Making Protocol Fees Operational

Proposal Discussion
22

Hi Uniswap Governance :wave: ,

“1. Do you agree that Uniswap does not need to offer substantial fee rebates to liquidity providers?”

Uniswap has established itself as a trusted brand. For liquidity providers who commit for longer durations and offer a wide tick range, it provides a sense of security. While reducing fees for LPs may lead to some migration, it could also result in increased fees for those who remain.

“2. Do you agree that a one-fifth fee tier is an appropriate starting point for a protocol fee?”

Starting with a one-fifth fee tier seems reasonable. It strikes a balance by not being excessively high while still being meaningful enough to have an impact.

“3. Do you agree that a new system for fee management is a good approach to carrying out the necessary maintenance surrounding the fee switch?”

The proposed approach makes sense as it lays the foundation for capturing fees, which is an essential first step towards determining how those fees could be utilized.

“4. Do you agree that implementing the system on Polygon is a good first step?”

Polygon has demonstrated its reliability and maturity over time, making it a suitable platform for testing the fee switch and observing the loyalty of existing LPs. Additionally, with Polygon’s growing partnerships with traditional corporations, there may be additional opportunities to explore the utilization of collected fees to users.

“5. What token do you think the DAO should trade fee income for to be held in the protocol treasury?”

It would be advisable for the DAO to consider holding fee income in tokens such as USDC, ETH, and UNI. USDC provides stability during volatile periods, while ETH showcases alignment with the development of the Ethereum protocol and the potential for staking. UNI tokens can be utilized for buybacks, creating a solid foundation for the development of the Uniswap ecosystem, as well as supporting the community and builders.

Regarding @guil-lambert’s proposal for the fee switch:

The proposal primarily focuses on the potential use case of fees for liquidity management and falls short in terms of establishing the necessary infrastructure for the protocol treasury.

A fee switch that functions as a pass-through exposes Uniswap to external rent-seeking behavior and may solidify the position of early adopters of specific fee utilization, such as liquidity management protocols or on-chain options companies.

In response to the question of how the fee switch proposal contributes to the sustainable growth of the Uniswap protocol (rather than solely benefiting the $UNI token), it’s important to note that revenue generated through fees enables Uniswap to serve the public good. UNI represents a unit of public good. Without fee income, the fees would solely benefit profit-motivated LPs. The same applies to liquidity management protocols or other protocols competing for a share of the fees, as they are primarily focused on their own token holders and user base.

While I am not a lawyer, passing the responsibility of the fee switch tax to other decentralized protocols would introduce additional complexity. Those protocols would also face uncertainties regarding legality, jurisdiction, and potential conflicts with regulations, especially if they are located outside of the United States. This approach could exclude protocols that are unwilling to navigate such risks, further entrenching the fee switch structure. Regulatory opacity should not hold back the DAO from being a leader in the space.

1 Like
23

First, thanks to @GFXlabs for pushing this.

Since last year, a number of positive steps have been taken in Uniswap DAO in the fee switch initiative, but these steps have mainly concentrated on the discussion and analysis around the topic. Discussion is always healthy, but we have finally come to the point where we need action on this initiative in addition to discussion.

Out of any possible revenue model, it is clear that the fee switch is the simplest and the most meaningful monetisation path, essentially just introducing a take-rate to an exchange business. Practically every single exchange globally has the same primary monetisation model, take a cut of the volume for the house and that’s it.

There are a few main aspects to the fee switch discussion:

  1. Why and how should the fee switch be activated?
  2. How should the captured value be distributed to UNI-token holders?
  3. Risks involved

Why should the fee switch be activated:
Uniswap DAO does not have a revenue model, but has a growing cost structure. It is a separate discussion how the expenses are allocated or which investments are made, the key point is that every single investment Uniswap DAO does is currently at the expense of UNI holders. It is simply ownership dilution. I do not want to discourage great initiatives like the Uniswap Foundation, I’m just pointing out how Uniswap initiatives are currently financed.

Let me attempt to fix the widespread misconception about treasury funds that @GFXlabs also pointed out: Uniswap does not have a treasury. I will re-emphasize this so there is no ambiguity around this topic: Uniswap DAO does not have a treasury by any reasonable definition of a treasury whatsoever. Uniswap DAO has a large genesis allocation of UNI allocated to “community”. Whenever this “treasury” is used, UNI is sold on the market, diluting existing holders. This point is particularly important, because it seems to be widely used as a counterargument against the fee switch: “We already have a treasury, therefore we don’t need a revenue model to generate more treasury”.

The only funds that can even remotely be called a treasury for Uniswap DAO are ARB-tokens and similar sizable airdrops, but everyone understands these are one-time occurrences now in the early phases of the crypto economy. The major point stands, which is that there is no sustainable revenue strategy to generate a treasury, and there is practically no treasury. These are the two key points answering the “Why”-question.

