Cross-Asset Collateral: Managing Risk and Solvency - Panel

Today we have like I think very interesting topic with uh best people who can like describe their experience in it and uh let's start with a short way of introduction. Further let's start with you what you building and like how you related to cross asset collateral risk.

Okay thanks. Hey everyone I'm uh further one of co-founders of resolve. Essentially what we are making is uh tkenized uh strategies like preliminary right now delta neutral strategies in form of stable coin but not only like around like crypto assets in the first place just to keep it very short

uh dh

um hi everyone I am dh at fluid fluid is a liquidity protocol on top of which we are building lending market decks and many more protocols in the future. Hi everybody. I'm I'm Allan uh founder of Covenant. We're basically creating creating a liquid credit market where anybody can mint and swap uh leverage and yield tokens uh backed by any collateral where the collateral itself is the liquidity of the market and on top of that build other stuff.

Hey, um I'm Monet. I'm the head of strategy at Spark. Uh we're a subdow under the sky ecosystem and uh we're trying to build like the liquidity and yield engine of the uh the internet.

Okay. So guys, what do you think we can learn from the past deep events and like

which one?

Yeah. What what exactly?

It was a lot of Yeah. But the last ones like uh what do you think what the like the main cause and like how we can uh change our ecosystem to be more resilient to it?

That's that's a fun question like literally discussing with with every second guy during I don't know last few days like everyone is happy with I don't know higher yields etc until it is blown up. So uh we for example at DOL we are pretty risk averse guys for example and in good market when everything is fine like everyone is asking hey guys why why we need you for example like while we can go to like project X I will not probably name specific projects right now uh well such projects like tend to die over time maybe but again there are like a big range of different kind of risks again just not take microphone for too long.

Um in my opinion it's very important to understand where you are depositing or what assets you accept as collateral on the lending market for example. Uh you have to be fully aware I mean fully fully insure that you can liquidate this position if if needed. And uh yeah in recent in recent times there were assets that were not backed by anything on chain. There was some sort of off-chain trust me uh strategy um which nobody could have verified and this assets they are very risky and of course like on the lending markets you cannot really have them. Um unfortunately there were lending markets where users were affected by you know by lending into the this assets or sometimes they were not they were not even aware of that because the asset that they had this collateral was exposed to some another asset and that asset was exposed to something else.

So yeah you just have to look into what you're using and what is the underlying of that asset. No, I think that's that's right. And I think in many ways we're going through uh kind of like the history of finance itself, right? A lot of stable coin DPEGs right now are equivalent to bank runs in the past. Uh and the answer is kind of like in finance itself is like diversification of collateral.

Uh moving away from volatile collateral uh like you can't even in finance volatile collateral you leave that to hedge funds and banks end up just using less volatile collateral real world assets and whatnot. And so I think we'll see that same uh trend where uh collateral is more transparent, more diversified and much much lower volatility. Even I would advocate for stable coins with high yield as collateral for for mass adoption stable coins. Yeah, I think there's there's been a few like big learnings from from the whole stream in uh DUSD fiasco, but um maybe one of the biggest that jumps out for me is uh just how there was a lot of uh sort of like drift of the strategy and the mandate of the different vault curators that were ended up being exposed to stream. I think I've spoken with a few people who um you know, luckily they were monitoring these vaults pretty closely.

So when they saw these assets being onboarded, they they pulled their funds and they were able to kind of avoid losses. But um we can't expect everyone to be, you know, babysitting their like the vaults uh and the strategies or stable coins that they're investing into. That's kind of the point of having like a a curator or a manager. So um yeah, I hope that going forward like uh these curators just kind of pick a lane and stick to it. So you shouldn't be putting in stream finance collateral into you know a sort of low or medium risk vault and surprising your users um and changing kind of the mandate.

Okay. And so how uh your protocols are secured on the design level from the this crosscultural risk and what do you think like is good options are there on the market?

Maybe I can pop in first. So Spark um we uh allocate primarily through this infrastructure called the Spark liquidity layer. Um we have granular control over each uh like each allocation uh how much we're putting into it, what are kind of like the uh deposit and withdraw triggers. Um so I think versus you know a classic like compound style lending market where everything is just globally rehypothecated, we have um you know a bit more control over that. there's other ways that you know some of the the folks on this stage are also addressing this problem but I think um isolating your risks um can be like very very helpful to uh to avoid one thing blowing you up.

