FIP-10 Curve Integration

Proposals
1

FIP-10 - Curve Integration

Status: Stage 0 - Discussion

Author: @arcology

Summary

  1. This proposal formalizes Curve integration with Fei and opens up the Fei-3Crv pool to the wider community.
  2. This thread also opens for discussion whether Fei Protocol should supply additional liquidity to the pool by providing additional 10k ETH (~4% PCV) to the underlying assets (Fei + 3Crv) and accruing yield back as revenue;
  3. This thread also opens for discussion the amount of TRIBE incentives to be provided into this pool.
  4. This thread starts a discussion on all of the above as a temperature check before advancing to Stage 1 (Snapshot Proposal).

Motivation

With most of the major bugs plaguing Fei out of the way and the recent successes with CREAM and Rari, further integrations with other major protocols will help drive Fei usage. Curve is one of the largest DEXes in the DeFi space with over $6.6B per Defi Pulse and major stablecoin throughput, Fei should also benefit with this integration to drive growth and further use case and adoption.

Per discussion with Joey, he has signaled support for this integration and the Fei-3Crv metapool has actually been created some time ago with approximately $550k of liquidity, but not currently initialized and made available on the Curve UI. External Addresses
Thus, opening it up for all to use creates the following benefits:

  • Allows Fei holders to create yield based on the 3Crv assets;
  • Allows indirectly, a stablecoin diversification of the PCV;
  • Allows a wider use base for Fei;
  • Allows indirectly, a liquidity provision to other stablecoins, but as a consequence, opens up exposure to centralized stables (USDT/USDC)

PCV Investment

One aspect that is open to discussion at this stage is whether Fei Protocol should provide additional liquidity. This would mean taking some additional ETH assets to either:

  1. Directly converting/buying 3Crv and supplying to the pool;
  2. Depositing ETH to a 3rd party protocol (eg Aave) and using ETH as collateral, borrow to provide the 3Crv liquidity required.

Per the model in another proposal written by @Bruno on Lido Staking, I propose that we draw down 10k ETH (about ~4% of PCV, or $20M at ETH $2000 as of this writing), which gives us sufficient amount of leftover reserve of PCV for peg maintenance. This amount is subject to discussion and can be further adjusted with DAO votes to raise/lower the cap.

Rewards Mechanism

Based on our initial discussion on Discord, the Core Team has been working on some code to allow TRIBE rewards to be diverted. In order to bootstrap liquidity provision, I propose that some amount of TRIBE rewards should be used to incentivize this pool, with the intent that the rewards emissions be tapered as time goes on. We can use part of the Treasury to fund the rewards scheme.
Also, the 3Crv pool itself has some nominal yield, so 50% would go directly to Curve DAO while the remaining is retained for Fei Protocol, which can be used for revenue generation for the PCV, with further dividends to TRIBE holders in the future.

Timeline for Implementation

However, please note that per Joey, any code finalized for this process will take about a month, so realistically any formal implementation of this Integration will take place early August at the earliest.

As always, open to feedback for any comments.

Do you support this Fei-Curve Integration?
  • Yes
  • No

0 voters

Would you support an investment into this pool by Fei Protocol’s PCV?
  • Yes, the PCV should be used to buy 3Crv directly (PCV Diversification)
  • Yes, the PCV should be used to borrow 3Crv using ETH as collateral (PCV Borrowing)
  • No, do not utilize the PCV

0 voters

Would you support additional TRIBE incentives for this new Curve pool?
  • Yes - how much to be incentivized would be further discussed
  • No - do not incentivize this pool

0 voters

6 Likes
2

I support this. I am more in favor of borrowing against ETH deposited in a 3rd party protocol. Perhaps even rari. I am also in favor of incentivizing the pool but further discussion is warranted.

1 Like
3

I’m in favor of the pool, but not in favor of using PCV to participate. If fei has a depeg event, nearly all assets contributed to the Curve pool would be held in fei, which would be detrimental to the PCV’s mission of adding stability to fei.

4

Thanks for your responses -

There was some discussion about this as well - @Eswak mentioned that the rate of borrowing will lead to a net loss for us - the borrowing rates for USDT/USDC/DAI is north of 3% on Aave and Compound, unless we can get a preferential rate. Any thoughts on how we can achieve positive cashflow?

Can you please elaborate? The Curve pool would hold assets in 3Crv also so there should be some component of this in each crvFei token? We also would not be deploying a large amount of the PCV so this risk is contained at least.

5

Amazing, on track become a decentralized stablecoin! I would love to contribute to this progress

6

If fei has a depeg event, nearly all assets contributed to the Curve pool would be held in fei, which would be detrimental to the PCV’s mission of adding stability to fei.

