IIP-XY: MVI Liquidity Migration to Visor Uniswap V3 Pool

IIP - XY
Title: MVI Liquidity Migration to Uniswap V3 Visor Pool
Status: Proposed
Author: @Matthew_Graham
Created: 26/11/2021

Simple Summary

After INDEX incentives on Uniswap V2 MVI-ETH came to an end, very little liquidity migrated across to Uniswap V3. We now find ourselves with a very small MVI:ETH Uniswap V3 pool, ~$1M TVL and ~$5.4M on Uniswap V2. This proposal presents an opportunity to spend $76K in INDEX and $42k in VISR to reimburse gas costs and provide a short 4 week incentive campaign to grow the Uniswap V3 MVI:ETH liquidity via a Visor pool.

Liquidity will be directed to the Visor’s MVI:ETH pool where deposited funds will be actively managed and optimised to minimise slippage for traders. Visor pools deploy capital into the 0.3% swap fee Uniswap V3 MVI:ETH pool and actively manages the liquidity thereafter. All rebalancing gas costs are paid for by Visor.

Abstract

This proposal presents the Index Coop community the opportunity to offer gas reimbursement payments to those that deposit into the Visor MVI:ETH pool. There are three transactions flows that are reimbursed which as detailed below:

  1. Remove their LP token from the staking contract, retrieve their liquidity from the Uni V2 pool and deposit their MVI & ETH tokens into the Visor MVI:ETH pool.

  2. Retrieve their liquidity from the Uni V2 pool and deposit their MVI & ETH tokens into the Visor MVI:ETH pool.

  3. Deposit their MVI & ETH tokens into the Visor MVI:ETH pool.

This proposal is to directly reimburse gas costs through distribution of the INDEX token via Parcel. Parcel offers free payment transactions. The Finance Nest will perform the appropriate analysis and oversee the distribution of payments.

In addition to reimbursing gas cost, this proposal presents a short sharp liquidity incentive campaign for 4 week duration offering a 20% APR on $5M Visor MVI:ETH pool. Index Coop will send 2,100 INDEX tokens to Visor, Visor will match Index Coop’s contribution of $42K and both INDEX and VISR will be distributed across the Visor MVI:ETH pool.

The total budget for this proposal is $76K, to be paid for with INDEX. All of the financial transactions relating to this proposal are to be performed from the Operations Account. Visor will also be spending $42K incentivising the pool and will incur all the gas costs in setting up and maintaining the pool’s liquidity distribution.

Motivation

Since LM rewards finished around 20th September 2021, around 45% of the MVI:ETH LP tokens have remained deposited in the staking contract. The 45% that remains in the staking contract, makes up 69% of the current MVI:ETH Uniswap V2 pool, or around $3.5M.

Index Coop debated the influence of MVI:ETH LM rewards extensively and it turns out the main benefit of LM is to draw capital into the contract. Once the capital is in the contract, in this instance the Uniswap V2 MVI:ETH LP staking contract, the APR needed to retain the user is low. With 2 months passing since LM rewards ceased, only 55% of the capital has been withdrawn.

We know from trying to migrate liquidity from Uniswap V2 to Uniswap V3 staking contract that incentives are somewhat effective. Reimbursing gas indirectly through liquidity mining rewards is easier to implement but not as effective as directly reimbursing gas cost. LPs that don’t transfer quickly to the new staking contract risk not being able to earn back their gas costs. By directly reimbursing LPs gas costs for moving capital around, we make sure Index Coop’s capital is more efficient by being very targeted.

However, we want to improve the Uniswap V3 MVI:ETH liquidity asap and for this to happen we need to incentivise people to act quickly. In an attempt to get people to act quickly, there needs to be a mechanism for rewarding early depositors. We know that once deposits are in the Visor pool, the trading swap fee APR and the gas costs in exting the pool will be a natural retainer. The swap fee income is greatest when the aggregators are running their trades through the Uniswap V3 pool and therefore we need to grow this pool such that it becomes the most efficient trading pool.

