggered what happened a liquidator will step in in my place will pay back the loan and get the collateral at a discounted price the the loan will be reimbursed and the lender should be good the problem with that process there is three problem the first one is for me as a borrower my collateral is going to be sold at a discount which is a profit for a liquidator but which is a loss for me it's also a problem for the lenders because if the sales proceeds are not
enough to cover the loan repayment then it will be a loss for the depositors and it can also be a loss for the entire ecosystem because if the Liquidator decided to dump the asset B in the market and it is a imagine it's a big liquidation then it will push downwards the price of the token B and can make other liquidation happens and this is what we call cascading liquidations so liquidation is a negative event that we must try to avoid as much as possible and we can find a collateral that is a good collaterality and can help you avoid liquidation so now let's try to think what makes a
good collateral I want to limit the expected loss due to liquidation if you remember your classes of probability there is two way to limit it to to minimize an expected loss the first one is to minimize the probability of liquidation and the second is to minimize the loss upon liquidation so let's try to minimize the probability of liquidation so I have a quiz for you you have a loan in East that will move the value in dollar terms will move with the ease price this is a red line okay and now you have to choose which collateral to put in front of this loan
you have the choice of a very stable collateral let's say usdc that will remain the same value in time and this is a blue line or you can choose for example BTC who has moved that are correlated to if which one for you is the best score at all um the second one good answer yeah right the liquidation will happen when the two lines will cross when the red line will go above the blue line and clearly a stable collateral doesn't help here what you want is a correlated collateral and that makes the two lines do not cross we can verify this intuition but doing what we do usually in the traditional
Finance industry is Monte Carlo simulation I'm gonna simulate a lot of asset trajectories with different levels of correlation from -1 anti-correlated to plus one totally correlated and I will count the number of liquidation the number of times the two line cross and the number of lines the number of liquidation divided by the number of simulations gives you an estimation of the probability of liquidation all right so I did the simulation of learn in a and a token and a collateral in token B with different level of correlation and you have this curve of liquidation probability and you can see that it decrease with the correlation
between between A and B so the more the loan asset is corrected to the collateral asset so less liquidation you will have in your protocol this is exactly so minimize the probability of liquidation is exactly the same as maximize the collision between the loan assets and the collateral asset second rule minimize the loss in case of liquidation so we've seen in the first graph that when the liquidation happens a liquid editor step in and will convert the collateral B in token a to pay back the loans and this conversion is what you pay the the Liquidator for if you make this conversion very easy
very seamless with no constraint on timing no constraint on size and no impact on price then you can give a small discount to the Liquidator and we'll be fine with that so the more you make the conversion easy between the collateral asset and the loan asset the lower the impact in case of liquidation you all with me okay so now let's go back let's go to liquidity backlending so we've seen most lending protocol you're gonna borrow ease and have a collateral for example in usdc and we've seen that in case of Market volatility this can cause liquidation which is always bad
in liquidity backlending this is what we do at secured we are going to use the same asset for the loan and for the collateral the only difference is that for the collateral we're gonna use is that are deposited in a decentralized exchange so we're not going to have is exactly in collateral but we're going to have a LP token is versus something okay cclp token has very good properties this is the token that uni swap will give you if you deposit is and usdc in the East usdc pool you will you will have a LP token is usdc and with this LP token at any point of time you can go
back to UNI Swap and get back the underlying tokens let's go back now to our Monte Carlo simulation so the blue curve we know it it's when you have a learning a collateralized by token B and we know that it decrease because with correlation if I replace the collateral B by the collateral LP token a b you embed much more correlation inside your collateral and you can see that you have a decrease in the probability of liquidation by 20 so it's much better and it's even better if you don't borrow only a but you borrow token a and token B with collateral a plus b with collateral LP token a b and in that case
liquidation is the yellow curve and you can see that it becomes a very very unlikely event so you can ask me now why would I borrow a plus b with LP token a b as collateral do you have any ID exactly leverage if I have if I have brought a liquidity to a liquidity pool at Uni swap for example for a value of 100 and I want to maximize my apy I can leverage my position by using borrowed fund more is and more usdc and deposit again in the liquidity pool so I can increase my position I have new LP tokens I can come back to secured get
more ease and more usdc and do a loop actually we can do the leverage automatized directly in a insecure so you don't have to do the loop you can do that in one transaction you can top up your position times five times ten depending on on the pair and you get this this leverage position so here we can see that LP token is a very good collateral for minimizing the probability of liquidation now let's look at rule number two rule number two if you remember this is minimizing the loss in case of liquidation and by making the conversion between asset the collateral asset and the loan asset as easy as possible if you look at LP token you don't even need
