Fireside Chat with Michael Egorov - crvUSD: Design, Peg Stability & Interest Rates vs Central Bank

hi Michael can you hear us yes I can hear now can you hear me uh yes we can you cannot see us but we can see you clearly okay good okay great um great thanks for joining us today really excited to get into this fireside chat so today we're going to be talking about uh curve USD um it's design how you think about Peg stability um and interest rates and specific specifically getting into a little bit of uh Central Bank approach comparison to how you guys have thought about this so I'm really excited to get into these questions I guess to kick us off um can you give us a brief introduction of yourself Michael and also curve youd and specifically I'd love to hear what you think is unique about its design yeah okay so it's already broad question so about myself I um founded uh Cur Finance uh in uh 2020 originally it was a stable coin um stable coin exchange and um uh by education I'm actually a physicist so uh applying uh the physics approaches in uh whatever whatever I'm building so and uh specifically um coming to curve USD um so um I think the main um the main reason why I actually decided to do it was that I wasn't very happy about uh liquidation mechanisms which exist in uh today's uh stable coins or like cdps or in landing platforms um because like when uh if let's say price of collateral goes down uh um to some critical price you actually get uh your position fully liquidated or partially liquidated and there is no kind of way to come back from that um and that's not uh algorithm how to make um essentially reversible liquidations uh where you actually don't really have to be scared all that much about liquidations well when you're close to threshold but that's kind of only one thing uh about uh CVUSD um there is another thing uh is how how it stabilizes and it actively uses uh well something we call peper which is uh maybe can can be compared with PSM but more efficient and uh Dynamic uh interest rate model which appears to be working um actually much better than I anticipated um especially when combined with uh with pack keeper so it's uh uh yeah basically appears to be um a combination of like CDP with this uh what we call soft liquidation mechanism which also Del liquidates you if price of collateral goes back up and the dynamic uh interest rate model uh with stabilization mechanisms which appear to give very very very good stability to to the stable coin so yeah I think maybe to tie that together so you're really focused on improving liquidation mechanisms to be able to have better stability of curve USD compared to other stable c yeah well I I would say it's two two different things uh liquidation are liquidation is um works on not necessarily to improve stability but more to give uh more Comfort to borrowers and um but but it actually ties into stability mechanism pretty interestingly so maybe let's dive into that stability mechanism a little bit more can you explain how you approach Peg stability more broadly I mean maybe maybe a good way to do that would be to compare to a mechanism um in in another protocol such as maker right so um uh one is our paker module which which is um essentially if it happens so that demand for stable coin is very high and uh Market tries to buy it so the price of stable coin goes tries to go above one then the P keeper actually um meants uh stable coin or well it's Prem mined but let's say it meant stable coin and uh deposits single PID in the pool um with this between this stable coin and uh some redeemable stable coin or several of them when I does that the price of uh curve USD goes down so it's like a little bit like an exchange really and the pack keeper keeps um liquidity in that pool temporarily but let's say if the price of um curve stable coin went down and it's below one then P keeper takes its liquidity and withdraws curve USD and let's say Burns it right um when it does that uh you like when you withdraw curve USD single decided the price of curve USD goes up because it's a little bit like buying curve USD and so it means that if if price of curve USD is too high it P tries to push it down if it's too low it tries to push it up but the problem is also that if um if a p keeper doesn't have any reserves let's say it's empty then you nothing can hold um Cur USD on the downside except for demand to to repay the loans but not everyone wants to repay the loans uh because maybe people have have reasons to to keep their loans up so um uh then uh to to help in this case we have um Dynamic um interest rate so um if price of Cur of USD goes really down then uh interest rate goes up and people really want to repay their loans or really don't want to borrow at least before the price of stable coin goes back up um but instead of just making this interest rate only depending on um on the price of stable coin we also made it dependent on on the reserves um of peper so if peper has lots of reserves those reserves partially back curve USD obviously um but if peper has lots of reserves then we actually want to encourage people to to borrow more to have less reserves in pack keeper because when people borrow they actually borrow curve USD and sell it for I don't know for redeemable stable coins or sell it for crypto when they leverage so um if you um make interest rate low people are incentivized to borrow more and sell and then the price uh well the price pressure is down so we want down price pressure If pepers have too many reserves but if reserves of pepers are low then the rate borrow rate is high and um like people want to buy Curve USD to repay their loans or maybe don't want to borrow so the price uh pressure is up and when the price pressure is up it also fills the P Keepers so we have this negative feedback mechanism which um like which is organized by both um this B keeper uh module which is uh providing a fast feedback and a rate uh Dynamic rate part which is providing a slower feedback and combination of these two uh feedback mechanisms actually stabilizes the um stable coin really really