Maybe im misunderstanding here, but cover_fee = min(cover_fee_max, 1 / u(CV)) is indeed a dynamic pricing model of the cover fee capped at some reasonable number (e.g. doesnt make sense to pay 100% premium) - and this follows a oft used curvature based on the utilization of the covered vault.
What’s wrong with the eq?
Since
then, using 1 and 3 as arbitrary ratios,
Regardless of whether the fee is an absolute number (0.5%) or a percentage of the yield (0.5 * net yield), it seems odd that when a CV is “under collateralized” it gives a lower fee to insurers. The protocol should be working to attract more people into that CV.
Did I misunderstand something here? I’m thinking the equation should be cover_fee=min(cover_fee_max, u(CV)) instead (without the reciprocal), unless i’m completely wrong.
I just saw the Terra news about UST on UnslashedF. I had been mulling around in my head from the beginning: How would Ozone do so something similar to what was just released on UnlashedF?
Ideally, it would be if the pegs breaks by more than 2.5% for period of more than 6 - 12 hours Ozone insurance would kick in with a buying plan. This would need more thought because liquidation in Ozone into UST could be tricky. The key idea is that if Ozone could work out some mechanism where Ozone buys huge amounts of UST at below this peg, it would greatly benefit Ozone stakers when it comes back to the peg. It’s a win win for the whole ecosystem.
Would like to hear more thoughts on this.
The thought of Ozone providing protection for the UST peg also did cross my mind. However, it feels counter-intuitive to have a protocol within the ecosystem try to protect the ecosystem. Unslashed works well conceptually because it’s a 3rd party provider - even if UST completely fails and Luna goes to zero, they will still be able to pay out the claim in another stablecoin or fiat. For Ozone to do the same, they would need to pay out in UST which, like you said, will be very tricky.
What you’re proposing sounds more like an automatic re-balancing mechanism that’s meant to stabilize the peg, rather than an insurance policy that pays out in the event of a claim.
The way I had originally envisioned this when proposing this protocol was similar to the insurance industry MCR and SCR. Simply put this is how I see the value of it and how it is working. The idea is the Ozone insurers need more yield than simply the return from the borrower on a 1 for 1 dollar basis i.e Insuree stakes 1000 oUST and insurer claims say 2% of that yield, there is no incentive without “leverage”. Simply put standard deviations used to price risk, ideally between 2-3 STDs like brokers do for margin for derivatives. If the insurer can take 5% capital as in a 2 STD risk calculation (insurance industry SCR it’s even higher, closer to 3 STD), they can cover over 95% of assets. This is how the insurance industry works and will allow Ozone stakes to have 90% plus yields if implemented in this way.
This is how I see it. Would like to hear more about this.
The problem is cover_fee = min(cover_fee_max,1/u(CV)) encourages more insurer capital when it is not needed and less insurer capital when it is needed most. It creates the opposite of the incentives needed.
Here is more detail:
And here is the model
https://docs.google.com/spreadsheets/d/16uoop_5eOW1qdEnJB7--WCn6w35ivGXfgbZTf_1JgCk/edit?usp=sharing
You actually could easily provide some PEG insurance in a CV.
If the CV had 50% insuree capital and 50% insurer capital, then the insuree is over-collateralized and the total pool of assets could suffer a 50% fall in value relative to a peg and the insuree would not lose a dime. You might imagine different CVs with different percentage mixes of Insurer and insuree capital and then allow market determined pricing for split of income from the asset pool to attract the right amount of insurer capital needed relative to the insuree capital in the pool. All you need to do is start by specifying a percent depeg insurance loss event that is protected, this determines the mix of the pool and dynamic pricing would adjust to attract or detract capital needed to hit the target mix percentages.
A capital stack could be created that offers higher yield and lower risk than anchor itself.
In our view, while this sort of model is a nice way of allowing users of a single protocol to trade on differences in risk appetites, it can’t serve as the basis for the more general insurance framework we are attempting to build with Ozone. For one, insurees would no longer be protected from more substantial failures, i.e. if the underlying protocol would suffer a complete loss of funds insurees would have no recourse. Additionally, our long-term vision for Ozone is for it to be insuring tens or hundreds of different projects, so a central treasury of funds that covers all of these is essential.