I understand your concern regarding the pool becoming 100% Fei in case of depeg event, which would make the PCV allocated become worthless if Fei goes to zero.
However, I think it could actually be a stabilizing force for Fei. The pool would provide a large liquidity to whoever want to sell its Fei because of the depeg. It would suck the Fei out of the market, therefore diminishing the selling pressure. Or am I wrong?

7

Since Curve is designed for pegged assets, it’s the nature of Curve pools to be very tolerant to imbalances within the pool in the pursuit of low slippage. In the event that FEI trades at $0.90, it’s likely that the [FEI-3crv] pool will be almost exclusively FEI. At prices lower than that, the problem becomes even worse. More info at the Curve.fi risks page.

The upshot is that depositing into a [FEI-3crv] pool is not diversifying the PCV in a helpful way. The PCV would become more correlated with FEI in an amplified way.

Much better in my mind is to deposit PCV into an alternative Curve pool like the sAAVE pool or the Compound pool. This would offer true diversification and yield for the PCV.

1 Like
8

I agree that using PCV to supply buy-side liquidity via a [fei-3crv] pool would add a marginal amount of support to the price. If that’s the goal then I could probably get on board.

My understanding is that the mechanics of using PCV to support the price are already established within the FEI protocol. My objection to using PCV for seeding the FEI pool is if it’s done in the pursuit of diversifying the PCV and generating yield for the PCV, which is a different set of objectives. In my view providing PCV liquidity to this pool is a poor strategy for meeting those objectives.

9

If FEI depegs, all liquidity pools become mostly composed of FEI, and that is what allows the protocol to take back FEI from circulation & burn them (where it market makes), returning to the peg. That’s how reweights work on the main ETH/FEI pool, and that could be how it works on the Curve pool, too.

As I see it, it acts as a very strong stabilization mechanism, precisely because of the large imbalance tolerance of the Curve StableSwap. If there is a strong market sell pressure on FEI, the Curve pool could start to get really imbalanced, while the price of FEI remains close to 1$. At this point, the protocol has opportunities to reweight, and take back a lot of FEI out of circulation, while still holding close to the 1$ peg. It can sense the signal before the price dumps.

The reweight conditions of the Curve metapool could be based on a pool imbalance percentage, e.g. if the pool holds 60% FEI and 40% 3crv , anyone can trigger a reweight. That way, the protocol can sense the big sell pressure & take FEI out of circulation before the price of FEI drops far from 1$. It also has the nice side-benefit that the pool is always almost balanced, so whales can perform huge swaps to/from FEI at very low slippage.

Honestly I think Curve should be the main pool where the protocol uses its liquidity to defend FEI’s peg… But that’s not where we are now, with almost no other stablecoins held in the PCV. When we want to get serious about defending a really tight peg (think .995-1.005), I think we’ll allocate a large chunk of protocol liquidity to market make on Curve (and ask their DAO to vote for a higher A parameter in our pool). Time will tell :upside_down_face:

2 Likes
10

It seems that there is broad support for the pool. I agree to the notion that the Curve pool would be one avenue where we could use for peg maintenance. My vision for ‘yield’ on this platform is that it would pay for any capital investment and initial setup fees, as well as to maintain positive cashflow throughout regular maintenance of the pool. Anything above any beyond that is a bonus.

Now, a second topic to be discussed are TRIBE incentives.
Currently, FIP-2 has increased staking rewards for the FEI-TRIBE pool, which intends to run out in 2 years. I propose a couple of options to incentivize this Curve pool, but will need some help from the community to analyze the economics of such:

  1. Divert the current TRIBE drip to the FEI/TRIBE pool into this new pool (say 25-50%). This is so that we can bootstrap liquidity into the new stablecoin pool for furthering peg stability. In this case FEI-TRIBE would be more redundant. [Rewards Diversion]

  2. Keep the current rewards rate for FEI-TRIBE but release an additional stream of TRIBE rewards from the Treasury for the Curve pool. This will allow more governance tokens to be distributed but dilutes the TRIBE circulation and creates further downward pressure on the price of TRIBE. [Additional Rewards Stream]

  3. Buy back current circulating TRIBE at the depressed prices and then use this to fund the incentives for the pool. I suspect this will take the most time to set up (initial buying period) and will be the most expensive, but it may have a potential effect of creating upward pressure for TRIBE prices. [Buyback and Re-distribute]

  4. My least preferred option, but we can mint additional TRIBE for this purpose. [Minting for Rewards]

In all scenarios, based on the initial tokenomics of 1B TRIBE, I would favor about a 4-5% distribution of this into the Curve pool for 2 years (same period as the FEI-TRIBE pool), so about 40-50M TRIBE, or up to about 68,500 TRIBE (about $41k/day at prices as of today) per day.

Again, would appreciate community and core team feedback on the technical and economic feasibility.

