Algofi Insurance Market

Temperature Check
1

Summary

The Algofi core developer team should implement an insurance market where anyone can lock liquidity for yield, earned by offering insurance against bad protocol debt.

  • Insurance against bad debt will increase confidence in STBL2’s peg
  • Insurance markets will make it easier to onboard more risky tokens as collateral
  • Insurance markets will add an extra way to earn yield on the platform, that can also generate more fees for the treasury.
  • Insurance markets could be a way to deepen liquidity for BANK

Motivation

Bad Debt

If you are familiar with how a decentralized market can take on “bad debt”, you can skip this section. For those who aren’t aware, it is possible for the protocol to enter a state of “bad debt”. This occurs when the price of collateral drops faster than it can be liquidated. For the sake of an example let’s imagine the collateral factor on Tether was nonzero and that the numerous amount of negative press on Tether (USDT) is accurate. One day the price of Tether starts collapsing. If Tether was useable as collateral, loans borrowed against it would start getting liquidated. However if the price dropped quickly to 0, nobody would want to liquidate positions because they couldn’t sell the tether collateral they liquidated. This would create a situtation where lenders may not be able to recover their funds.

Black Thursday

This “bad debt” situation is not a hypothetical. It happened to MakerDAO during “Black Thursday”. To summarize, a problem with the liquidation mechanism led to zero bids on liquidations, during the March 2020 ETH price crash. This left the protocol with uncollateralized DAI which is called “bad debt”. Users were made whole by diluting the governance token, MKR, via new issuance. The issue with this solution is one of incentives.

This worked out because MKR has a lot of whales (Rune, VCs, etc) who could force that vote through at the time. Its unclear if that situation occured again, if MKR holders would actually vote to dilute themselves, given they may not have a financial incentive to do so. A DAI surplus buffer was introduced to address this, but that’s no guarantee that it is large enough to cover potential debts.

Proposal

Open Insurance Market

My proposal would be to create a special insurance market on Algofi. Users could stake any combination of the following LP tokens:

  • ALGO/STBL2
  • ALGO/BANK
  • STBL2/BANK

Staking in this market would be time locked, similarly to Algofi governance locking. Stakers would divide fees (fees explained later). Earned fees would be immediately withdrawable. The staked funds would be completely locked and only withdrawable in two scenarios. First if the time limit has expired, the user could withdraw their funds. The second would be in the case where the price of collateral falls, such that any loan in the lending market has higher debt than the collateral backing the loan. Any actor would be able to execute a special liquidation smart contract in this condition. This contract would allow the liquidator to withdraw from the staked insurance fund, swap the LP into whatever the loan is denominated in, and then repay the loan all in one composed action.

Insurance stakers would earn fees derived from a cut of the interest the borrower makes on their collateral. For example, right now someone could be lending $1000 worth of USDC and borrowing $500 worth of Tether. Their collateral would be earning ~1.60% at current rates . In this world, the USDC might only earn 1% and 0.6% is reallocated to insurance liquidity providers. BANK holders would vote on what these rates are for each form of collateral. Low risk collateral like USDC would have a low premium but something like Tether could be given a very high premium.

Insurance stakers should be happy to earn fees denominated in the same unit as what they are insuring. Fundamentally, their bet is that the token is safe. If I’m willing to insure Tether, it would be silly for me to think there is too much risk in earning Tether denominated fees.

Onboarding Riskier Tokens

There has been quite a lot of interest in adding new tokens to the lending market as collateral (Agrotoken, Defly, Tether, goMint, XET). An insurance market would allow BANK holders to set insurance rates for each token separately, based on risk. Because BANK holders already vote on the collateral factors, it would be on BANK holders to adjust those factors based on how robust the insurance market is for a given token.

Further Discussion

Control Insurance Exposure

Ideally insurance stakers could choose which tokens they are willing to insure and only collect premiums from those tokens. I imagine there are people who would want to insure Tether but not goMint.

Solution to centralized tokens

MakerDAO has been going through some struggle due to the conflict over USDC backing DAI. Specifically, what happens if Circle freezes funds due to sanctions. Circle freezing collateral can depeg a stablecoin like DAI or STBL2. If the Algofi DAO finds that 10% of the USDC on the platform could be exposed to freezes due to sanctions, it can vote on a risk premium that would attract enough insurance capital to cover any bad debt that arises from that situation. This would give users of STBL2 confidence that even though STBL2 is partially backed by USDC, Circle’s antics can’t depeg it. This could be a pathway to making STBL2 more resilient than even USDC.