We are and we in a very early stage. We're just starting these permissionless trenching markets, but um the stable we're going after is not having anybody actually be managing what the the ratios are uh or the distribution, but um instead of having a stable coin with no price volatility and just yield, having a little bit of price volatility. And so then let normal sort of decks price and market arbitrage dictate uh which other collaterals to keep or not keep given how things are going. Um in our case there are several things. First of all it comes from the protocol architecture that um we basically give credit lines to every protocol on top of fluid and we control exposure uh against any asset.

And second thing that um we always have like also have a very robust oracle system and automated ceilings on the protocol which do not allow to withdraw liquid to drain liquidity from the protocol if something happens. And uh probably the most important thing is risk management. You just don't get this assets on your platform and you don't get you know wrecked after that. Well, I totally agree with everyone like previously in terms of like proper uh assets risk management like diversification etc. But again, it still can be just a question of time when like black swan event can happen with I don't know counterparty specific asset like or whatever.

And one of the ways we are solving similar like uh addressing kit at resolve is using transient models. So essentially like isolate certain risks as a separate product essentially which might be suitable for more risky guys if like it's required but in case something happening like there is like public contract everyone know like what he is like using

okay

ideally if just one last thought which is being able to mint money is creates a very big temptation to anybody that has the authority to do so. And in history in the long term, there's been government after government that has at times uh overuse that. And so you could argue part of what happened with curators is exactly that. They've faced a tough moment and then used that power uh to over mint or change collaterals or whatnot. And so how do we and I think we are in a very good place to somehow make it permissionless and get that out of the way hopefully.

uh we're still getting there but that should also be how how we ensure that governance is not tempted and can be kept in control.

Okay. And uh what is your like hands-on checklist when you adding a new collateral into your protocol? Like what you should think of about the like what risk you think of when you adding new collateral to your protocol? Uh there's yeah there's a whole checklist of stuff that you have to uh you have to go through to make sure that your collateral is is not going to blow you up. But um I think everything from like sort of governance, custody risks, key management uh depending on the type of asset um all the way down to liquidity.

Um I think yeah we at Spark we usually start the process with like fundamental research um and just try to identify all the different you know risk factors that are kind of like 0ero to one like um where it could potentially wipe you entirely um and then only after that do we you know we have verified the asset is fundamentally sound then we start looking at other other factors market etc. No, I think that's right. I mean, there's a long list of sort of risk factors, counterparty risk if you're thinking about real world assets, uh, liquidity on chain, um, how much do you accept on your stable coin versus the liquidity indexes and whatnot. Um, historical volatility, and even then, I think at the end of the day, it's diversification because you'll never be able to fully capture everything. Yeah, in our case I think it's quite similar to what monet uh described.

We have full checklist of things we are uh looking at including the custody what the rights of the custody the liquidity the governance risks uh I mean like basically everything uh we're checking everything to ensure that when needed we can liquidate the asset that's you know this is why this RWA assets are not coming into defy yet because in our case we cannot on board them as long as we are not sure if they're liquidatable. Um, and for every crypto asset, the same only. We can only on board if we're extremely confident that the asset is safe for our users. We can liquidate and nobody will hurt on on on our platform.

So, pretty similar to what guys already mentioned like basically liquidity on the market ways to liquidate essentially how like volatile and mature in general the asset. That's one of the reasons why we are using uh at least for now only BTC and eure and like not really considering something like alcoin or mium coins or whatever and even in some future if we consider it somehow it won't be like absolutely random things uh for meaningful allocation. Okay, it's good that we have a consensus here regarding risks and uh but do you think will it be possible to like integrate assets with zero or like very low liquidity but that they have like good underlying uh that uh like culturalize them into like current markets or it's it's too dangerous right now. And if for example if you have a agreement with some like liquidity provider or market maker that they will buy this asset if something happens is it like enough or is it too dangerous right now? I think uh in in some respects it's already common to uh to do this you know to greater or lesser degree in the DeFi space where if you look at how much is lent against uh a lot of like liquid staking tokens as an example um there could be you know hundreds of millions billions uh used as collateral lots of stable coin loans out against them or or like sort of looping um and the onchain liquidity is quite quite poor um particularly for like the the longer tail ones.

So um what makes that work is there you know there's a redemption queue um you can kind of model out like a reasonable discount that it might trade at. Um, so I think if you [snorts] know if like sort of the the primary market like the minting and redeeming is kind of tractable where you you have a good understanding of how that works and you're confident in it. Um, it's possible to to lend against uh or just kind of collateralize assets even if the secondary market liquidity is is not very good. Um, but then it you know it brings a whole host of other uh challenges and sort of like you have to adjust your approach to oracles and and many other things. sort of agree.

