[Proposal] 🌲 Delegating some PCV management to TRIBE staking pools

This proposal is probably a longer term proposal (quite some code to write), but all the proposals I read so far can be accomplished with this one.

Background

We are seeing a lot of proposals here about what should be done with the PCV :

  • A redeem mechanism in FEI for ETH
  • A redeem mechanism in FEI or FEI + TRIBE for ETH
  • A redeem mechanism in fixed $ amount
  • Farm in a Yearn vault
  • Liquidity in a Balancer pool
  • Liquidity in a Curve pool
  • Stake in LIDO
  • ETH Options
  • etc…

It is good that some actions are taken globally by DAO, at the protocol level, but I also believe there should be some room for competition and “permissionless innovation” in the management of the PCV, to discover what works and what doesn’t, at what thresholds, etc.

What is your proposal ?

The proposal is to allow TRIBE holders to borrow value from the PCV at no interest rate, using staked TRIBE as collateral for the loan.

Like a forest :evergreen_tree::evergreen_tree::evergreen_tree:, TRIBE holders will allow multiple management strategies to grow.

What does it accomplish ?

  • It gives a first utility to the TRIBE token beyond governance.
  • The staked TRIBE contributes to improving the collateralization of the protocol.
  • PCV is partially protected from volatility of the underlying assets (currently, just ETH).
  • It allows fast iteration & competition in the management of the PCV.

How is it accomplished ?

On the protocol level, a new loan mechanism is implemented :

  • A portion of the ETH PCV is assigned to a new EthPCVLoanManager
    • Determined by DAO vote, e.g. 20%.
    • There could be multiple LoanManager once PCV is diversified into multiple assets.
  • TRIBE can be staked to borrow ETH from the PCV.
    • LTV (loan-to-value) thresholds are determined at protocol level, by DAO votes.
    • E.g. 3$ of staked TRIBE allow for 1$ loan from the PCV (33% initial LTV).
    • E.g. Liquidation can happen if LTV raises above 70%.

The net effect for the protocol is : collateral increases by ~2$ for each 1$ borrowed (1$ taken out, 3$ staked), and the volatility risk/opportunity of ETH is transferred to the borrower.

If the LTV raises above a given threshold, the position can be liquidated to the benefit of the caller and the protocol (at the expense of the TRIBE staking borrower).

The loans are denominated in constant dollar value, and can be paid back either in FEI or in the borrowed asset, ETH (for the protocol, gaining ETH or burning circulating FEI is the same).

Borrowers could be large TRIBE whales, or smart contract systems aggregating the TRIBE stake of smaller TRIBE holders.

Example 1 : ETH value soars (opportunity cost to the protocol)

In this scenario, the value of TRIBE stays the same, so the LTV stays at 33% since loans are denominated in fixed $ value.

  • 3M$ of TRIBE are staked to borrow 1M$ of ETH from the PCV
    • PCV increases by +2M$ (-1M$ of ETH, +3M$ of TRIBE).
  • ETH doubles overnight :exploding_head: the borrower now has 2M$ worth of ETH.
    • PCV is still at +2M$.
  • Borrower pays back the loan using half their ETH, and keeps 1M$ of free ETH.
    • PCV is at +3M$ (-1M$ of ETH +1M$ of ETH + 3M$ of TRIBE)
  • Borrower unstake their TRIBE.
  • PCV is the same as before the loan.
    • The protocol did not profit from the raising value of ETH.
Example 2 : ETH value crashes (the protocol is protected)

In this scenario, the value of TRIBE stays the same, so the LTV stays at 33% since loans are denominated in fixed $ value.

  • 3M$ of TRIBE are staked to borrow 1M$ of ETH from the PCV.
    • PCV increases by +2M$ (-1M$ of ETH, +3M$ of TRIBE).
  • ETH crashes -50% overnight :scream: the borrower now has 0.5M$ worth of ETH.
    • PCV is still at +2M$.
  • Borrower needs to buy more ETH or FEI to pay back the loan.
  • Alternatively, a mechanism to market sell the staked TRIBE for FEI and reduce the value of the loan could be implemented.
Example 3 : TRIBE value soars (improve protocol collateralization)

Nothing special to say here. At the time of taking the loan, LTV is at 33%. If the value of TRIBE increase, the borrower can borrow more ETH from the protocol. If the borrower does not increase the size of the loan, the rise of TRIBE value benefits the protocol by increasing the collateralization ratio.