How should the fee switch be activated:
There should be a gradual rollout to ensure that risks are minimised. Further, there have to be some predefined metrics on turning the switch off due to possible market share losses in volume. While I personally do not see Uniswap losing market share as a result of the revenue model, this issue is not binary in the sense that the answer might be a -1%, or -5% decrease in the comparable overall volumes long-term, when we have a system-wide fee switch in place.

How should the captured value be distributed to UNI holders:
The best way to begin is to simply accumulate the revenue into the Uniswap treasury in the form of stablecoins and/or ETH. In practice, this could be 50% USDC, 50% stETH.

Once we have run this strategy for a while, some of the revenue needs to systematically be distributed to UNI-token holders. We are not fans of the burn model in general and hope to see a staking module implemented, in which participation is voluntary (also legally more sound). This would create an organic yield for the UNI token in the form of directly distributed USDC / stETH tokens. In addition to this, there needs to be an option to stake delegated UNI, so that delegation is not disincentivized.

For GFX’s questions, we agree with steps 1-4 and for 5th: 50% USDC, 50% stETH.

Addressing the risks:
The main counterarguments around the fee switch seem to circle around 1) regulation 2) the idea that all the “LPs would fork Uniswap and leave”.

Regulatory risks need to be carefully balanced with the financial reality Uniswap DAO is currently at. In practice this means that risks need to be understood, the best model forward needs to be determined, and then proceed to move forward with that model. We are definitely not regulatory experts with @mhonkasalo, but as non-US people we clearly see how US-centric the regulatory discussion is. We want to see Uniswap DAO evolve into a truly global project. This means creating its own global sustainable financials outside of any specific country.

As for all the LPs leaving Uniswap due to this initiative, we have @WintermuteGovernance, a large Uniswap/DeFi MM, expressing a positive view on this thread. While this is only one data point, due Uniswap’s market position and reputation, we would probably not see a mass exit of LPs. Moreover, Uniswap users have trust in the current battle-tested contracts. The volume loss risks can further be minimised by the gradual rollout of the changes.

Conclusion:
Finally, I’d like to make it clear that our delegate platform represents the interests of UNI-token holders, which includes financial interests. We are not comfortable seeing UNI declining until eternity due to the previously mentioned imbalance in the revenue/cost structure. The key challenge here is that everyone has a different view of what Uniswap is. Especially to @guil-lambert’s points:

The Uniswap infrastructure is a public good. Turning on the fee switch in this manner would lead to the deterioration of the Uniswap protocol’s public good nature (in a very tragedy of the commons-like manner) and would undoubtedly push the $UNI token closer to being a security.

We do not see Uniswap purely as a public good or a non-profit entity. Uniswap has elements of a public good, but it also has elements of a traditional exchange business. We have to be able to critically look at the Uniswap DAO financials and understand that the decision to do nothing means driving the UNI token ultimately to 0, long-term. Turning on the fee switch would lead to long-term prosperity (not deterioration) of the Uniswap DAO and all the related protocol infrastructure.

The Uniswap protocol is currently valued at $5.3b. Who here truly thinks that this value is based on the market’s willingness to finance a public good infrastructure on a charitable basis?

We fully support proceeding with the fee switch implementation with the delegation we have been trusted with and want to see Uniswap DAO and the UNI token prosper long-term. This initiative is ultimately not only about a revenue model. It is about deciding whether Uniswap is inherently a global project or a US-project.

6 Likes
24

Had a thought, will there need to be a new proposal for new pools that fall under the above criteria after this proposal is deployed, or would any new pool that meets the above criteria automatically have 1/5 of its pool fees diverted?

1 Like
25

Excellent proposal and discussion so far.

But the DAO might just not have enough data to make a good decision at the moment, even leaving the legal issues and conceptual differences aside. How about running some kind of randomized controlled trials (RCT)?

By just enabling the fee on all active pools within an ecosystem the DAO is not going to get a causal relationship between the action (fee switch) and the results (reaction of LPs) because there might other factors (market sentiment, appearance of new v3 clones) that also make the LPs to leave Uniswap. Inter-ecosystem comparison can be done, but the number of ecosystems is probably too small to get meaningful results.

Here’s a sketch of a potentially better study. The DAO could randomly select some pools, but not others for the fee switch. Let’s say there are actively traded coins A, B, C, D. A coinflip selects B and D. The next step is that all ETH and USDC pairs for B and D have the fee switch enabled; then the behavior of B and D LPs can be observed and statistically compared with A and C LP behavior. With a sufficient number of samples, and the assumption of rational LPs, the results should be pretty informative both about the size of the effect and the likelihood that the LPs reacted specifically to the fee switch, rather than something else in the market.

