Marek Hauzr / DCxPrague

Transcript

um hi guys uh I'm mik um from Carmine Finance uh we deal with def fire risk directly on chain and yeah uh I'm going to talk about insurance uh and the problems around it uh and how to how we can kind of solve them um before I start has anyone insured anything on chain up to this day like show of hands really did awesome awesome yeah uh so like yeah there was just one hand for those that didn't didn't see it uh yeah so that's huge problem at the moment there is very very

little use cases for insurance on chain um but before we get into the use cases what actually has to happen to to have the insurance right to have something that you would call insurance uh one is the insurance itself the smart contract that we describe that another one is hedging uh the reason behind that is that it's very easy if you have very specific things that are scalable to do directly insurance for them but things that are a little bit nuanced and you can do one or two things of those uh instead of like pricing those you you usually hedge those uh because it's much much cheaper uh on your end as an

insurer to do and another thing is monitoring right uh because at the moment if you were to talk to def5 protocols quite a lot of them would tell you hey I don't want to insure I don't want to hatch uh and I don't want I don't know what the risks are right that's actually the the case that uh we are observing as as Carmine Finance right cool so uh what insurance does exist on chain right uh as I said there's not a lot uh there's niche market that is surprisingly big and and that's actually Insurance of blockchain things but it doesn't happen on chain actually those

are uh legal contracts that happen off chain uh and what you can do is you can ensure the security of your smart contracts right so you do audit and with the audit uh you can ensure sure you can ensure against hacks and exploits uh so that's probably like the biggest use case of insurance of anything onchain there is uh there is insurance for real world assets uh which what which is what J was mentioning uh that he ensured uh that his flight is not going to be late uh right so that's one of I would say three use cases that are available in terms of real world assets uh the sorry the other

two are ensuring uh against crop failures which is uh as essentially parametric insurance against that which means that if there's not enough rain you're going to get get paid off no matter how your crop was good or not uh so it so if there is good uh good rain good good good amount of rain and your crop fails you're screwed essentially um and there have been few ad hoc things that different protocols have been trying as far as I know uh so for example insurance against flooding uh and a few other things but uh those kind of died uh as far as I know

uh another thing that has been happening recently uh but hasn't received uh Market direction that I assume it will is hedging or insuring against uh impermanent loss right because this is relatively easy to do uh there's bunch of different designs how to how to deal with this uh but we have our own too uh and from when I talk to amm for example uh a lot of them and I mean a lot of them look into this and they are just like getting ready to to deploy these things uh but again like there hasn't been there hasn't been a lot of Market

reaction yet um and so there's not that many use cases that have been or will be deployed soon uh but what can we actually do right uh so one thing is to to take care of kind like to onchain stuff uh the risk that you you receive directly onchain which means you can completely do that within smart contracts and uh there is kind of like the other side where you have to talk to talk to the real world right uh where you have to uh do either some legal paperwork you have to do kyc uh you you have to make sure that uh data you are pushing through Oracle uh have a

secured uh data source and so on so in reality there's today uh there's not that many things that are happening behind the scenes just because all of these problems uh so in reality the biggest one that will be that will happen soon is the impermanent loss and companies are trying to push few more things which is for example um uh the insurance against uh uh delayed flights uh in Kenya and uh one more African country they have insurance against crop failures which in reality the way this uh happens is uh that you get percipitation data through an oracle on chain and based on that you pay or you

don't pay uh the farmers or the entities that ensure right and yeah uh there's not that much that can happen right um so yeah I mentioned the the impermanent loss another cool thing is uh lending protocols they face the risk of under theater loans right they don't talk about it a lot uh they don't advertise it a lot but if you were to look for example on AA uh not sure what the number is today but few few months back they had I think $300 million in a security pool to make sure that they can pay off the uh the liquidity pools to in case of uh under thewater loans

happening right uh but if you don't know what under the water loan is it's essentially a loan that can be profitably liquidated right um and actually the insurance against that is pretty pretty simple uh the only thing you have to know is the amount of the insurance that you want to buy and then the insurance is nothing else than in most cases long put on that given pair which means what you borrow against what you borrow in few cases it could be long call but based on the data most of most of the cases are long put options uh and nothing else uh and yeah uh mentioned those two uh you can ensure and it's been happening slowly the against crop

failures and also uh the same way you can ensure against flooding parametric parametrically ensure against flooding uh you just uh push uh height of the river through an or on chain and suddenly the insurance is is relatively easy uh and now to the to the to the main point what is actually the issue right um so if you look at defi most of the things that we have are fully collateralized which means that there is huge cost of capital uh yeah uh that there is huge cost of capital uh so imagine that you would ensure against the flooding for example

uh and the probability of the flooding happening would be I don't know 1% in a given year right so in 100 years chances are your house will be flooded once right so the premium you're going to pay is 1% of the insured Value Plus some margin right on chain because it would be fully collateralized at the moment the cost of capital is roughly 3 to 4% suddenly you're not paying 1% plus margin you're paying 1% plus margin plus the cost of capital which would result roughly in 5% and now imagine you would go to the end consumer and tell the end end consumer look you're not going to pay 1.5% for example you're going to pay five that's not

going to happen right so cost of capital is is a massive thing another one that was on the previous slide is that in traditional Finance these products are pretty complex right you have to describe them well you have to measure them well well you might have to measure the events well you have to price those well and so on and so on uh and most of the people most of the developers that I've talked to they tend to uh look at this in a way that they would kind of copypaste what is happening in traditional finance and put it on chain right which means suddenly the cost of development the cost of deployment is pretty high

and connected to the cost of capital on uh blockchain doesn't have any reinsurance right so what this means is that if you insure against flooding for example your insurance your insurance company that you insure with is insured within within the insurance company and it's insured so that it can pay off in terms of too many too many events uh would happen right so imagine there would be insurance company in pra um selling insurance against flooding only in pra on vava River chances are if the event when the event happens everyone is going to get paid right so ideally this company this insurance company would had