Example 4 : TRIBE value crashes (liquidation)
  • 3M$ of TRIBE are staked to borrow 1M$ of ETH from the PCV
    • PCV increases by +2M$ (-1M$ of ETH, +3M$ of TRIBE).
  • TRIBE value crashes -50% overnight :scream_cat: LTV is now 67%, better close the loan soon !
  • TRIBE value crashes even more. LTV crosses the >70% threshold :frowning_face: let’s say that the staked TRIBE now only are worth 1.25M$.
  • Anyone can send 1M$ to the protocol (in ETH or FEI) to get the 1.25M$ worth of staked TRIBE.
  • The address who liquidated the 1.25M$ worth of TRIBE can market sell them to get >1M$ of ETH or FEI back instantly, making a nice profit while keeping the protocol safe.

Who does it benefit ?

  • It benefits FEI users because the collateralization is reinforced.
  • It benefits FEI users because the price will stay closer to the peg (see below).
  • It benefits TRIBE holders because it gives an additional power to their token.
  • It benefits FEI & TRIBE holders because more integrations with other protocols will exist.
  • It benefits FEI & TRIBE holders because they can stake their FEI or TRIBE in multiple competing strategies that will give staking rewards.

What are the costs and risks ?

  • It will require new developments, and with new code comes new risks. I suggest the Fei Labs team is closely following the development of that feature (or that they do it themselves), and that Fei Labs finances audits of these core protocol changes.
  • If the TRIBE value drops 99% and liquidation does not happen, or if no one operates liquidation bots, some PCV could be lost.
  • TRIBE staking pools (borrower smart contracts) do not present a risk to the protocol, since all loans are safeguarded by a collateral of TRIBE. But TRIBE stakers could be at risk if they ape into a flawed/scam/unaudited strategy.

Why should the community vote for this ?
Because it’s awesome !

  • TRIBE tokens will help secure the FEI stablecoin.
  • TRIBE tokens can be staked to earn rewards in multiple strategies.
  • TRIBE tokens can vote to implement the successful strategies happening at small scale (and protected by TRIBE staking) to be part of the core protocol strategy.
  • Anyone can build anything and I’m sure we’ll see some unexpected uses of the PCV.

Here are 3 smart contracts that I would like to code using the new TRIBE Loan feature :

Arbitrage of FEI when price is below peg ⚠️READ THIS⚠️

Essentially, this proposal allows staking TRIBE holders to always be able to redeem FEI at the exact peg price, so staking TRIBE holders can perform risk-free arbitrage and stabilize the price of FEI :

Assuming 1000$ of TRIBE staked :

  • Take 300$ ETH loan from the protocol.
  • Instantly swap those 300$ of ETH to > 300 FEI.
  • Pay back loan with 300 FEI.
  • Keep the extra FEI.
  • Take a new loan, repeat until FEI price reaches 1$.

Staking TRIBE holders can profit from underpeg FEI and reduce the need for reweights.

Price floor of .995 DAI to redeem FEI tokens

Assuming 1000$ of TRIBE staked :

  • Take 300$ ETH loan from the protocol.
  • Instantly swap those 300$ of ETH to 299 DAI.
  • Allow up to 300 FEI to be redeemed for 298.5 DAI.
  • Pay back loan with 300 FEI.
  • Use the remaining 0.5 DAI to buy TRIBE on the market.

The pool’s quantity of TRIBE will grow, offering staking rewards for TRIBE stakers.

Curve liquidity farming

Assuming 1000$ of TRIBE staked :

  • Take 300$ ETH loan from the protocol.
  • Instantly swap those 300$ of ETH to 299$ worth of other stablecoins.
  • Provide liquidity on Curve.
  • Accumulate CRV tokens and trading fees.

Pay back the loan with the earned interests.

Anyone can code any other of the strategies proposed (redeem FEI for X, farming in yearn/balancer/lido/…). Some strategies could decide to give some (most?) of their profits back to the protocol, instead of individual TRIBE stakers, and the DAO could vote to incentivize those strategies by distributing TRIBE from the 40% DAO treasury. Or just replicate the strategy using the main PCV pool.

28 Likes

This is undoubtedly a great idea, and I support it!

8 Likes

Few people can make such specific suggestions. You must have paid a lot for this. The community should thank you. Maybe some airdrops are right for you.

7 Likes

This is an amazing idea, and will eliminate the headache that will inevitably come when TRIBE tries to govern use of the PCV. Instead of having to sort through a million different ideas you get to go do it yourself(and profit yourself, and increase PCV)

This is the perfect decentralized approach to PCV diversification. Now to work on implementation/feasability…

6 Likes

Great idea!

If I squint, this looks to me like a multi-strategy TRIBE Vault. Is there any reason why FEI cannot be used as the collateral instead? Should a governance token be used as collateral? How will the yield return to the PCV, and in what format? As the PCV all consists of ETH, it will be interesting to see how the addition of other assets will affect it.

1 Like

I need more time to fully understand this but I am following :slight_smile:

9 Likes

Fei Protocol please hire Eswak. He has been quite contributing, and this should be rewarded. There are other fellows too. I suggest making sth like a core group. These are the powerful resources.