I think I think right now a lot of the um yield bearing stable coins are very dependent on an external anchor of USDC, USDT or whatn not from a liquidity point of view like secondary markets or whatnot. Um I think the more we move away of that and and the al algorithms themselves do not need that external anchor uh it will be more and more possible to create markets where all liquidity is in uh basically minting a stable coin against that asset.

I would say it's already happening now. uh LST is the best case we like users can borrow ETH against ST for example and even in the case if ST doesn't have onchain liquidity there is still a way to redeem the asset and get Ethereum in return um for assets like if you want to lend and borrow uncorrelated assets it's becomes tricky and uh what I mentioned before about this RWS some of them are permissioned and we already discussing uh with different issues they say okay we can we have like a private liquidator he will come and liquidate but then you have the stress assumption you know what's what's the beauty about uh defy that you know liquidations are open for everyone there are thousands of liquidators we don't know them they come and liquidate but if it's a permissioned system and there is one white listed actor there is a lot of trust assumption will he come if something happens I'm sure he will be liquidating then you know everything is good but when everything is bad like he will not um likely. So in this case you want to have at least several disperion liquidators or maybe you want to have some legal um contracts that they have to liquidate or they will compensate or something. So it's tricky questions, difficult one uh and I mean nobody knows an answer that's why we don't have this assets uh in in defy widely know. Well, maybe if to touch not only defy domain, it's uh like also pretty normal to have something like I don't know redemption cues or just waiting a time until certain asset not necessary even defy or crypto related asset until can be like liquidated.

uh if to talk about RWA and many other like uh scenarios it's like normal cases and I think uh in defy domain some of the like uh learnings from traditional finance can be adapted to basically that's it

like the reason why I'm asking is the we'll see a lot of RWA adoption in the like coming years right and we'll not be able to build enough liquidity on the market for such different assets but we'll need to culturalize them at um point uh and like what do you think like how what will be your approach to it? How you will be able to like say okay we can integrate this but you should have this this and this what it will be

well not really different to what I said depends on uh like strategy for the diversification as salon for example mentioned before depends on the type of the assets like the like highly likely won't be very unified rule or like strategy how you can liquidate all of them like unless of course you are making very specific narrow thing. So and it is either liquidity or allocation management inside of uh your like collateral or just standard things like just you need to wait I don't know uh 24 hours or something like that before like you can be liquidated.

Um

or or guys like fluid can can help.

Yes. So we're discussing yeah with different asset actually first of all I want to say that I hope I really hope that all this like thread fire will move on chain in one day I mean defy is so much better I mean like one day they will realize that and they think that you know everything is transparent everything on chain everything is verifiable every block you don't have to wait for quarterly report to see that there's nothing left there you see every block what's happening with underlying assets There is so much composibility but okay we're speaking about like far future probably uh currently a few things that we're exploring. So first of all private uh liquidators second thing where like we want we would love them to have onchain secondary liquidity and this uh this is where fluidex is also shines because they can create this liquidity at fraction of cost. we're trying to, you know, to present them this idea and this like the benefits they can get and they look they seem interested in this um but because of the regulations they're not sure if they can do that or not. Maybe they need to rep assets but then if they wrap assets there is some other issues and so on.

So yeah like two options I see now secondary liquidity and private liquidators if um permissionless liquidations are not possible but I want to have at least three or more private liquidators to be more you know confident that they will come one day. I think we need to move away from uh a liquidationbased CDP model uh because that's for margin loans for um collateral that has a price that has secondary liquidity. Uh if we're going to move to less liquid collateral, you need to move to a model where the effectively the equity layer is more mutualized. Uh and I think actually um this the star system sky is going in that direction. Um but effectively that that's what you need to do.

Yeah. Yeah. The Sky ecosystem I I don't know if uh people might have seen the uh announcement of OEX which is like a new um initiative at Sky to onboard kind of more interesting like bespoke um private credit assets uh deepend stuff of that nature. But um yeah it's not it's not like a crypto token that you can just yeet into the market uh 24/7. So the approach that you have for that needs to be a lot different.

Uh but I think that's it's like a green field for um you know we we can do like cool mechanism design solutions to some of these problems like one uh one example so uh USDA uh they're investing into uh or will be investing into GPUs which or GPU backed loans which uh amortize over like three years. So you're not getting your money back like really anytime soon. Um but their solution to allow secondary market liquidity and and kind of make make the asset uh able to be redeemed is uh each month when people make payments and they unlock that liquidity uh there's like a a queue um or not a queue actually it's like a bidding process where if you want out then uh if it's not overs subscribed you can just get all your money out at once. If it is then you kind of bid how much of a discount you're willing to take. Um I think we'll see [clears throat] a lot of solutions like that where um it might not be efficient to rely on market makers or uh various you know stuff like that.