The experimental design needs to be considered in detail, and involve some statisticians, but in general running a RCT is the gold standard of establishing causality.

3 Likes
26

Unfortunately, each new pool needs to have a fee set. Fortunately, our design can separate the process of setting a fee such that this isn’t a managerial pain.

2 Likes
27

I thought that someday Uniswap would add an option to pay fees with UNI and receive a discount on swap.

1 Like
28

GFX has posted three polls. We invite all voters to express their opinions by voting on these three polls. Once the polls are closed, we’ll use the polling data to propose a formal Temperature Check.

Polls

Poll 1: Fee options for V3
Poll 2: Initial deployment chain
Poll 3: Asset to be held in the treasury

5 Likes
29

Thanks to @GFXlabs for the detailed post regarding the Uniswap Fee Switch. We share the excitement for expanding the design space for protocol-based revenue.

As many know, we have consistently advocated for ensuring the appropriate legal and technical structures are in place for the Uniswap protocol prior to turning on fees at the protocol level. Additionally, it’s important to consider the fee switch within Uniswap’s broader growth strategy. From a business perspective, what are the risks and rewards associated with the fee switch and LP behavior? Could turning on the fee switch now present an opportunity to a competitor DEX? Has the community designed a creative and elegant way to incorporate the UNI token into the economic structure of the network?

We believe patience is warranted due to these business, technical, and legal variables: (1) the overall risk-reward regarding Uniswap’s long-term business strategy, (2) the lack of programmable technical architecture for the flow of funds stemming from activating the fee switch (and the UNI token’s role in accessing such fees for performing work on behalf of the project), and/or (3) the need for a proper legal entity in the event fees pool at the protocol level.

Ultimately, if any action relating to the fee switch could result in a tax obligation, then there needs to be an ability to pay such obligation. Simply reserving funds is not sufficient. A legal entity could accomplish this and therefore greatly mitigate the risk associated with potential tax obligations for the Uniswap DAO. However, we recognize the legal and regulatory complexities associated with implementing an appropriate legal structure for the Uniswap DAO.

In the absence of a legal entity, it is important to reduce tax risk by using a programmatic flow of funds directly to token holders who are performing work on behalf of the DAO (as the Uniswap Foundation previously recommended here). A programmatic flow of funds could help ensure the taxable obligation rests with those users instead of the DAO.

Importantly though, the Uniswap v2 and v3 smart contracts are not explicitly designed for a programmatic flow of funds. They require governance participation to claim fees, whether from the FeeTo contract in v2 or a governance vote calling the collectProtocol function in v3.

The @GFXlabs’ proposal tries to address this limitation through their Fee Management solution “that allows anyone to pay the gas to collect the fees, sell the fee tokens automatically via the Uniswap auto-router, and then send the proceeds to the protocol treasury.” While we appreciate the innovative decentralized fee routing, it does not address the underlying issues regarding tax considerations for the DAO. In fact, it potentially exacerbates them.

@GFXlabs states they, “have separated the protocol fee discussion into two distinct conversations. One is the technical implementation of accruing fees in pools, collecting the accrued fees, and selling the accrued assets to a protocol-designated asset to be held in the treasury, and a separate conversation of how the assets in the treasury could be distributed, spent, or otherwise utilized.” This analysis fails to consider or appreciate the tax implications for such actions and is likely no better than just simply turning on the fee switch. Ultimately, any collection of fees at the protocol level without a new mechanism for the flow of funds could potentially result in a taxable event, one which the Uniswap DAO is not currently equipped to address. If you’re building a house, you don’t want to turn on the water until the plumbing is completely installed and checked.

The Uniswap Foundation alluded to the same line of logic in a follow-up post to the most recent fee switch discussion:

“Much of our research has led us to the conclusion that implementing a programmatic distribution of fees directly to persons, for instance to those who have “opted in” through taking some beneficial action for the Protocol, may be a superior path forward. This kind of mechanism would represent a significant mitigation of risk for the Uniswap DAO because tax obligations would rightfully be the responsibility of the individual in their respective jurisdiction, among other potential benefits. For instance, properly designed, this mechanism could further incentivize productive behavior to support the Protocol.

There have already been many comments in the forum exploring different uses of fees, including but not limited to: gas fee rebates, incentivizing LPs to lock their liquidity, and distribution directly to UNI token holders who actively “opt-in” through taking some beneficial action. Programmatic distribution of fees directly to stakeholders would relieve some, but not all, of the uncertainties associated with fees directly routing to the DAO.”