11 Likes

Amazing, and really concrete idea.

2 Likes

Eswak! This is wonderful work! It looks like you’re managing to synthesize the concerns of many stakeholders into one proposal. There are a lot of moving pieces. And I imagine that under inspection this will need to be tweaked in one way or another. But in principle, I’m on board with the idea of giving loans with tribe as collateral. It’s simple. It’s in the spirit of using PCV creatively rather than as collateral. And, if you’re right, it could solve many of the problems that the protocol is facing at the moment. Thank you for doing this work! And I hope you get compensated for it beyond the compensation you’ll be able to glean from implementing the strategies you describe near the end.

At the moment, only one potential vulnerability comes to mind. And it’s likely that you’re already implicitly assuming the solution. So I’ll start with describing the solution. Tribe staked as collateral for a loan can’t be used for on chain votes.

Otherwise, I could do the following. (I’ll be ignoring fees just to illustrate my point. Fees shouldn’t make a huge difference for sufficiently large versions of this scheme). Stake $1,000 TRIBE in exchange for $300 ETH. Buy $300 TRIBE. Stake $300 TRIBE in exchange for $90 ETH. Buy $90 TRIBE. Stake $90 TRIBE in exchange for $27 ETH. And so on. In the limit I’ll accrue as many as

\frac{1}{1 - 0.3} = 1.42

dollars worth of TRIBE for every dollar worth of TRIBE I put in. If I can vote with my staked TRIBE, then I’ve increased my voting power by 42% without spending a penny.

And, of course, being able to vote on the liquidation ratio of your own loans might cause problems.

6 Likes

I love this. A decentralized yet flexible way to utilize the PCV without having to vote on every investment decision.

Just a few thoughts: I’m not sure if each loan increases the collateral (in the sense of PCV that can be used to back FEI), since the collateral for each loan should always be always redeemable by the borrower. Or maybe you are thinking that in a crisis, the protocol can pause repayments and use the borrower’s collateral to redeem FEI.
But perhaps the PCV can charge a small interest rate that is lower than the market rate, so that the interest helps the PCV grow in the long run. And this could offset the risk of default. Default would not directly lower the value of each TRIBE since the staked TRIBE is burned, but it would lower the size of PCV.

Also, wouldn’t allowing TRIBE holders to redeem FEI at $1 make the protocol quite similar to a reserve-backed stablecoin?

3 Likes

Brilliant thinking here. Love the clarity and fully support the idea.

1 Like

Good point. Maybe we want the staked TRIBE to be ineligible to vote. Not sure about the technical details, but don’t you currently have to have TRIBE in your wallet to vote? So unless we take extra steps to give the “debt tokens” voting power corresponding to the amount of TRIBE that they represent, they would not be able to vote on DAO votes, just as TRIBE-FEI liquidity providers cannot?

1 Like

I definitely am liking this idea. I would definitely be interested in seeing something like this implemented. Great stuff as always, @Eswak!

1 Like

Incredible work! I really appreciate YOU putting large amounts of effort into finding elegant solutions to make this a great protocol . B/t you and fei_saver, I have strong conviction for this project moving forward by this (among other) remarkable proposals that I really hope the core team will soon acknowledge and take action on. This is really awesome ! Thanks again.

1 Like

This is a great idea and I support it, however I don’t think any headway can be made on this until we get past the current crisis and also manage to maintain peg stability.

Oh I see. I was confused. I thought that you could vote on DAO votes with staked TRIBE. But I had just misread this part of the FAQ:

Q: Can I vote with my staked TRIBE?
A: Yes, you can vote with your staked TRIBE in the snapshot vote (only for signaling purposes) but not in the DAO vote which is the final vote

1 Like

I think your second point is right. For example, the arbitrage strategy (as described in the original post by @Eswak) amounts to the protocol spending $300 worth of ETH on 300 FEI. It does it indirectly, so it looks like it’s not using PCV as collateral, but on closer inspection it is doing exactly that. I still think the idea of using tribe as collateral for loans of the PCV is a great idea. But I’m not sure it’s a good idea to let the loans be paid in FEI at a valuation of $1 regardless of whether FEI’s on peg. Of course, without this feature, it’s also not possible to implement the price floor at 0.995 DAI. @Eswak, I imagine you already gave this some thought. What do you think?

:heart: haha. I’m not so anonymous, and if you stalk me you’ll see I have a pretty sexy day job already. But I’d love to contribute to the protocol & get more TRIBE in a way or another, this project rocks.

Yeah, definitely a complex feature that we need to take some time to mature & discuss. Some obvious question I see : what if a strategy becomes so popular that no capital remains to be borrowed by other strategies ?