Insurance Fund denominated in LPs vs Tokens

Personally I think that the LPs are better than just the straight token. Staking LPs would earn swap fees, lending fees, and insurance fees which would make them the highest yielding farms on the platform. That creates a wonderful incentive to participate in the insurance market. Because of lending pools, this is all very capital efficient. If we used straight tokens, the insurance funds would essentially be locked away.

Lending demand destruction

There is a valid question to ask, which is if lowering yields on lending (because of insurance premiums) would decrease the TVL. Personally, I don’t think this would be significant. Because the insurance fund uses LP tokens (specifically lending pools), staking there would likely have higher returns than just staking on the lending market. imo a lot of liquidity would move there. Because the liquidity is lending pools, it would still be available for borrowing and would still be a part of the TVL. I think the biggest risk are people holding USDC who are not interested in staking anything else but USDC. goETH and goBTC are currently too small to really matter and ALGO vault users are earning more in governance than anything else. Setting insurance premiums too high would push USDC lenders to other platforms, so the key is setting them low enough to be competitive but high enough to drive liquidity to insurance staking.

BANK Rewards

I think BANK rewards naturally fit here. The point of dispersing BANK to users is to enable the users of Algofi to govern it. Insurance stakers would be critical stakeholders in the ecosystem and should be a part of that. Emissions are already being dispersed to anyone who stakes the 3 LP tokens I listed in the farm section of the platform. All I’d be advocating for here is to require participation in the insurance market to keep receiving those emissions. Essentially it makes those farms provide more direct value to protocol.

BANK Tokenomics

Insurance funds being pooled in the 3 LPs I listed would specifically deepen liquidity around those tokens. In particular, that could have an effect on the price of BANK. As I’m fairly clear about, I will fight the DAO making decisions with the specific intent of raising the market price of BANK. However, this proposal might change the tokenomics in such a way that has consequences on the protocol itself or its governance. I’m curious if anyone sees any potential problems with a lot of liquidity between BANK and ALGO/STBL2? One thing to consider, is that in this world it might be possible for BANK to be used as collateral, which would increase its utility.

Exit Poll

Would you stake funds in an Algofi insurance market?
  • Yes, I’d insure anything the DAO thinks is safe
  • Yes, but I’d pick and choose which tokens to insure
  • No, I think its too risky
  • No, There are more lucrative strategies I’m pursuing with Algofi
  • No, I don’t have interest in staking ALGO/BANK/STBL2
  • No, other
  • Other

0 voters

7 Likes
2

dont have time to read it fully right now (i will tho) but wanted to ask of you know about nimble? its an insurance protocols thats on testnet which could be used for stuff like that

2 Likes
3

I have high hopes for what they are building. I’ve played with the testnet and also have been trying to get them to run a participation node.

Claims are paid when users of the protocol suffer covered financial losses due to failures in either the protocol code, economic design, governance set-up or oracles.

The above is from the testnet. Bad debt would fall I guess fall under “economic design” but I’d need to verify that with them. I have doubts they could ever set premiums for that on their side, at least not statically, because the risks shift as the Algofi DAO changes what collateral is acceptable and the collateral factors. They are also a private insurance provider where individual holders would be buying policies. The kind of insurance I’m proposing is more to keep the entire protocol solvent than a specific user or wallet. I could see some world where they integrate with a feature like the one I’m proposing and become a bridge between CeFi insurance providers and an Algofi insurance protocol. I could also see a world where Algofi has a protocol level insurance system for solvency but individual users buy from nimble marketplace to insure against other types of risks (ex smart contract failure).