A short burst of incentives should encourage depositors to act quickly. The joining incentive campaign by Visor and Index Coop over a 4 week duration should be enough to attract a target $5M of liquidity into the Visor pool. The APR will start off very high and then will decay as more capital enters the pool. When the pool reaches $5M, the APR should be around 20% and the initial $1M deposited will earn around 100% APR. The high initial APR is a great marketing tool for promoting and encouraging people to deposit their MVI and ETH into the pool.

Visor 2 is an active liquidity management protocol. Concentrated liquidity pools bring increased capital efficiency, but they also bring increased sophistication and management. Visor takes on the overhead of managing liquidity such that it stays effective and continues earning trading fees. Visor takes 10% of earned fees, which get distributed to VISR stakers - the other 90% of earned fees will be automatically invested into the LP position upon each rebalance. Visor covers all gas fees for rebalancing positions and re-investing earned fees.

Funding

In order to implement this proposal, I developed the following cost estimate:

  • Eligibility: Depositors in the Index Coop Uniswap V2 MVI:ETH LP contract
  • Eligibility: Deposit liquidity in the Visor MVI:ETH pool within 30 days of Index Coop announcing liquidity incentives
  • Retrieve Uniswap V2 MVI:ETH LP token from Index Coop staking contract (0.0069 ETH rounded up)
  • Remove MVI and ETH tokens from Uniswap V2 MVI:ETH pool (0.0108 ETH, rounded up)
  • Approve depositing into Visor Uniswap V3 MVI:ETH pool (0.0074 ETH, rounded up)
  • Deposit into Visor Uniswap V3 MVI:ETH pool (0.0282 ETH, rounded up)

Whilst reviewing the various transactions on etherscan above, there is some variance in the ETH spend. The above transactions equate to around 0.0533 ETH and due to the variance in transaction costs, it would be prudent to draw a higher budget estimate of say 30%. Ie: 0.06929, rounded up to 0.07 ETH. Do note this is an estimate and Index Coop will only be reimbursing actual costs up to 0.06 ETH per LP wallet. This provides a cost ceiling per LP depostor and should lead to some funds being unspent.

Based upon my calculations here, there are around 78 unique wallets that hold more than $100 USD in the Index Coop MVI:ETH staking contract. 78x0.07 = 5.46 ETH @$4,300 per ETH, that is $23,478 which is approximately $24K. A snapshot of the eligible wallet address was taken on the 26th November 2021. Ie: see spreadsheet attached for eligibility.

  • Eligibility:
    • Passive MVI & ETH holders
    • Uniswap V2 LPs not deposited in the Index Coop staking contract
    • Uniswap V3 LPs
    • Deposit liquidity in the Visor MVI:ETH pool within 30 days of Index Coop announcing liquidity incentives
  • Remove MVI and ETH tokens from Uniswap V2 MVI:ETH pool (0.0108 ETH, rounded up)
  • Deposit into Visor Uniswap V3 MVI:ETH pool (0.0282 ETH, rounded up)

To encourage non Uniswap V2 MVI:ETH LP stakers, who are already providing liquidity or just holding MVI and ETH, we offer a budget of $10K to reimburse gas costs for migrating liquidity from Uniswap V2 and depositing into the Visor MVI:ETH pool. This will be administered on a first come first served basis. LPs who remove liquidity from Uniswap V3 will remove liquidity at their own cost and shall only receive a refund for gas spent depositing into Visor MVI:ETH pool. A refund of 0.039 ETH per wallet is recommended. A snapshot of the Uniswap V2 LPs was taken the 26th November.

In an attempt to accelerate this process of creating liquidity within the Visor MVI:ETH pool and also attracting fresh capital into the pool that isn’t already providing liquidity to the MVI:ETH pair, I recommend a short burst of incentives. A $42K spend provides a 20% APR over two weeks on a $5M pool. The Uniswap V2 MVI:ETH pool is around $5M in TVL.