well okay so it sounds like the main difference there is that you're adding another element to how you determine the price which is the reserves of the peg keeper is that right yeah it's not like determining the price it's determining the borrow rate okay um yeah so yeah so it's uh it feeds into Dynamic borrow rate and it's not governance which changes which changes what the borrow rate is it's it's really the algorithm uh and and um somehow it it happens that volatility which would have been otherwise in the stable coin comes into volatility of borrow rate so borrow rate can go up and down and it does on average it's about 5% but sometimes it goes to 10% sometimes it goes to below 1% and uh yeah so um maybe uh maybe it's not uh great for some borrowers but it's actually better this way than if this volatility comes to price of the stable coin so you you need to push the volatility somewhere and when you push the volatility to borrow rate it appears to stabilize stable coin so you mentioned something that I think is pretty important there that this is automated um it's done with algorithms and not byy any vote of the Dow and I think this is an important difference to how um stable coins and particularly curve USD approach Peg stability compared to traditional currencies like uh ones backed by central banks um I mean so central banks they're not quite as automated um and I'd love to hear from your perspective if you think it's possible to create a robust and stable currency with this full automation that we see like fully reliant on you know automated market makers here or if we'll see more nuanced mechanisms come up um that allow for more integration with like existing Financial systems yeah I think it's not even like just automation it's actually um also what how exactly Central Banking currencies are created and what the what the key rate over there does compared to borrow rate and the important difference is that uh the rate which central banks uh sets it it's actually not the borrow rate necessarily it's the rate which central banks pay uh pay to um let's say to commercial Banks uh when they deposit uh their reserves into Central Bank so essentially Central Bank prints money if you deposit money in there right so uh let's say if the price of um I don't know US dollar right so the buying power of US dollar goes down right uh tries to go down and you let's say you don't want that you in uh start uh increase the key rate and start printing money to uh for banks to remove the liquidity from the market and uh you know push the price of so buying power of USD up by doing this and uh well this helps temporarily but uh you you might end up uh in um Terra kind of situation and to not have that you have the whole like manual uh control of how Central Bank Works to have this rate not very high to kind of compare it to economic growth and uh when you measure how how quickly economy grows you don't don't want to fool yourself and otherwise things things can go ugly and in some countries they do so anyway it's uh in CDP model not just for curve USD but I guess for anything it's uh it's upside down it's um you you um set the borrow rate so if you want your the um um like you you you don't print money you um to for for those who deposit you actually uh have people borrowing funds and you uh you by by changing the rate you control um returning the loans essentially right or control not borrowing more so and that uh that's a better mechanism because it's actually um more more sustainable there is no no Ponzi uh component in it like it cannot become Ponzi that's the most important thing and that's why it can be automated because like if you automate something and let's say it it's fully autonomous but um let's say if it can become a Ponzi then it's not good so you don't want to create such a system um and but if it's uh so you don't want positive feedback loops you want negative feedback loops which stabilize things um but that's just one part another part which I didn't really think of much at first is when you mix it with soft liquidation mechanism uh for curve stable coin and maybe I need to explain how it actually how this soft liquidation Works um yeah um because um um normally uh in typical CD P you have a loan or backed by let's say cryptocurrency so you have some value of cryptocurrency and you have some value of your loan and as long as value of cryptocurrency is higher than your loan then you are good right so for but and let's say if it's if if value of cryptocurrency drops CR critically low you have liquidation so what happens here if value of your collateral which is cryptocurrency ether or Bitcoin becomes critically low it actually starts converting this uh cryptocurrency to um uh to stable coin this curve USD stable coin via an amm and if the price of um collateral goes really too low then you then you have all this collateral fully converted to curve USD and you basically have let's say $100 curve USD loan backed by 110 curve USD right so um you you can actually you already sold your collateral at at this time when if it dropped by for curve USD and actually bought more curve USD than than your loan is so it's fully covered but if you have uh price of collateral going back up then this you know special amm it it starts buying back your buying back collateral and um if the price of collateral went fully fully up above liquidation price then you appear like fully converted back to your collateral so if price of of cholesterol dropped and then went up again then you you have uh your collateral back maybe with some loss but this loss is usually like nothing in comparison with losses you have in like traditional liquidation so and um this mechanism is pretty good for borrowers because let's say if price of your collateral went fully down you you actually have have it like stop loss being executed and you you have your position is in stable coin and it's fully covers the loan you can pay off the loan and uh take the the difference um because you you've been over collateralized but let's say if price of collateral went back up then you