4

This is a very good idea :+1:

1 Like
5

I need to continue thinking on this before I can contribute. Before that I wanted to commend you for making a well thought-out proposal. The community really has come a long way in a short time since the very first community proposal, “Hey what if we had lofty tokens… (refuses to elaborate)”

5 Likes
6

really like the idea of building an insurance fund, naturally i think a part of the treasury would be used in case sth like bad debt happens but with the TVL rising the treasury alone might not be enough.

the numbers were probably only an example, but in that case i think almost 40% of the supply APR going to liquidity insurance providers is way too much. people might leave the protocol if the cut for normal users is too high. also a potential cut could be discussed of the 17.5% the protocol earns right now from the borrow interest to go to liquidity insurance providers instead

i think this part would hurt the protocol a lot, potentially kill it. making people lock up their LP tokens to farm them + the risk of them being sized is too much. it would lead to people just leaving the LPs imo

2 Likes
7

the numbers were probably only an example, but in that case i think almost 40% of the supply APR going to liquidity insurance providers is way too much.

I’m flexible on the numbers but it really depends on market dynamics. We see frequent fluctuations between the rates on Folks and Algofi and most users (especially the whales) aren’t shifting around their stacks because of that. I agree the first cuts would come from the Protocol’s earnings but there is room to charge from the lenders, especially because many are there to borrow, and aren’t overly sensitive to the yields on lending. But starting small and slowly increasing is better than introducing a huge fee.

i think this part would hurt the protocol a lot, potentially kill it. making people lock up their LP tokens to farm them + the risk of them being sized is too much. it would lead to people just leaving the LPs imo

I disagree this would kill the protocol. The protocol will eventually not have any BANK emissions. So it needs to be able to survive and provide enough fees to LP holders without them constantly selling BANK. The vast majority of the current TVL is wrapped up in the vault and being used in the leveraged governance trade. Because governance rewards are high, that creates enough lending APY that ALGO paired Lending Pairs are quite profitable. ALGO/STBL2 and ALGO/BANK already have yields above governance without taking BANK emissions into account. So the incentive to try to earn is there without the emissions.

The BANK emissions are to seed the membership of the DAO. Imo the users willing to take the lockup are aligning their incentives with the DAO. If one doesn’t want to take that risk fine, just stick to earning LP fees like you can on any AMM. Those with locked up LPs are disincentivized to sell the BANK they earn. They are incentivized to use BANK (for its actual purpose) of voting on protocol changes to ensure that the protocol never takes on too much risks and they lose their position.

8

BANK emissions continue forever essentially since its a geometric series. but yes they will become less and less but cutting them already for farmers who dont want to take any extra risks is definitely the wrong way to go imo

again people who provide liquidity to algofi should have a say in the DAO, forcing them to take the extra risks is wrong imo

1 Like
9

again people who provide liquidity to algofi should have a say in the DAO

I think that’s fair, because they should be represented. Maybe a compromise is giving a higher rate of reward to those in the insurance market. Because insurance market users are taking a bigger bet on the health of the protocol than straight LP providers are.

1 Like
10

Algofi should focus on its core competencies imo

1 Like
11

imo this is a core competency. As the DAO, we are already tasked with evaluating and managing the risk of collateral allowed in the lending market. This proposal is simply about creating a hedging mechanism against the risk that we misjudge the risk. An insurance market helps to provide liquidity for that mechanism. STBL2 is a flagship product. Insurance on the collateral that backs STBL2 will increase confidence in the peg because it offers a hedge to risks like clawbacks/blacklisting of underlying collateral (USDCa for example). @DragonFi do you have reasons you think this would be technically difficult to implement or would create negative externalities for the platform? Or do you see this as competing with other features/development that you think is more important?

2 Likes
12

Not a technical dragon. It’s the second part that I think is important. I think there is more work to do on the dex and UI/UX and decentralization. I’m no authority, try to stay a humble dragon. My 2 cents

1 Like
13

I think there is more work to do on the dex and UI/UX and decentralization.

I can respect that perspective. I don’t know that I’d say this is the highest priority thing to work on either. But I do think its something that should be considered for the 1 to 2 year time horizon.

3 Likes
14

I don’t fully understand the Insurance proposal. I think @pescennius is right to think of products and to suggest products that add additional protection to the protocol and it’s users. Personally, I don’t think I would participate in insurance as I feel like Algofi already has a conservative approach.

I think the best next step is for this proposal to be amended for the core dev team to consider the pros/cons, feasibility and where it would fall on the roadmap.

I was thinking how the dev team might feel about governance proposals. For governance to be successful the relationship should be bi-directional.

I would absolutely vote YES for this proposal if it was amended so that the dev team can evaluate and report back.

