Yinsure.finance a new insurance primitive

General Chat
#1

Hi Guys Saw this post from Andre on yinsure - seems like us getting into the insurance game is a good idea but I think there may be problems with liquidity and claims:

  1. Liquidity - the return for LPs will have to be large enough to entice them to put up full collatoralization on the cover (that’s a big ask in this space) - I think Nexus has fractional reserve to 6x the staked cover on a contract and where this doesn’t cover the claims - overflow is pulled from the capital reserve of the protocol - I think we can manage this if demand for yinsure is high enough cause the market will sort this out.

  2. There are incentives in the fullness of time for LPs to down vote claims as the protocol grows will increase. As I understand Nexus the actual LPs who stake against the contract to allow for the protocol to provide the cover and are at risk of having it burned aren’t necessarily the ones voting on the claims assessment - because all NXM holders can vote on a claim. I think this is a reputational problem for YFI because if peoples claims aren’t paid a lot of the good will we have built up will be lost. We could perhaps solve this by bringing claims assessment inside our governance structure.

Alternative or in addition - if the problem we are addressing is KYC-less insurance - could we just act as a reseller of Nexus by buying cover for our vaults on behalf of people who want it and taking a premium - almost no risk for us but huge money to be made in our cut? We could even just be a decentralized reseller for all protocols backed by our good name of YFI!!

5 Likes
What does yinsure insure?
#2

Good points. My initial thoughts…for something as critical as insurance we should have a foot hold in that area…it is totally fine also offer other providers as alternatives(competition benifits the users). If we solely rely on third party providers acting as reseller…we are staking our brand name against third party products where we dont know/control all the moving peices…its a higher reputational risk imho

If we start the insurance vaults on a small scale and slowly scale up as the ecosystem grows we will be able to identify and address issue as it grows over time…the biggest risk i see is growing too big too fast(which we cant control with thirdparty insurers)…as long as we are selective about what kind of assets we insure and scale it up slowly we can always pause and monitor/address risks as they come up…just my 2 cents

1 Like
#3

Also some relevant thoughts on on-chain insurance here,

1 Like
#4

This is fascinating. I wonder what type of claims people will try to make. I’m worried that this claim governance by voting may lead to a lot of complaining and FUD from people who have their claims denied. It also might require a lot of work on the insurer LPs part, especially if many people make claims for relatively small amounts.

Ratking, I like your idea of being a reseller of Nexus if they allow that. That way, the arbitrator could be a third party.

But I’m definitely open to hearing about how this new financial primitive will evolve.

1 Like
#5

I’m definitely concerned about this as well. If I provide insurance for y, I have to lock up funds equal to the amount being insured. If y fails, I lose money, if it succeeds, I gain a premium paid by the person I am insuring. It sounds like I’m exposed to the same downside risk as y itself, but with the premium as my upside rather than interest. This implies that the insurance premium is naturally going to reach parity with the interest rate of the asset I am insuring (otherwise I’d just go use y myself rather than selling insurance for it).

In terms of claims voting, do insurers of all asset types vote on each claim, or do just sellers of y insurance vote on y? I think we want it to be the former, since most failure scenarios would involve all y holders.

2 Likes
#6

I like the idea, not sure if the fees are high enough for the insurers. I don’t like: “Should LPs largely decline valid claims, insurers will simply move their funds out”. In case of a larger incident it would be cheaper for the insurers to deny the claims, rendering the insurance useless.

The main reason I buy insurance (I also bought a contract with NXM) is for the event of a major failure of any of the involved protocols and this should in most cases reflect a substantial part of the whole pool (e.g. yVault). There may be events where only a small part of a pool is affected but personally I would like to be insured for both cases.

And if I get a vote as a LP to accept or deny a claim, e.g. in case half the yVault funds are gone, then I would rather deny it and keep my funds instead of accepting the claim and lose everything.

The alternative that I can see would be a joint insurance vault, where a percentage of all yearn products get’s invested into, something like a joint insurance treasury (which itself would not generate that much of a return). Or you could set up the yearn products to act as cross-collateral with each other, any risk is shared among all pools, basically ruling out a major loss for a single product.

And I would be willing to pay more for the insurance, especially if a third party is validating the claims, instead of the LPs themselves.

3 Likes
#7

Great idea! I may be missing something, but it feels that the only way an LP would want to provide insurance is if they can get a higher return than by investing in the pool they are insuring. Otherwise, they can just go to the pool themselves. The risk profiles are the same.

With a 1-1 relationship between the reserves and the notional amount covered (I am assuming this is what’s being proposed), I don’t think the fees will be enough to provide meaningful returns to LPs. Realistically, capital/reserves only need to be a % of the notional amount insured. I think for this to work leverage will be needed (insured reserves/capital are only a fraction of the insured amount), fees need to be higher and the LP capital needs to be invested in a uncorrelated asset (another pool?) that provides meaningful yield. Fees should cover on average how much $ will go out in claims, maybe a bit less depending on investment returns, but LP capital should be mostly be there for tail events. Looking forward to how this turns out.

Thanks for all the work!

1 Like
#8

Another thought…what if we allocate a small portion of the rewards distributed to the YFI holders into an insurance pool(may be 3%)? This might start out small but It will grow and compound over time as the ecosystem grows. Even though this might be a small tax on the YFI holders it would be in the best long term interest of the project. This could be in addition to other insurance pools. Perception of safety is paramount for the yearn ecosystem.

1 Like
#9

I do like the idea of having an insurance pool, especially one that fills up to a certain x% of TVL, before then spilling over into rewarding YFI token holders.

My biggest problem with having an insurance pool is that it means we YFI holders, will have to take smart contract security into consideration, when green lighting new strategies.

In a design where insurance is strategy specific, then each strategy has it’s own risks/rewards, but it isn’t impactful on the wider YFI ecosystem. It becomes up to the yield farmers and insurers to establish risk.

1 Like
#10

Also it could be a combination of both…Andre’s efforts on Yinsure will help the ecosystem gain foot hold in an important future growth area…its important to deploy and iterate on insurance strategies starting now. As long we keep the size of the insurance pools small in the early days while we iterate/refine them, the risk should be manageable. But its important for us to get a foot in the door strategically speaking…

The common YFI insurance pool could be another layer of fallback safety…

I think a combination of the two approaches will be very powerful.

Andre’s Yinsurance intitative is a calculated strategic risk we take…we back it up with a common yfi insurance pool. May be we can say the Yinsurance insurance pool can never out grow the YFI common insurance pool? That way the risk is always under control. This way we can cap the allowed insurance slots, creating scarcity in the early days and allowing it to grow at a reasonable pace…once the community feels that the insurance is starting to work well we can remove that restriction?

2 Likes
#11

An approach to making insurance more attractive to the insurers, is to allow them to insure multiple assets with the same collateral. There would have to be some requirements for orthogonality between the assets. It’d allow for covering multiple assets with a smaller amount of funds, while also juicing the returns of insurers.

A system like that works, as long as there are no cascading or simultaneous failures.

2 Likes
#12

Related to insurance.
a proposal is in progress with an example for YFI insurance pool and Cover, deep dive in.

3 Likes
#13

https://blog.etherisc.com
@andre.cronje @banteg , they are Backed by ETH Foundation
What everybody real-life people want.

image
image875×492 86.5 KB

The non kyc idea is excellent and the insurance funds too.

1 Like
#14