We understand the desire to turn on the Uniswap fee switch, but we think it’s necessary to take into consideration the nuanced factors in how it’s done. We are actively researching different methods and technical implementations that create a programmatic flow of funds to users and UNI holders who opt-in. We are preparing to provide an update next month on possible solutions. In the interim, while we appreciate GFX’s efforts here, we do not support this proposal and will vote against it based on the reasons above.

Porter, Miles, and Eddy - a16z

3 Likes
30

It is an important distinction between a proposal with an implementer/builder to enable the fee switch vs a future proposal that seeks to provide guidance. Would a16z’s future proposal have someone who can implement its fee switch solution?

One of my worries is that the overhang of legal uncertainty is being used to kick the ball down the road. Governance is deflated when there is inaction. This can be seen in corporate governance structures all the way up to large/small country governments.

Looking back at the last major thread on the fee switch here many ideas are put out as uses. The great thing about the fee switch is it opens possibilities.

The idea of an opt in, and having the burden of tax go to the user’s is a good idea. I do not see how this proposal prevents that from happening. Rather it is a first step to creating the infrastructure of contracts to do it in the future. Perhaps the independent parties who are pursuing the fee switch implementation/credit should collaborate.

Is the a16z’s proposal that is to be released next month/June a continuation of the unreleased proposal that was stated to be made available in “late January or early Febuary”?

1 Like
32

1. Do you agree Uniswap does not need to offer substantial fee rebates to liquidity providers?

We believe LPs need to be compensated with fee rebates, it’s not an uncommon practice in centralised exchanges as well. Market making entities and trading firms that are able to reach high volumes usually maintain negative maker fees.
At the same time, we believe that there is space for fees to be distributed to the DAO considering the importance of revenue generation.

2. Do you agree a one-fifth fee tier is the appropriate starting point for a protocol fee?

We believe that a 1/5 fee tier could be aggressive for LPs . With that said, we acknowledge that it provides be a good balance between DAO treasury & LP rebates, so we encourage the community to consider a conservative or progressive approach where fees start from 1/10 and run test periods for each tier until 1/5 to observe differences and potential behavioural changes from LPs.
1/5 fee switch could potentially squeeze less “sophisticated LPs” but overall we are not worried about diminishing liquidity. It is of course something that needs to be monitored by the DAO.

4. Do you agree implementing the system on Polygon is a good first step?

Yes the risks there are considerably low and we believe it is a good first step.

5. What token do you think the DAO should trade fee income for to be held in the protocol treasury?

We believe a split between stablecoins and ETH should be reasonable. We would lean towards a 60/70% stables and the rest ETH approach. Markets can be volatile and so treasury management can break or make protocols.

33

We greatly appreciate the effort and research put into the proposal and the discussions surrounding it. After thorough consideration of the research conducted on fee enablement, we support GFX’s preliminary designs for the fee switch options.

However, we acknowledge that these discussions are still in the early stages, with Snapshot polls and forum commentary around key questions. To fully endorse this change through to execution, we would like to see a finalized technical specification that includes a detailed plan for implementation, auditing, and other necessary aspects.

Furthermore, we have sufficient concerns regarding the tax and legal implications, which necessitate a more in-depth examination before we can provide complete support. While we understand that addressing these issues may not be the proposer’s responsibility, the regulatory gray area motivates our opinion that it is crucial to have a system with optional mechanisms in place to handle any potential implications. Uniswap is a vital protocol within the ecosystem and has high visibility as a dapp.

We recognize the substantial work already invested in this endeavor, and we understand that some believe these matters can be resolved at a later stage. However, we strongly feel that they are important considerations to have properly analyzed and addressed before the fee switch is activated considering the potential significant impacts and implications of the action.

4 Likes
34

FRAGMENTED VOTE

It seems like the activation fee switch was in majority, but the vote was fragmented.

A better options for the vote could be:

  • No, do not activate
  • Yes AND activate at 1/10
  • Yes AND activate at 1/5

In case “Yes” options wins, the more popular among 1/10 and 1/5 will be chosen.

Alternatively, to allow for all edge cases, the following voting options:

  • No, do not activate
  • Yes, no preference over which fee settings
  • Yes AND prefer 1/10 over 1/5 (both options correct)
  • Yes AND prefer 1/5 over 1/10 (both options correct)
  • Yes BUT only 1/10 (if 1/5 wins then NO)
  • Yes BUT only 1/5 (if 1/10 wins then NO)

I’m genuinely unsure. Appears overcomplicated, not intuitive. I presented these options for completeness. Are there any other edge cases? :octopus:

EDIT / UPDATE:

Once the polls are closed, we’ll use the polling data to propose a formal Temperature Check.

I see… These were polls. Before temperature check. Still more way to go…

2 Likes
35

fee of 1/5 is too high. That’s a 20% tax on LPs.

36

Any news about fee switch ?

37

image
image625×568 77.7 KB

6 Likes