2 Likes
15

Nice, thoughtful idea @pescennius and I really like the general concept to facilitate broad participation in the liquidation process. That would be a good thing for AlgoFi lending health and another potential market for users, as you say.

I’ve hoped for a vault-like thing for USDC or STBL to be used for liquidations, which is very similar to your proposal. There’s an existing 7.5% liquidation incentive already which is enough imo, but could change based on liquidation demand… increasing as the need rises. It would be ideal if the liquidated asset was swapped to the user’s preferred liquidating asset (USDC or STBL) in the same transaction as the liquidation. That would be my preference, but maybe that swap could be a choice. I’m not sure this sort of transaction is technically possible, but maybe the devs can say?

It seems to me, using USDC or STBL token for this purpose will remove some risk and complexity so I’d favor those rather than the LPs mentioned.

2 Likes
16

It seems to me, using USDC or STBL token for this purpose will remove some risk and complexity so I’d favor those rather than the LPs mentioned.

There are a few reasons I think the LPs are better than using only USDC and/or STBL2. The first is capital efficiency. If the collateral is the direct tokens, they are locked out of use in swapping or lending which is inefficient. Lending Pools are incredibly efficient because the tokens represent a claim on assets that are actually getting used. Next is the fact that both USDC and STBL2 need to be insured. USDCa has a clawback risk and STBL2 has a depegging risk. So at the very least a 3rd token (imo should be ALGO) should be used.

1 Like
17

I don’t believe a lock-up period is necessary. Rather, a liquidator would commit and remove funds similar to the lending and vault contracts. The liquidation incentives (based on demand) should be enough to attract funds for the opportunity to liquidate. One could still borrow against these assets too, but doing so would reduce the assets available for liquidating, and may introduce the possibility that a leveraged liquidator becomes liquidated themselves. The risk of liquidation would be reduced by using non-volatile assets and the ideal swaps I mentioned:

I am not necessarily against using Algo or LPs but I wonder if, or to what extent, a liquidation event could affect the price of the LPs being used for liquidation? I think this would be a factor for any non-USDC/STBL2 LP. Also if using LPs, there will be more swaps needed so trading/slippage cost would be another factor… might be worth it, could be reduced with more liquidity, but something else to consider.

18

its not so easy to detect real bad debt i think so no lockup would mean people could just remove their liquidity before the liquidation happens

1 Like
19

Hello, interesting proposal imo but I am not sure I understand everything. So please correct me if my summary is not correct.

  • The idea is to create a LPs vault that could be use as a last resort liquidator in case there is no more bidders.

  • People would earn extra APR for depositing in the vault.

  • Use of the vault for liquidation is not automatic (only the user can do it).

If that’s indeed the proposal I have several questions/ remarks.

  1. I don’t understand why stakers get extra fees. They would already receive a premium for liquidation.

  2. I think liquidation should be automatic in case of bad debt. And then the tokens would be distributed relatively to the shares. If users can choose to liquidate or not, it’s not really an insurance.

Otherwise I think it’s overall a good idea and I would be interested in such a product.

1 Like
20

The idea is to create a LPs vault that could be use as a last resort liquidator in case there is no more bidders.

exactly

People would earn extra APR for depositing in the vault.

yes

Use of the vault for liquidation is not automatic (only the user can do it).

No it is automatic the way liquidations are automatic today. Any user onchain can hit the contract and draw from this fund to execute a liquidation. This can only happen if the position has become under collateralized, not just when it is in the red.

  1. I don’t understand why stakers get extra fees. They would already receive a premium for liquidation.

My answer to the last question may have resolved this. The user’s pledging the LPs may not be the ones that execute the liquidation. As well, the liquidation occurs when the loan is under collateralized so the liquidation is happening at a loss. They are receiving extra rewards to compensate for the risk of that loss. If the DAO properly handles risk, that loss may never happen.

  1. I think liquidation should be automatic in case of bad debt. And then the tokens would be distributed relatively to the shares. If users can choose to liquidate or not, it’s not really an insurance.

Again agreed and I also believe it should be automatic (as in any arbitrager can execute it). Because yeah nobody is going to voluntarily execute a trade at a loss. The goal is to create a pool of liquidity that can be used to backstop bad loans in a crisis. I think this past weekend demonstrated exactly why that is necessary. If USDC can depeg that fast then virtually anything we allow as collateral could.