a liquidator when you have to liquidate LP tokens you just take to CLP token you go back to any swap you give them the ELP token and you can withdraw East and usdc and you can pay back your two loans there is no conversion no restriction on timing you can do it immediately no impact on price which is very important because you're not causing any cascading liquidation as we've seen with other protocols so it's a very seamless a liquidation process that you can do for any size without doing any impact on the market so it's a perfect collateral for rule number one and for rule number two conclusion it's a perfect volatile
based on this design you can build very strong value proposition both for lenders and Borrowers for lenders you can build the safest lending protocol because there is no liquidation so they will have no loss on their deposit so you can give them some Capital protection actually if you dive a little bit in our white paper you will see that we add three other level of capital protection to this one but what we want to do and what we we target is really to give a product that is for risk adverse investors just want to find a safe place for their token engineering passive income
for borrowers you give them Capital efficiency meaning that if I have only 100 but I want to generate the maximum apy because I'm a professional market makers then I can use this protocol to leverage my proposition and find the right balance between the risk and the return that I want to have so it's much more adaptable to my position this protocol will be live in the coming months so if you're interested but we will building please check check our website and have a look at our white paper and if there is any question I'm happy to enter them thank you foreign
how are you calculating parameters of your protocol I mean like collateral factor and so on okay so that's a very good question actually uh the parameters on the so the liquidation Threshold at which level I will uh liquidate Your Position will be based on uh two factors we have a rating methodology to rate LP tokens based on the volatility and the liquidity of the pool if you have a very stable pair let's say usdc usdt you have a super AAA rating on the volatility side and if the pull is large you can have a AAA rating on the liquidity side in that case I can
give you like 20 times leverage so your collateral Factor will be around 105 percent okay if you have a much more volatile uh pool let's say is usdc then you will have less leverage possible it will be like time five any any other questions yeah hi um are you guys also planning to support a single asset collateral types like a tokens C tokens or deposits in like Maple Finance for example actually any token where I can get back the underlying tokens to pay back my loans work with that Design This is
really what I'm looking to find correlation plus easy liquidation if this fits in in that design I'm thinking about some tokens for strategy Volts for example it can works as well so this is providing leverage for this for this kind of strategy yes awesome we are also planning to uh so we have a test version on univ2 and univ3 is coming in a in the coming weeks as well hi I'm curious what kind of oracles do you use guys and the second question uh even if you have esense staked East for example liquidity for state is apparently less than for the East and in
this case you can still manipulate State East and you know the correlation will break uh what what do what's your perspective on that tissue uh on or I call um I'm not uh to worry about the the price of the assets because this is the same assets that I will use to pay back my loans what's important for me is what is the quantity of East and usdc I will retrieve if I liquidate my LP token so this uh I can rely on the on the data that I get directly from the Decks that will tell me at any moment any block how I can how many tokens I can retrieve by unwinding my position of
LP tokens so I don't need external Oracle to validate that yeah my question like in the past we noticed some of the major liquidation was actually caused by some kind of security incident when the underlying pool itself is broken have you considered this as an extreme test case because when that happened your LP token cannot be easily convert back to the thing you want true so there's two answers for that the first one is we're going to focus on very famous decentralized exchange that have went through all the security audits that we feel comfortable with the second answer is we're going to use
diversification meaning that if you have deposited is you will not be collateralized only but by the lp token is usdc from uni swap you will be collateralized by all glp tokens that contain is in the pair meaning that even if there is one LP token that goes best it will impact only a small portion of your loan so we'll minimize the loss in case of liquidation by doing this diversification a diversification for lenders as a lender you will be Diversified in terms of collateral but the ratio stays the same for the borrowers the borrowers will have one LP token and one Loan in face
as it so so each LP token will be uh will be managed separately for for Borrowers any other questions um oh my God I'm sorry okay man come to me if he comes back the borrow asset so is that based off of like a similar model to like Ave compound where people have to come in and lend the asset and they earn interest based on like a utilization rate model or how does that work Oh you mean the interest rate model yeah for like if I'm depositing a weath
correlated LP token and borrowing with like where does that actually come from so exactly so your interest will come from the Borrowers and uh the level the interest rates that the borrowers will pay will depend on the utilization rate but we have a different formula from avian compound we're using the fact that if the rate is interest rate is at zero percent the utilization Factor should go to 100 and if the interest rate is like 100 you think that the utilization rate should be at zero so you have two fixed Bonds on your utilization curve which is a zero one hundred and one hundred zero and then you can make some linear
interpolation of your interest rate so that you can converge to an optimal utilization why do we need this utility optimal utilization you don't want all your funds to be Borrowed by borrowers because if lenders want to go out there is no liquidity in the pool and you don't want to have zero utilization because otherwise you won't have any passive income so we are targeting some optimal utilization around 70 percent so friends if no more questions thank you thank you very much thank you foreign
[Music]