"Fee Switch" Design Space & Next Steps

Following up on some of this conversation, I agree with @brenner and @leighton that targeting a specific number or a time bound period for fees to be activated would be an appropriate way to proceed.

I don’t think it’s necessary to specify a use for these funds, but some obvious things come to mind. In terms of adding funds to existing initiatives, options are:

  • Funding next version of Uniswap Grants Program with fees.
  • Adding funds to the bug bounty.

In terms of new initiatives, there is an enormous amount of value to be gained in funding dedicated entities to take on certain aspects of protocol governance work that currently is not happening because none of us are getting paid.

  • Designing a paid delegation program.
  • Creating a special purpose vehicle to act as legal counterparty on dao2dao transactions and contracts, e.g. like those that have happened with Voltz, the Polygon team, Moonbeam, and so on. (We have another report coming out soon that details some of the accountability issues that we’ve seen in the Uniswap ecosystem that covers this in greater detail).
  • Independent team focused on procurement work along the lines of what we have been doing with our crosschain governance initiative. This is especially critical now that Uniswap Labs has increasingly stepped back.
  • Hiring new development orgs to make new contributions to the protocol, like what Scopelift has done with its recent grants from UGP.

Even without deciding a larger “mission” for Uniswap, these opportunities are pretty obvious and I daresay necessary steps for growing Uniswap’s competitiveness.

@leighton I think a viable next step for moving this conversation forward would be writing the code for a parameter change that takes the preceding conversation here into account and posting it here as a concrete artifact for debate and more discussion. Is there someone on your team who would be willing to take that on?

1 Like

Yep! 100% agree. I actually drafted this last night (pretty much along the lines of what you just outlined). I will post it today or tomorrow so we can move to next steps!

  • Lp can either buy UNI for revenue compensation or let UNI deflate.
1 Like

Hello all, excellent discussion so far. Some thoughts as a humble Uniswap tourist.

  1. I don’t quite get why LPs are particularly concerned with this experiment on a limited set of pools. Either (a) the margins on the pools are really low and they leave, (b) margins are high for some LPs and they stay while others leave, or (c) LPs leave the pools that have the fee switch turned on en masse. In any case, LPs can vote with the purse, and I don’t buy that turning the switch on is an irreversible transgression against LPs.
  2. @Game_Theorist, you and I share a name and a passion for algo game theory! Your approach aligns perfectly with a rational-self-interested-agent approach, which is where my dispositions lie as well. But I see the point of this proposal being to test just how efficient liquidity provision is on Uni V3. I love hypothesizing from theory as much as the next guy, but to fully reject this proposal on your grounds, one would need to assert that these markets are efficient. And who the hell would dare rest their name on a claim that crypto markets are efficient?
  3. I think there’s been precious little said about the legal concerns with this. I’m no lawyer, but I at least know there’s stuff I don’t know, and it’d be great to get someone who knows one to start thinking about the intricacies of the fee switch legally. This seems much higher priority than squabbling about LPs’ opportunity costs for a short period of time on a limited number of pools.
  4. In a way, the fact that so much debate has occurred already on the fee switch is perhaps a valuable prior for understanding the probability that turning on the fee switch (for some pools) is “good”. When optimal parameters are not known, it is necessary to search the param space to hone in on optimal parameters. Of course, if this exploration is extremely high-cost, e.g. in the form of securities backlash or bitter LPs, then there’s reason for pause. But if LPs’ capital allocation is nearly as rational as the naysayers claim, then their equilibrium strategy will be to return their liquidity once the experiment concludes, same as before (assuming these LPs don’t have enough power to grim-trigger the pools, which I think is a fair assumption). Under a consequentialist framework, which seems to be the one most agreeable to economics, taking no action is an action itself. And if legal concerns are resolvable, the search for optimal fee switch obviously requires turning some on.

As I said, I am but a tourist, and I’m likely missing context here. But if this experiment is tractable legally, the conversation should be on how it should be conducted, rather than if it should be conducted, due to the apparent lack of clarity that the community currently has regarding the effect of a fee switch.

I’m not saying reject the proposal, I want the fee switch on, but this experiment will exaggerate the effects of turning on the fee switch which may actually hurt progress towards turning on the fee switch. I think we need to think carefully about conducting the expiriment in a way that minimizes spillovers for example we should do the experiment on an entire chain of pools, since bridging is expensive and can be considered enough of a moat to contain most of the spillovers.

I would recommend turning on fees on Arbitrum system wide for all pools and all fee tiers.

Going back to the casino starting to rake pools, imagine you own a chain of casinos. you want to figure out how raking will affect your attendance and the sale of ancillary stuff like hotel rooms and drinks/food. So instead of turning on rake on a few tables in every casino, you turn on the rake in a few casinos. As long as there is sufficient geographic distance, you will have contained the spillovers pretty well.

I think markets are pretty localized here so for example it wouldn’t make any sense to have a market for a bunch of arbitrum tokens exist on polygon or visa versa.

1 Like

Hey everyone, I’m from Fluidity and we have experimented with an alternative solution to this, which uses both the fee switch and our Transfer Reward Function (TRF). Fluidity works by rewarding users for spending their crypto, including swaps on DEXs. For every reward that is paid out, it is split between sender and receiver, which means that the Uniswap treasury would earn yield every time somebody performs a transaction on Uniswap using fluid assets.

We have done some calculations with this, and they have shown that Fluidity would generate approximately the same amount of revenue for the three pools that have been chosen for the test run. Because of this, we propose a hybrid model that uses both fluid assets and the fee switch, where fluid assets would be used on pools that have lower LP fees and where it would otherwise become unprofitable for LPs to provide liquidity. For pools where this is not the case, the fee switch can be activated. With this, Uniswap would prevent liquidity leakage while still generating revenue for the treasury that can be used to incentivise token holders. It also discounts any increase in trading volume due to incentivising utility.

For example, we can assume around 200,000 daily transactions on Uniswap V3. If Fluidity captures only 5% of this volume, we can estimate the mean daily revenue for the treasury to be $1,100, or $400,000 annually. If we compare this to the DAI-ETH-0.05% pool, it would have earned the treasury an average of $136 daily in the last 7 days.

For some reason I can’t include images in this post, but we have written an extensive blogpost on how this would work as well as this thread.

This is first and foremost meant as a thought experiment and not an official proposal, nevertheless we would love to hear your thoughts on this!

Edit: embedded links

Hi everyone! All being equal, a successful fee-switch would mean LPs help finance development costs, right? It would be good to have some a-priori statistics of LP impermanent loss, for different pairs, liquidity concentrations, time-scales up to 120 day max bins. Data should exist that if show LP are mostly in profit, would be good to know.

If I read this right, with the fee switch turned-on: a small fee
percentage on every pool Tx would go to a treasury fund. But what then of these funds? Can they be matched by the long term released UNI team funds to fund development? Although this might put buy pressure to Uni holders, it might be difficult to take for LPs that have encumbered significant impermanent loss on liquidation so though LPs might not be too deeply involved in this discussion, if loses are significant they may still want to pull-out of those LP pools, against the best interest of the protocol so caution seems good advice.

Lastly: What about Pros/Cons of choosing either ETH/WBTC or WETH/WBTC pools in the pilot switch study? (For statistical purposes might be useful, since both are inextricably tied to “bitcoin”, the commodity!)