The total cost is $24K + $10K + $42K = $76K

The first $1M of capital to deposit into the Visor pool will receive a juicy >100% APR. I believe this is sufficient to create hype/interest and get people’s attention with the added kicker it is free to deposit MVI and ETH tokens into the vault.

The Uniswap V2 MVI:ETH LP staking contract started receiving tokens on the 7th April 2021, the peak number of LP tokens in this contract was around 24th April which is a little more than 2 weeks after launch. Two weeks is a lot shorter than the promised runway of incentives compared to the launch promise of MVI. However, the gas cost reimbursement should be very attractive for small token holders in this environment. We will also need to create sufficient visibility around the APR deposits in the Visor vault generated from swap fees and utilising more active liquidity bands.

Pool Visor MVI:ETH
Decentralised Exchange Uniswap V3
Target Pool Size $6M on Uniswap V3, $5M in the Visor MVI:ETH pool
Target APR 20% at end of campaign
Total USD Cost 76,000
QTY INDEX 3,800 (2,100 to be distributed across depositors in Visor pool)

Success Criteria

Increase the size of the Uniswap MVI:ETH V3 pool to $5M which enables a 11 ETH, ~$43K, trade to incur 2% slippage or 4.6 ETH, ~$17.9K, trade at 1% slippage. The pool SALE:ETH is $5M in size and was used to provide the figures mentioned above.

Specification

  • To use the Visor’s hypervisor to actively manage liquidity in the Uniswap V3 MVI:ETH (0.3% fee) pool
  • Operations Account to transfer 2,100 INDEX to Visor to distribute across deposits into MVI:ETH Visor pool
  • Stakers, Uni V2 LPs, Uni V3 LPs and passive MVI and ETH holders will need to act accordingly
  • Finance Nest to monitor transactions and refund gas cost at the end of the 30 day period

Voting

FOR:

DO proceed as detailed above, $76,000 spend, reimbursing gas and providing incentives for depositing capital in the Visor MVI:ETH pool

AGAINST

DO NOT proceed as detailed above, $76,000 spend, reimbursing gas and providing incentives for depositing capital in the Visor MVI:ETH pool

Copyright

Copyright and related rights waived via CC0.

6 Likes

Managed Uni V3 pools seem a better option than relying on holders to spin up their own positions, especially on L1. Looking forward to the initiatives here, I expect we’ll see their effect pretty quickly. The spend seems easily justified given the product’s state and trajectory.

1 Like

Hi @jdcook,

With the formation of the liquidity pod and this proposal request spend through a mix of gas reimbursement and incentives, how shall we proceed with this proposal?

@DarkForestCapital @verto0912

Thoughts on this proposal as it relates to MVI ?

So at a high level, we agree that something needs to be done to address liquidity for MVI.

These numbers don’t quite cover the full extent of the situation. Since incentives came to an end, we have lost nearly 80% of MVI liquidity based on the number of MVI units in the liquidity pools across different DEXs.

You’ll notice that the withdrawal of liquidity accelerated around the 28 of October, the day of the Facebook name change. As MVI began to meaningfully outperform ETH, impermanent loss in the pool became painful for LPs.

We would also agree that V3 is the best option for MVI. But the execution numbers below are off.

A v2 pool with $5m in liquidity enables 11 ETH trade with 2% price impact. A v3 pool should be much better, especially when liquidity is actively managed. If we are not getting a much better execution, then providing incentives makes less sense.

In terms of Visor vs. Gelato, we would personally lean towards Gelato. This mostly has to do with the recent oracle exploit at Visor vs. Gelato’s G-UNI tokens being used by Maker as collateral and currently going through the Aave process. Further, we are unsure if there’s a need for active management of liquidity for pairs like MVI, which should be highly correlated with ETH and face limited IL. Do we have any evidence that this would outperform a relatively conservative range position on a net basis (after Visor’s 10% cut)? Perhaps a fairly conservative range on Gelato makes more sense and gives us the ability to do short-term incentive campaigns for more concentrated ranges around a particular market event. For instance, when we saw significant trading volume and interest in MVI in November, we could’ve incentivised a more concentrated range to unlock very deep liquidity for the market. From my understanding, incentivising G-UNI pools is much easier for us as it requires a simple staking contract that the IC would fully control.