Welcome for all feedback and comments.

What type of Incentive mechanism would you support?
  • Rewards Diversion
  • Additional Rewards stream via Treasury
  • Buyback TRIBE and Re-distribute
  • Mint new TRIBE for rewards

0 voters

1 Like
11

Having a large amount of liquidity on Curve would be great for FEI and the peg.

I think we should incentivize this pool by diverting a portion of the existing TRIBE drip to this pool. A large allocation of the rewards to this pool would also make sense, but should be phased in so that there isn’t a large shock to the FEI-TRIBE LP rewards all at once.

As far as utilizing the PCV we can consider both the borrowing option and the direct purchase option. The main question would be how to acquire the stablecoins. The DCA code would make sense, we can go from ETH-USDC on Sushiswap. This would introduce other stablecoins into the PCV but if it is scoped to a target percentage and focused on liquidity provision it should be fine.

Reweights on Curve would be great for the protocol, we can consider the current model of only withdrawing and burning the protocol’s stake. We could then mint and buy when the demand recovers on the other side. This can potentially be something to invest in separately from the initial deployment.

12

I think part of why FIP-6 failed was that there was no purpose for the DAI/USDC purchase, and we talked at length about the benefits for the PCV’s “portfolio volatility” etc, like it was an investment. But we need these stablecoins for liquidity providing and reinforce the position of FEI, not for “investing”. I’m not an expert in central banking, but I think all countries hold foreign currencies to defend the value of their currencies internationally. For us, to defend the position of FEI, I think it would make sense to have a bunch of other stablecoins.

That would be an interesting approach to bootstrap liquidity :

  • Spend part of the PCV to acquire other stablecoins
  • Wrap the stablecoins to 3crv
  • Use 3crv + minted FEI to provide liquidity in our Curve metapool
  • Start a liquidity mining program to bring more people in the metapool
  • As reweights are needed, burn the PCV’s FEI
  • Over time, the protocol won’t hold much 3crv anymore (it will have spent most of them during reweights), but hopefully the metapool will be sizeable by then

During and after the “trial run”, we can monitor how closely FEI sticks the peg. If it is tighter (I think it will be), we will draw the conclusion that holding 3crv in the protocol and actively market make in a Curve pool is beneficial, and we can take action to sustain the number of 3crv held by the protocol for the longer term to keep this mechanism active (like deploying an advanced version of the market maker that can do reverse reweights when FEI is above peg).

I also think this is the best approach. In the short term, we have Fei Labs that will continue to provide enough liquidity in the FEI-TRIBE pool (their LP tokens are vested). So as much as possible, we should use the incentives to guide everyone else towards pools that benefit the protocol more directly.

13

All, the Proposal is now on Curve’s governance forum. Please voice your support!

UPDATE: voting will go LIVE on snapshot on July 2 12:00.

3 Likes
14

Great initiative, @arcology ! Very happy to see all the energy invested here!

I support diverting part of the rewards to this pool.

In relation to PCV deployment, I have some doubts in allocating a relevant amount of it and relying on this pool as the main tool for reweighting. One of the reasons for holding FEI, a decentralized stablecoin, is to be less exposure to the risks USDC and USDT have. By providing liquidity to this pool, we would indirectly hold these two stablecoins and their related risks.

This impact of a depegging could be amplified as @jcliff pointed due to the Curve pools nature:

"Permanent loss of a peg

If one of the stablecoins in the pool goes significantly down below the peg of 1.0 and never returns to the peg, it’ll effectively mean that pool liquidity providers hold almost all their liquidity in that currency."

If, for example, USDT has problems in the future, we would stay with worthless USDT in hands. So, if any of the stablecoins in the pool have problems, USDT, DAI or USDC we would be considerable exposure to the most problematic. I think the reason why the decentralized stablepool has been widely discussed in the market is to provide an alternative to the risks from the centralized stablecoins.

I think these risks reduce FEI value proposition as a decentralized stablecoin.

So while I think this pool is important for FEI, I am not sure about PCV deployment, but would support some TRIBE rewards for it.

1 Like
15

@Bruno, Thank you for your responses.

I am of the opinion similar to @Eswak and @trx314 about using this as a stability mechanism. While the concerns about centralized stables (USDT in particular) is well noted and discussed, for Fei to be the stablecoin for Defi we will also need to recognize that we need direct on/off ramps to them, rather than the current status quo ‘Cstables->ETH->Dstables’. In fact, you can relate USDT as a fractional-reserve backed stablecoin, kind of similar to another decentralized project we all know of…

As with previously mentioned, if we want to take a ‘reserve bank’ approach for Fei, then we will have to hold some amount of other stables, and unfortunately by design the Curve integration with their usage of 3Crv means that holding USDT/USDC/DAI is part and parcel of this necessity.