So we'll try and figure out kind of like cryptonative mechanism design solutions to make the uh the primary market provide that liquidity uh that we need.

Yeah, I totally agree with everything you said. Uh and uh what do you think about delayed liquidations? Are you like pro delayed liquidations or like you think it's it's bad like betting on some asset to restore some value at some point? Is it okay or not? Maker uh Maker was like the uh uh the OG to do this in D5 only for an hour but uh the original Maker vaults they had this um yeah 1 hour delay basically where you could save yourself uh from getting liquidated and very popular feature um you know as a borrower of course you'd love delayed liquidations um as a lender like certainly for crypto collateral I I don't think it makes a whole lot of sense like there's got to be some level um that you are just going to autom automatically sell because we don't know who our borrowers are.

Um we have no recourse against them. So there's there's nothing else other than uh the collateral protecting you and you can't just like rely on like hopes and prayers that it's going to go back up. But um I mean it's a yeah it's definitely like a very very in demand feature. So if if someone can crack it, that'd be super popular.

I guess as I was saying before, no liquidations. uh are more into an equity model where there's a layer of people that have first loss uh and it becomes kind of continuous. You don't need liquidators, you don't need people holding that. It's just all combined. H delayed liquidations I think in threat why uh it's like in private lending they always have this like you have say four days window to like top up repay do something to restore your say health ratio whatever they call it in private lending space.

uh but they can do that only because they have agreements with some other counter counterparties that they will come and liquidate and they are working based on the contracts in defy we're working uh you know with smart contracts if we can create a system where it will be enforced by smart contract that the collateral will be liquidated even like after 4 days delay for example uh and there's a counterparty which is forced to liquidate no matter what then Yeah, that's that will be a very popular thing. Uh and yeah, probably we will have in the future some you know um some variations about this delayed liquidations. Currently I think there is nothing uh like maker walls was the first first one to create this um delayed liquidations. I don't think it ended up well uh some in some cases. Uh so yeah like probably in the future it will be will be a thing.

Well or opposite if everything will be like closer to defy like you are saying in some future so probably we won't really need uh such such such functionality on practice but until we are there so if it it will help to resolve any like practical I don't know issues without ideally adding new ones for I don't know speculations or whatever it's like why not

if it's done properly

but what if we like add additional insurance layer that can be built with risk takingaking protocols for example and if you have like some value of like uh not bad debt but like some position can be liquidated but it's still in the range that kind of safe and insured and you give some time till it like being liquidated. What do you think about it?

Sounds interesting. I mean, yeah, I think like any solutions that um provide like more safety uh for the user is going to like be a big benefit. It's just a question of like how much extra are the users willing to pay for it and um or is it Yeah. Does it make enough sort of profit for the the resters or the capital providers um who are you know basically stepping in in place of when typically we'd liquidate. Um, I think reststaking is like a really interesting approach to it potentially cuz restakers have like a very low hurdle rate.

They're kind of willing to just yeet their capital around everywhere. Um, so that that could be a good good jumping off point.

Yeah, could be a good solution.

Um, it's interesting especially for the risk takingaking protocols that don't have use case. So I think the complexity will be like pricing of this feature because you know you basically accept all the like bad depth uh and yeah how do you price that um and there were different ways like how people were thinking how to protect users from liquidations let's say different like some sort of inverse range orders unis swap or um like through the options or something I think like in some time we will come up with solutions how to protect users nothing clear I mean there's no clear idea now but it will be a valable valuable feature for sure well with every added layer like some risks also to be added at least like counterparty risks etc and it also have to be considered as like part of risk framework we just discussed uh a bit earlier other than that like makes sense

I I think also like there's yeah there's an approach where you can you can deal with some of this at like the UX layer where so I actually I got margin called like maybe a few months ago and somebody from my broker like literally called me up. I just wasn't paying attention and they were like hello top up your margin. It was very good UX. I liked that a lot. Um and for the risk manager it's also it's also good because if somebody you know cures their own margin call um you're like guaranteed to not lose any money.

you only like have the possibility of losing money if the position like fully goes bad. So, um I think that's probably underexplored like um you know there's some OG projects like people like DeFi save uh I mean instant app did you guys have uh automation like uh okay anyways there are some OG projects who are who are kind of addressing this but it's seems like it's still not uh not not common for some reason

given they can always liquidate you I guess that's sounds more like a growth strategy like hey put some more [laughter] dollars in your Uh okay. And uh I think we can switch a bit. Uh if we are talking about like CD5 protocols that have some uh collateral offchain and they're like doing something, do you think it's like it's important to build some framework to assess this to bring it on chain somehow? I don't know like to to g give like more transparency to users to protocols like and uh will help the industry or like it's it's not uh that relevant right now