Regarding timing, now is probably a pretty decent time. Some of the hype has cooled off so IL shouldn’t be too significant. Trading fees are going to be less than over the last 4-5 weeks, but still meaningful. We are probably looking at an average daily trading volume of $1m or ~22% APR from fees. Boosting that with another 20% in incentives would be attractive to large LP providers, but probably not retail + smaller holders. Which could be fine if we can get guys like this to transition to v3.

To sum up, we think the proposed IIP could help address MVI’s liquidity problem. However, the decisions around execution and incentives are fully in the hands of IC.

2 Likes

Hey @verto0912 thanks for your response. I’m Brian from Visor Finance and I just wanted to address a few of the points that you made.

As MVI began to meaningfully outperform ETH, impermanent loss in the pool became painful for LPs.

Hey I totally agree with your assessment here that impermanent loss should be a big factor in having an active manager on Uniswap v3. With respect to the MVI-ETH pool that we are currently managing, we are utilizing the AutoRegressive strategy, which takes into account price prediction and past historical volatility in determining ranges. How this works in practice is that as soon as the algorithm detects volatility, it widens the range extremely wide to mitigate the effects of impermanent loss. As price volatility lessens, the ranges narrow significantly to take advantage of the higher fee multiplier.

This strategy is particularly advantageous to MVI-ETH because we have a decent history of trading data on Uniswap v3 which will allow us to really fine tune the strategy to further mitigate the effects of impermanent loss.

A v2 pool with $5m in liquidity enables 11 ETH trade with 2% price impact. A v3 pool should be much better, especially when liquidity is actively managed. If we are not getting a much better execution, then providing incentives makes less sense.

https://docs.google.com/spreadsheets/d/1wq-W6VBgqq2RwsT_KYU_VA7Fud5ihfiEF9wGm5RjzU0/edit#gid=360826019 Based on our simulation here, an actively managed Uni v3 pool will have roughly the same amount of slippage with roughly half of the liquidity on Uni v3 so long as the price is within range.

In terms of Visor vs. Gelato, we would personally lean towards Gelato. This mostly has to do with the recent oracle exploit at Visor vs. Gelato’s G-UNI tokens being used by Maker as collateral and currently going through the Aave process.

There was an incident on Thanksgiving day in which one of our public position’s was attacked. $600k was exploited due to the extremely abnormal liquidity composition of both the managed position and the overall pair. The issue has been fully resolved. Please see our official statement here: uniswapv3-risk-mitigation/Notes on Uniswap v3 Risk Mitigation.md at main · GammaStrategies/uniswapv3-risk-mitigation · GitHub

The exploit opportunity was largely mitigated due to pre-existing safeguards such as floating deposit caps and total global caps on the overall pool size that we’ve implemented. However, this wasn’t enough in completely safeguarding the vaults.

Our solution has been the following:

  • Enforcing dual-sided deposits at the time of depositing assets. This will have the added benefit of significantly increasing the amount that can be deposited at one time.
  • Utilizing a TWAP oracle as a check against volatility. Essentially, the LP tokens will still be minted at the spot price, but a prerequisite to depositing would be that the variance between spot and TWAP price is a low enough amount. When spot price has been manipulated, the variance between TWAP and spot would be large enough to prohibit any deposits into the vault.

Security is a top concern of ours and will always take precedence over other matters. We had intentionally capped deposits, global pool size, and TVL since the inception of our vaults out of an abundance of caution. These safety measures were extremely significant in mitigating the effects of price manipulation exploits, and we’re moving ahead with the additional safeguards to further protect our vaults and increase the amount that can be deposited into them safely.