Thanks Michael.
I’m intrigued by your comments and appreciate the response. Can you expand further on the mechanics for protecting insurees from losses “if the underlying protocol” suffered “a complete loss of funds”? Best I can tell, under the current proposal and in a slight adjustment to the simple example I modeled, there are potentially three sources of capital to protect insurees from losses:
You can insure against losses in excess of (contributed capital + retained earnings) by allowing the protocol to mint tokens to absorb the excess losses. In this way, the token of the Ozone protocol would share a resemblance to Luna in the Terra-Luna protocol in that the Ozone and Luna both act as a shock absorber by minting more tokens during adverse events. This is really no different than the tools traditional companies have at their disposal for absorbing negative shocks, but the nice thing here is that it can all be done algorithmically which removes much of the cost structure a traditional company has for administering something similar. This should mean much better pricing for insurees and much better economics for the other stake-holders.
Ozone is a wonderfully ambitious project. It is also a bit complex and perhaps still ambiguous in parts. Any additional description on how the mechanics would work under different circumstances would no doubt be highly appreciated by all in the community. Maybe even consider releasing a little model in google sheets so folks can get a more concrete sense of how all the pieces fit together?
While Ozone as a whole will be leveraged with regards to total liabilities, any individual protocol insured will be more than covered. So if a CV suffers very significant losses, in excess of what could be covered under the framework you proposed, insurees will still be made whole. This is a little unclear I think because at start we will only be insuring a few projects, but the long-term vision is for Ozone to insure many different projects. Imagine as a toy example there are 20 CVs, each holding 10 UST, with the treasury containing 50 UST. The loss of all insuree funds from a single CV could easily be covered.
From what you wrote and the detail from the proposal, it sounds like the structure is something like this:
So, for example, if Covered Vault1 had an adverse event in excess of A1 (see the box highlighted in yellow), more capital could be drawn from the Global Insurance Pool to shore up any shortfalls. Is that right?
Can you expand further on how leverage is created at the Global Insurance pool? The proposal describes a target_leverage_ratio:
target_leverage_ratio : Ratio by which Ozone’s AUM will be leveraged to provide coverage. For example, if N is set to 4 and Ozone’s AUM is 2B, Ozone will provide coverage up to 2B * 4 = 8B. Initially set to 1.Question: If leverage is determined at the General Pool level, how do you calibrate for different kinds of assets at each covered vault? For example, a covered vault with aUST assets might be far less risky than covered vault with mTesla assets. You would think that the allowable leverage on the mTelsa CV should be lower than the allowable leverage on the aUST CV. Is this possible under the current proposal?
Hi guys,
What is the most reliable and easiest way to get insurance on Anchor? I want to put some size into the stable UST and want to just have some peace of mind and protection from protocol / exploit risks etc before I do… any recommendations?
You have the “Get Insurance Coverage” right there on Anchor web app, one for UST peg and several other for smart contract issues.
Thanks Joao, just had a look and the rates are not too bad.
hi with the news on Ozone to come
“Risk Harbor Announces it Will Take Charge of Terra’s Ozone Protocol”
Link
May I ask where can we view the latest concept paper? or is this still the one that will be in place? Thank you
Thanks for the write-up @dokwon! You mention under the “buy coverage” section that –
“If there are any remaining losses, they are withdrawn from the remaining insurance pool, “slashed”, with insurers being compensated in new emission of $OZ tokens vested over [1 year].”
→ Does this mean that L(CV) > O(CV) in this scenario? Additionally, what is the “remaining insurance pool?” Is that the global insurance pool? I’m also struggling to understand what mechanics exactly enable Ozone to guarantee levered insurance coverage. I’d greatly appreciate some insight on these matters.
keen to have a read too if there’s one
Are there docs on how Ozone in its launched state works?
Also I remember talks about OZ as a token. Any news on if that is still the plan?
Since this is levered coverage there is a possibility of a black swan event occuring and the losses being greater than the coverage provided. How do you deal with a liquidation event like this? Presuming everyone in the pool is treated equally, is it a pro-rata distribution of the remaining funds or a first-come first-serve grab for the capital?