100% I would say like transparency like even judging by my own experience like transparency like requires some effort like if you are like reporting something more publicly or communicating like it's it is added operation operational effort but at the end of the day it's solves like quite a lot of other potential questions like uh how you are managing guys like like helping with I don't know due diligence and research to like your clients essentially and like just uh over the time or not just over the time like uh helping to trust as for the specific framework for like integrations across like different uh of chain uh and maybe other onchain just from different chains for example it also like can help you to simplify like uh I don't know management of the assets evolution of the of the product etc. That's also what we are doing and like at the end of the day I think also at some extent uh like uh from official like from regulatory perspective also some more specific frameworks will be will be there or should be there probably at some at some extent.

I don't remember what was the question but I I think I I get the idea. So

sorry for long answer.

Yeah. Yeah. Yeah, I like we already have this offchain assets uh in defy which I use as a collateral including resolve and Athena and the only reason why they accepted widely in defy is because they have very clear transparency. Um in case of resolve it's even better because they have this RLP which is a junior branch like we love that you know it's basically a tokenized insurance fund. So Ethereum also has insurance fund but like it's not tokenized.

This this is easier uh you know for lending markets to it when when it's tokenized. So yeah and what I was saying the only reason why they got adoption in defy on the lending markets is because uh they have very good transparency and everyone like I mean uh risk teams can assess this assets and uh yeah and then like accept it for the users.

I think um as as crypto matures there's going to be more and more transparency. Um but at the same time I think even now you got a lot of people looking for yield and so in some ways less transparent markets have higher yield and as perverse as it sounds that's what people end up doing or investing in insurance layer or whatnot. So but [snorts] yeah this more transparency will lead at least to the sort of like the more senior trench is to have less less risk and should go that way. Yeah, I think um it it hasn't really worked this way yet, but I I do think transparency is going to be a a big competitive advantage for some of these sort of u these projects that are like straddling some onchain components, some off-chain components. Um and I think it'll be driven by like more large scale capital allocators.

So, uh vault curators, stable coin issuers, like the sky ecosystem, I think is is um really addressing this uh head on. And um you know if you can be more transparent and people know everything you're doing offchain um you'll get more allocations you'll get lower capital requirements. Um so yeah I think even though for kind of our retail driven market or like sort of proumer driven market it doesn't seem to matter. It I think it will become really really important.

Okay. Yeah [clears throat] totally agree. Totally agree. So we are like kind of running out of time not having so much and um yeah the last question would be like what security feature helps you sleep comfortably at [clears throat] night in your protocol? Well like internal probably more streamlined and explicit security or operational security process like external audits I'm talking not only about I don't know purely smart contract audits but auditing like how you are managing assets probably from economical perspective is also important and also like helping building trust yeah probably has take a lot of time

yeah so for us first I think it's mixed by audit uh like guarantee that everything is safe in the protocol. But actually um you know for us for fluid and previously for instance we have never had an exploit uh lost funds bad depth nothing like that. I think we're the only lending market on Ethereum that has never had an issue. have never lost any money, have never got bad depth or any exploits. Uh and one of very few protocols overall in defy that never had those.

I think like unis swap and lido uh never had this issues. Yeah. And like we you know we know that we cannot lose users money. That's our like first and the most important priority. So we treat security very seriously.

We have audits, dozens of audits for all of our products. Uh internal internal audits and uh very strict risk management for the protocol and yeah and then if you're speaking about architecture of protocol, it's very very advanced to protect you know us and users. Uh yeah, that's that's what I think about.

Big respect guys.

Pretty much the same. I think everybody's focusing on audits pre-launch, operational procedures and then post-launch just keeping tabs of what's happening and being able to react quickly.

Yeah. Um if I had to pick out a few uh we use redundant oracles. So we use like a quorum of different oracle providers which uh it's one of the sort of top things that could go wrong is Oracle prints a bad price. So that's very good. Um, we have rate limits and like sort of automated supply cap adjustments, so we never are like leaving too much capacity out into the markets, which I think is also good.

Like even if something goes wrong, you're only going to lose a small additional amount. Um and then uh yeah we're we're integrating you know alerting and uh hyper native all this all this kind of alerting infrastructure so that we can um hopefully be the first to know if there's any sort of market moving event where we need to take action and and automate as much of that as possible.

Okay, good good choices I would say. Okay, uh thanks for this great talk guys. I think like that's it.

Thank you.

Thank you.

Thanks.