Further, we are unsure if there’s a need for active management of liquidity for pairs like MVI, which should be highly correlated with ETH and face limited IL. Do we have any evidence that this would outperform a relatively conservative range position on a net basis (after Visor’s 10% cut)? Perhaps a fairly conservative range on Gelato makes more sense and gives us the ability to do short-term incentive campaigns for more concentrated ranges around a particular market event

As we have seen MVI & ETH, while correlated, can suffer severe impermanent loss during periods of high volatility like we have seen earlier where MVI strongly outperformed ETH. Incurring impermanent loss in this situation has the ability to damper profitability quite a bit if the ranges are not managed effectively. Gelato from my understanding is more of a heuristic strategy which takes fixed ranges and only rebalances when out of range. The Autoregressive strategy would be a lot better in terms of mitigating IL during these price swings due to taking volatility into effect in determining price width of the bands.

I’ll also work through a report which will show our performance in the MVI-ETH vault thus far and share with you after I’m done. Happy to answer any further questions as well!

As promised, here is our performance report for the current MVI-WETH pool that is currently running our Autoregressive strategy: MVI Analysis - Google Docs

image
In the chart above, the Net APR, which takes into consideration impermanent loss is 75.32%. The LP Value vs. HODL is an outperformance of the LP position vs. HODLing MVI & ETH externally. Over the period of 25 days, the LP position outperformed HODLing MVI & ETH by 3.09%, which is pretty good considering the extreme price divergence between MVI and ETH in late November.

image
Here you can see how the strategy adjusts in late November to widen to account for the price divergence.

image
Here, you can see that the price divergence between MVI & ETH was handled pretty well given the circumstances. In the long-run, the LP value should further outperform HODLing the assets.

image
Lastly, you can see the position composition of WETH and MVI in the pool. Although the position becomes slightly more weighted towards WETH due to the price increase in MVI against WETH, this is a pretty good outcome given the circumstances.

3 Likes

I asked Brian to chime in because he is much better at articulating Visor’s value prop than myself, as well as to bring in the actual performance report for the MVI-ETH pool.

To build on that, I think we may broadly be under-appreciating the value of active management right now. Especially in comparison to G-UNI tokens (albeit stable pairs) proposed as collateral assets on Aave / Maker. I think we likely will have both, in time, and that is great. But our MVI and DPI liquidity pools are really poor from an efficiency perspective. Visor vaults will do more to provide deeper liquidity with the same or fewer assets.

Also, I am not convinced that an MVI-ETH G-UNI token is even attractive as a collateral asset. If you are losing in your position relative to holding due to IL because the position is not actively managed, why would you use that as collateral? We have not seen evidence of our products being valued as collateral to date anyways, so I would like to see us optimize for active management and more efficient liquidity than AMM listings.

All this to say, I would still like to see us try and push liquidity to a Visor vault and capitalize on the active management - improve liquidity for our users first and foremost.

cc: @verto0912 @Matthew_Graham

note: my assumption is that the liquidity pod will take over this initiative

4 Likes

Chiming in to add one piece of evidence that highlights the need for active management in v3 LP positions. The DPI / ETH pair is looking absolutely brutal as LPs have not updated their ranges though the market moved against DPI since they created their positions back when we introduced liquidity mining incentives.


source: Uniswap Info

2 Likes

Hi @TheYoungCrews

We have been working on DPI :ETH with the Gelato team for some time now. A similar proposal to this one is to be draft using the Gelato G-Uni pools on top of the Uniswap V3 DPI : ETH pool. Based on discussions to date and expectation to proceed with a G-Uni pool, Gelato has already created the pool for us which can be found here.

The action is currently on myself here to draft up the IIP for the DPI : ETH pair. I hope to get this on the forum soon and I plan to include gas reimbursement, but I am not 100% sure about incentives at this point in time. I do think they are needed, but it will likely only be Index Coop that provides them. The management keys will be retained on the G-Uni pool so the liquidity can be rebalanced.

As @jdcook mentions we do envisage having both Gelato and Visor pools for our products.

4 Likes