A few thoughts on the risks around Uniswap v3 LP and staking

I thought I would share a few thoughts on Uniswap V3 and how the coop can use staking contracts to maintain our products’ liquidity. Since we did our initial research on v3 a month ago there has been some progress, within DeFi, but there are still plenty of unknowns and challenges remain, particularly for pairs that are not correlated (i.e. all coop products).

This follows on the work done with @verto0912 @Matthew_Graham @jdcook @catjam and @Pepperoni_Joe and discussions with @dylan on the staking contracts that @ncitron has been doing.

Risks to LP’s

The key advantage of v3, is that it allows liquidity to be concentrated, and so capture more fees for a given $ size position.

However, by focusing the liquidity the LP is exposed to a number of disadvantages:

  • If the price ratio moves out of range there is no fee income.
  • Divergence loss is also concentrated compared to v2.
  • It takes time, expertise, attention and gas to reposition.

While most of our product pairs are ones that people may be happy to hold either token long term, the impact of having a position converted to 100% of one token and then trading back is not pleasant. [ I’ve experienced > 10% losses compared to a simple buy and hold for MVI:ETH within 3 weeks. ]

Unlike v2 or sushiswap LP’s each position is unique and needs attention and while it’s possible to conduct back tests on historical data, that is no guarantee that even wide limits will stay in range for an extended period.

I think the following best sums up v3 liquidity provision:

1 “It’s a whales game”

2 “unlike v2, No-one will get average returns”

#1 is due to the more active nature of being a LP, and the gas costs associated with positioning, claiming fees, withdrawing and repositioning. Positions less than say $50 k are at a disadvantage.

#2 is the very nature of the new protocol:

  • If you go wide you will get less than average fees.
  • If you go narrow, you risk being out of range
  • Even if you have a great strategy, a more active LP can reduce your income by being better.

There is still research being done to improve modelling and backtesting of different strategies:

https://uniswapv3.flipsidecrypto.com/

Staking in v3.

I’ve not yet seen a good explanation of the process required to stake a v3 position. My understanding is that the flow is slightly different:

  1. Position liquidity
  2. Register NFT in staking contract (or contracts)
  3. Claim fees / withdraw liquidity / claim mining rewards maybe three different transactions.

Risks to the coop

While the staking contracts have been prepared and audited, there are still risks with deploying a contract that has not been battle tested by others. To date, no major liquidity mining campaigns have been run using v3. So, while there could be kudos in being the first, there is additional work to educate potential LP’s.

While the technical risks are likely understood, there is an additional reputational risk to the coop. If we are the first to run such a campaign, and there is significant price movement, a number of our LP’s could suffer unanticipated losses.

This leads me to the question:

“Do we want to encourage LP’s to get micro rekt?”

In addition, if the coop were to be the first time people have LP’s on v3, then they could come to associate the coop (and not v3) with their losses.

If LP’s move to v3, have problems and then leave the market, the depth of our liquidity at market price will suffer.

How we can mitigate those risks?

At the moment, most of the attention has been focused on the DPI:ETH pool as this is the largest cost for the coop, and there is already a significant v3 pool in operation. It also has the advantage of having many more options for mitigation:

  • Delay liquidity mining on v3 until others have run large campaigns and many of our potential LP’s have learnt what is required.
  • Increase liquidity mining rewards on v3 to offset any particular losses.
  • Educate LP’s of the risks of moving to v3 and the different strategies (wide passive / narrow and active)
  • Highlight the availability of other liquidity options.
    • Sushiswap
    • Balancer v2 (ETH:wBTC:DPI)
    • Other incentivised pools TBC.
    • Stay in uniswap v2 and see what residual fees income is available. (may also have tax advantages)
  • Highlight other options to make assets productive:
    • DPI as collateral on Cream
    • DPI as collateral on Alpha (unlikely to stop when Liquidity mining stops)
    • ETH staking / collateral / lending
  • Encourage automated strategies for coop products - once the technology has been proven.

There is of course the option to looking to migrate the MVI liquidity before DPI as it is a smaller pool.

Timing

At the moment we have ~ 18 days before the end of the DPI and MVI liquidity mining contracts on the 13th of August 2021. So we have a couple of weeks to discuss the options, agree on a plan, and run IIP’s to set up any continuation of LM. We will also need to communicate our plans with the community and LP’s.

I’m looking at the following timing:

  • Share the idea for delegation of Liquidity management to a small team.
  • Post blog on my experience as a v3 LP.
  • Share some thoughts on v3 LP’ing [This doc]
  • Review the current liquidity for our products as a continuation of these posts 1 and 2
  • Have a community call on Thursday, July 29 · 17:00 UTC
  • Agree a plan and draft IIP’s for v2 LM / v3 LM / Delegation - [by 30th for vote w/c 2nd].
  • Communicate the plan with the community
  • Manage the liquidity after the end of the current cycle.

So, what are your thoughts on v3 LP’ing, staking and the coops strategy going forward? :thinking:

4 Likes

Thanks for the great write-up @overanalyser

From a business perspective, I think this is absolutely correct. If DPI (and other Index products) are intended to be held by beginner-intermediate crypto enthusiasts, Uniswap V3 liquidity providing is definitely not the place the DAO wants to send them IMO. This reflects a very important decision IndexCoop has to take everyday which is: Should efforts be focused on beginner or experienced users?

1 Like

I would separate this out as LP’ing is fundamentally different from basic holding.

LP’ing is complex and should only be done by those who understand it. Should we give sufficient warning and education I have no issue with that.

4 Likes

I have been saying this since the beginning of the v3 dicussion, and I think it is important to echo again here given that the bulk of this post is regarding LP’s…

The LP’s on v3 will likely be very different to the LP’s on v2 - and that is perfectly fine. I think we are spending too much time worry about how a move to v3 effects LPs. LPs are service providers - they provide liquidity in exchange for fees. DeFi has taught us that where there are market dynamics to capitalize on, smart market actors will take advantage. That is the beauty of v3, imo. It creates a much more efficient market that better serves the Coop.

So, I don’t think we should be worrying about the risks to LPs (aside from some basic warnings as mentioned). The risk we need to consider is whether or not we will get new LPs on v3 once we remove incentives on v2. And, imo, the history of v3 so far shows that if more volume gets pushed through v3, the LPs will arrive… an “if you build it, they will come” type effect - this is where we could spend more time/resources analyzing.

I would love to start focusing less on individual LP-ing, and more on the market dynamics that serve the Coop when it comes to v3. I understand that a lot of us here are LPs so this move directly impacts us, but that should not become a core component of the Coop’s decision to move liquidity to a market that is much more efficient.

9 Likes