To help limit our risk I am suggesting that we hold a small amount of PCV, at a proportion even less than FIP-6. I think if the objectives of FIP-10 are met (we provide liquidity to other stables via Curve, and how much yield we can capture for the PCV and for Fei Protocol), then we can re-evaluate whether we should add or remove PCV liquidity further down the road.

1 Like
16

To reflect some earlier discussions with @joey on Discord, it was suggested that the FEI-TRIBE rewards will be diverted so that approx 40% of the eventual rewards pool (70M TRIBE) will go into this Fei-3Crv pool.

Another question is to decide on the timeframe for the redistribution of the rewards. The 2 options discussed were:

  1. A gradual phase-down of TRIBE rewards to the FEI-TRIBE pool tied with the same phase-up of rewards to the FEI-3CRV pool, over a period of 2 months to reduce the shock of sudden liquidity shifts: 100% → 70% → 40%. [Phased Rewards Reallocation]
  2. Instantaneous reward pool shift from FEI-TRIBE to FEI-3CRV at a given date [Instant Rewards Reallocation]

All of this is contingent on the implementation of the MasterChief Rewards system.

How would you like the TRIBE incentives to be Redistributed?
  • Phased Reallocation
  • Instant Reallocation

0 voters

17

As a Curve community member and a large veCRV holder, I wasn’t aware about the snapshot vote that occurred (not sure how widely it was publicized)? With that said, Fei can always create a factory pool with 3crv as a base:

and potentially be listed on the main Curve UI once you meat the following requirements:

Pools which meet the following requirements will be listed on Curve main UI.

  • At least one audit

  • Marketcap of at least $3m

  • Fortnightly volume over $3m

  • Pool liquidity over $500k in 3CRV

  • Pools on Factory v2 only

See https://resources.curve.fi/faq/pool-factory#pool-factory-curve-listing-requirements

1 Like
18

Hi @glenn, welcome to the community.

The initial proposal was to simply list our metapool (already established, just not listed) onto the Curve UI but unfortunately the signal vote did not reach quorum.

For next steps, we are going to try again with a more robust proposal, which needs voting in the Fei DAO, but more or less in the following manner:

  • Fei Protocol purchases an amount of CRV (still internally discussing how much, but initially the quorum percentage) to be locked into veCRV which would get us to quorum.
  • Fei Protocol would make a purchase using 10k ETH (about 4% PCV) for 3crv using the PCVDripper;
  • Fei Protocol would (heavily) incentivize the Fei-3Crv pool to bootstrap further liquidity and trading volume;
  • Using our resources, petition the Curve community to allocate a gauge and further CRV rewards weighting to the new FEI-3Crv pool.

I think the core team will explore the technical implementation of the Dripper module, the CRV purchase amount and from which markets we could purchase from, and the rewards module.

Open to more suggestions and ideas!

1 Like
19

I’ve done some more thinking on this, I believe a smaller purchase makes more sense, as allocating over 30m to reach quorum on this doesn’t seem like a wise investment with such a long time horizon in a rapidly evolving landscape. It can simply be a show of good faith and can help us get on the map in some form maybe $2m for example.

Has anyone done any math on how many rewards we could hypothetically purchase for either ourselves or FEI holders by making the CRV investment?

Fully in support of this, I think using a DAI bonding curve to bootstrap the liquidity will make the most sense.

20

I’d also proffer for someone with more economics knowledge to help to calculate this. Conceptually, the more veCRV you have, the more voting power to reweight a gauge with a higher CRV emissions proportion to the project at hand (in this case Fei).
From there, the larger veCRV holdings would allow you to get a higher proportion of CRV rewards for as long as the veCRV is locked (typically 4 years).

One of the problems as demonstrated by the snapshot vote at Curve was that there was no quorum reached, which means that the big veCRV holders are not voting, nor are they incentivized to do so. In this case I would ask the core team (perhaps @Brianna could lead this) to talk directly to Curve and the wider Curve community to do a much higher profile marketing campaign to rally support for this.

My intention of buying larger quantity of veCRV is for capital appreciation, and reduces the uncertainty of getting additional stakeholders to buy in to vote for Curve listing and further gauge acceptance, as the latter will definitely require a DAO vote. If required, we can hopefully redeem our costs by selling CRV at the end of the lockup period.

Therefore, to mitigate the risk of lack of quorum (and lack of wider support) then we need to really spend on a marketing campaign, leverage the core team’s VC connections to spur buy in on Curve, or simply buy enough CRV to gain quorum by ourselves. A lack of integration to Curve (and its derivative projects) is a big chunk of the stablecoin market we simply cannot miss.

Perhaps others in the community can pitch in on the strategy we should adopt for purchasing CRV (or not at all).

1 Like