I agree it’s not really “PCV”. It is controlled by the protocol, in the sense that as long as the loan is open, the staked TRIBE are held captive, but it is not controlled in the sense that it can be moved & used as the protocol sees fit. It’s part of the TVL (securing the protocol), but not part of the PCV.

But what matters most IMO is that it increases the protocol’s ability to secure the peg (by transferring volatility risks of its collateral). We won’t find many hedges against a drop of the value of ETH that are as cheap & effective.

Liquidation would lower the price of TRIBE because the liquidator will get the staked tribe as compensation for keeping the protocol safe & probably market sell them to get their spent ETH/FEI back.

I’m not sure of the exact implementation, but I had a quick look at the dao contracts and it seems that vote can be delegated (and a contract holding 100 TRIBE can delegate 50 TRIBE of voting power to an address, and the remaining to others). That’s what they use for the timelocked (vested) tokens. Therefore, I think the staking (loan) contracts could delegate back to the staking addresses. I think it’s desirable because we don’t want TRIBE unstaked every time there is a vote.

Why would that be a problem ? The TRIBE does not come from nowhere, and if someone cares enough about the protocol to go on a leveraged position on its governance token, why not let them vote ? :stuck_out_tongue:

If the protocol does not believe FEI is worth 1$, who will ? :stuck_out_tongue:

Joke aside, it is true that it makes FEI a reserve-backed stablecoin. I am pretty convinced that this peace of mind is required for stability (which does not mean 100% of the reserve should always be available to back the stablecoin). But it’s just an opinion.

Fei protocol is still much more than just a reserve-backed stablecoin. The remaining PCV can be used for a variety of other things. Keep in mind that loan system would only be allocated part of the PCV. I read somewhere that keeping ~20% of AUM liquid for people wanting in / out is considered safe. That means long term, probably around 20% of the PCV will be used to back FEI on the market. We still have to figure out what to do with the remaining :slight_smile:

4 Likes

Two vulnerabilities come to mind: (1) it would reduce the cost of a 51% attack. It would become a 36% attack! Hehehe! Sure, there’s an argument that this would be good for tribe holders. It raises the secure price floor from around 2 times the PCV to around 3 times the PCV. But that “secure price floor” is only a floor in theory. We don’t know how well it will support the market at the moment because TRIBE isn’t fully liquid yet. And a lot of the tribe that can already be voted with (read: the tribe owned by the team and seed investors) cannot be sold. We’ll see what happens in the future. (2) The bank is owned by its debtors. All I can say about this is that it feels risky! :sweat_smile: On the other hand, some people might read “the bank is owned by its debtors” and think that it sounds like the epitomy of DeFi. I haven’t come to a conclusion about this. I’ve tried to come up with strategies that would abuse this power, but I haven’t been able to come up with any. I’ll wait for people that are better at breaking things than I am to give it a try.

I agree with you here. I think many of us (myself included) have a bit of a dogmatic resistance to the idea that FEI would be reserve-backed. In reality, the ecosystem would probably benefit from having a faucet (even if an indirect one) where people can sell FEI for ETH at $1. And what’s nice about your solution is that the size of that faucet will be dynamic, but capped at 20% of PCV in the case that the only use of the loans is arbitrage.

Stake FEI to borrow ETH ? I have hopes that, down the road, that will be possible on Aave/Compound/etc, and that would be better for the protocol because the borrowed ETH wouldn’t come from the PCV in these cases.

Technically, that would also be possible. It would also protect the PCV from volatility. And it secures the protocol in a similar manner because it locks circulating FEI for a smaller quantity of PCV (even smaller than that, though, is if FEI is locked in 3rd party protocols like Aave or Compound).

If the protocol always believe that FEI is worth 1$, the collateral wouldn’t fluctuate in value, so that would also be a lower risk loan to take for the borrower : the protocol could charge a bit of interests for that.

I like the idea of staking TRIBE better though, because it’s supposed to be “the token that controls the PCV”. But maybe setting a 2nd lending pool where FEI is staked to borrow would prove to be a fruitful experiment ? In the end, the protocol takes minimal risks by doing this, so it’s worth trying to see if that creates value for the protocol. If the code is written for TRIBE staking, the code could be adapted for FEI staking at minimal effort.

I don’t like the idea of charging interest rate for TRIBE staking though. But that could just be a parameter adjustable by DAO vote, and we can try with and without interests to see what works out. As you noted, some use cases are not possible if the protocol charges interests.


One thing that came to my mind : we probably want a “maximum duration” for the loans, after which the loans could be liquidated. For instance, 1 month. That forces borrowers to repay & reborrow their loan on a regular basis, and if we decide to shut down new loans, for whatever reason, we know the PCV will come back to the protocol within a month.