100% collateral for that in reality they don't they have 10 maybe 20% something like that and they Ure themselves uh with their insurance companies right that can diversify the risk right um what this means is and we don't have that at all and we're not even close to to having this on chain um and in reality what we we could do on chain is that we would just transfer the risk right so you can measure the risk you can estimate the risk or you can have a market that is estimating what the risk is and then you can when you know this you can have an insurance company or an insurance protocol that would insure you

and then this protocol would be insured uh within reinsurance protocol which means just transferring the risk to some someone else essentially a risk of risk onto someone else and what this would mean is that within the insurance company the diversification is big enough so that they don't have to have full collateral uh against everything right and then your insurance company because it's insured that doesn't have to have full collateral right which means cost of capital goes down and cost of the insurance goes down and actually this is equivalent to the traditional Finance um but we're not even close to that and we to be honest we need that

and we will need that because if defi have in to for defi to compete with traditional Finance without this without removing the cost of capital it's not going to it's not going to be competitive at least for like a normal person that doesn't care about blockchain at all which is most of the world right uh yeah so uh there is quite a lot of quite a lot of opportunities uh that can happen uh re insurance companies are insanely big right uh there's not a lot of them uh the insurance itself uh since very very few Protocols are doing that

it's essentially open market and insurance Market in traditional world is huge as you probably know so there's a lot of opportunities and yeah that's what I wanted to go through and yeah uh questions please thanks when you mentioned risks uh isn't it a finally use case for the prediction markets yeah yeah it actually is um so one of the things that we do is we have implemented options amm uh in reality what this means for us as car mine is

that we look at options as if statements right and any insurance is essentially if statement right so if stuff happens until some maturity you're going to get paid off options are nothing else you just replace the the part of the what if right so instead of if price is higher or lower you're going to go if there is more water in the river or if there is less rain right and yeah and the amm actually works kind of like the prediction Market one related question is uh how do you deal their with the Oracle problems yeah so there is bunch of different ways how to how to how to do that and like to be full

honest this is very very com could be very very long answer right so in case of Carmine we uh we have an oracle just for the underlying price or the underlying value right so so for example if it's options for if versus USD we look at the sport Market of if versus USD and then the volatility parameter that we use in uh is moves up or down based on the market realized market supply or Demand right so if someone buys it moves the price up and we do that through moving the volatility up if price if someone sells it move it should move the price down and we do that through moving the volatility down right so we use uh the spot Market

uh which is pretty well used across different protocols in terms of the and then it gets there are different designs obviously uh there are option amm that query the or use the volatility or price of the options directly from an oracle then you have a problem of mispricing and blah blah blah uh and there is one more problem with real world assets where essentially imagine you would have the case of uh insurance against uh uh flooding in prag right so you would want to query the height of the River from some endpoint now what you have to make what you have to make sure happens is that that Endo is secure right which today how do you do that like it's not

properly right and how do you do that you have to have some certification around it you have to have it deployed on multiple multiple sources by multiple parties and so on so on so yeah uh and it could go on the answer could go on so just very briefly yeah bunch of different problems we at the moment like us we look at some specific things and yeah that's how we deal with them yep hi thank you for the presentation uh first of all do you run your own oracles or do you use some third party yeah so uh us as a company we just implement the code base and the protocol is run by decentralized entity by the way

connected to the to to to the talk that was the first one uh that's how we do things uh and we don't have a controlling part over the protocol we have at the moment roughly one3 of the votes uh so yeah that's that's how we do it okay and uh regarding uh putting the information regarding fing con for example could it be that the source is trusted but you could have like a decentralized SC like like claros at the top of it where a user could dispute the decision yeah that would have to happen uh because like we can't guarantee that the that the Oracle is correct so uh how we do things now for for the crypto

assets is that once the option matures immediately after that you can get your money out with r of assets what we would do is there would be for example a week till you could get your money out right to after the maturity so that there would be period where someone could come in and dispute that say Hey this is wrong right the Oracle was uh uh yeah was hacked for example do you have solution for that already in place short answer is yes we can do this we have but we have to have few things one is the security on the endpoint side right then we have to convince uh oracles to provide the data which would be I don't know 10 minutes of convention

for us and on their side few hours of implementation uh that's what we need and then we would have to put in the delay for uh for getting the money out which again is I don't know one two or three lines of code into into the protocol which in reality would mean I don't know week till till it would be approved by the governance okay thank you yeah welcome uh just one more thing to this we wanted to do this very very early on but the market is far from ready for this so when we were discussing this everyone said hey not interested uh yeah and it would be way more expensive than uh you going to a traditional traditional insurance

company cool