have back your collateral and uh you you can still on it and uh that's um so so that's good you don't really you can sleep well and not being worried about being liquidated while you sleep um but uh the good thing uh the most interesting thing is that when you convert your collateral to stable coin let's say your collateral goes down it's a bare market and um your uh it gets sold for stable coin right when that happens you stable coin gets bought so it price its price tries to go up pepers try to um to Mint uh care of stable coin and deposited to stable coin pools to push the price back down to one and when they do that the interest rate the borrow rate uh decreases so if it's a bare Market if let's say economy well crypto economy is going down everything is bad then borrow rate go rates go down and this encourages people to borrow more um and uh that's if you think about it you could apply it even to traditional economy imagine such such a system work working in traditional economy and Loans backed by having loans backed by I don't know company shares companies right so if economy goes down then uh interest rates go down and that's uh I guess not something you typically see so you kind of probably would support the economy by that so that's kind of an interesting thought I had but anyway this is um for now we observe that just in crypto and um yeah indeed uh when crypto went down um a little bit then our borrow rates went down and this encouraged encouraged more loans and that that was interesting yeah this is actually how central banks are designed um definitely like uh interest rates are there as a mechanism to be able to either ignite the economy or or try to um calm it down and so when you see yeah yeah but the the problem is that usually in Central Banking um the uh manipulation with the rates uh it's very indirect you're right yeah it's it's yeah they try to to stabilize the currency and uh in their case simly the goals of stabilizing the currency are in a little bit in a conflict with the goal of supporting the economy and here these goals are kind of simly aligned so that's um that's a kind of interesting observation yeah another super important thing here that I think is um really interesting for the future is that you you're basically able to borrow directly from a central bank in this case you know as a borrower you can go directly into you know curve USD and you it's almost like you would be able to borrow out directly from a central bank as a person and I think this yeah you know shortening of that um you know yeah but you don't you you still need to think about collateral like with what do you collateralize your loan so and you probably want some prime collateral uh which is uh which is not then then it would mean that it's not a person who borrows from Central Bank it would still be large companies so yeah it's still kind of otherwise it's just not not as safe not safe to back the currency by by this collateral right yeah so that's um yeah but you are short that space in between the individual and being able to interact directly with that mechanism which I think is for me super interesting about these but another thing that I wanted to touch on with you is um you mentioned you know these stable coins being over collateralized and I'm yeah really curious to hear from your perspective if you think that it's possible to sustain these stable coins into the long run being heavily over collateralized in this way just because being over collateralized means that it's more expensive to maintain these well I mean you over collateralize them by volatile currencies and um well if you have this soft liquidation mechanism it appears that you can afford quite High ltvs so the highest you can do is uh in our case is uh 89% um with ethereum or Bitcoin collateral uh and this uh like LTV doesn't mean that you like let's say if you borrow maximum and price of your collateral goes down by 1% you you don't get like liquidated instantly you you have this soft liquidation thing which uh which means that you actually can can get close to uh to to maximum um utilization of uh well to to like maximum size of the loan so you don't really have to leave that huge of a margin over there and well I probably wouldn't really go or borrow 89% but some people actually do it um I would uh I kind of found for myself like a good regim where you can make the safest possible loan or you you can actually change some parameters when you borrow and then with 71% it's um you can actually be pretty pretty safe like in in any place whether in soft liquidation or above so um you you actually so it means that you you don't really have to be be 200% or 300% over collateralized I think uh being at I don't know somewhere about 130 is uh is okay and um so you you do increase Capital efficiency this way I very sadly think we're at time I kind of uh want to round this conversation out with you before we move on to this next panel and just ask you how you see curve USD developing from here and what you think is going to be most important for it to continue to strengthen right so I guess uh the most important part for curve USD is to to get adopted as a stable coin because you know borrowing is good but you uh cannot borrow more than than it is used uh in in reality and um yeah so this adoption as a stable coin is is very very important apparently um and uh another thing is of course um good to support multiple collaterals including liquidity tokens and um maybe it's good to have lendon markets uh which use the same U sort of soft liquidation algorithm um that that kind of stuff is is probably good to explore um and um because like we only tried uh this mechanism with prime collateral like what happens if you have more volatile stable coins not necessarily backing uh sorry M volatile collaterals not necessarily back in stable coin so this is probably something we would look at uh next okay look forward to seeing that um those efforts to increase curve USD adoption um thanks so much for joining us Michael and explaining it a little bit more to us um and thanks for listening