A completely new $UNI token model for V4

[mjc716]

@eek637 from the UF asked me to start a discussion on protocol fees and V4 after this post. The conversation the last few years has revolved on 1) whether the DAO should activate the fee switch, 2) which pools should charge fees, and 3) what share of LP fees the protocol should take.

With the V4 code now public, I think this conversation is no longer relevant. The fee switch token model could always be described as rent-seeking – if activated, token holders would add no incremental value to the protocol but would now be extracting value from other participants. In V4, the token model is also unenforceable, making the whole discussion moot.

I wrote more about this here (tweets 38-48): https://twitter.com/mjayceee/status/1673394949188976642?s=20

The protocol has gotten better and better, while the token has gotten worse and worse. This is before even considering that the lawyers at Uniswap and its investors are very unlikely to allow the fee switch to be activated, at least until there is more regulatory clarity. We could easily be on Uniswap V7 at that point.

I’m proposing a completely new UNI token model below. It requires no changes to the core protocol. It also may be more defensible from a regulatory perspective. I think it makes a lot of sense, but it also may not work. Still, I think it’s worth a shot while the token is still worth something.

Where can $UNI add value?

While Uniswap currently dominates on blue chip pairs, the true magic of the XYK model is the ease with which it creates markets for long-tail assets. The expressivity of V4 is a great step towards defending Uniswap’s blue chip dominance onchain, but it’s an open question whether algorithmic models like XYK can beat out conventional order books over the long term for markets with deep liquidity.

Without neglecting the blue chips, a new UNI token model should lean into Uniswap’s importance to the long tail of liquidity pools. Right now, most long-tail assets pair with ETH. This has been the logical standard for a long time, for a variety of reasons.

The problem with this is that ETH has an increasingly high opportunity cost. With staking, it’s about 4%. With re-staking on the horizon, it may go higher. ETH is too precious at this point to be the go-to asset pair for shitcoins.

There is an opening here for UNI – a token with essentially zero opportunity cost – to supplant ETH as the asset pair of choice. UNI can serve as the liquidity connector and hub, linking an extensive portfolio of long-tail assets.

If this sounds like the Bancor V1 token model, that’s because it basically is. But UNI has advantages that BNT never did. First, Uniswap is the preeminent DEX in a far more mature ecosystem, doing billions in volume. Second, UNI has a scale and market cap that BNT never had. Third, projects would be using UNI as a pair asset because it benefits them, not because it’s imposed by the protocol.

UNI also brings the advantage of being more correlated to long tail assets than ETH is, meaning UNI pairs are less susceptible to impermanent loss.

Bringing this section together, UNI is preferable to ETH in terms of both opportunity cost (staking yield) and capital risk (impermanent loss exposure).

How would this work?

The transition from V2/3 to V4 may take a long time, but the clean slate of V4 is the perfect opportunity to enact an initiative like this. I’m proposing a two part plan to transform UNI from a mostly useless governance token into a focal part of the protocol it represents. The two parts
complement each other, with the goal of bootstrapping a new sustainable token model where UNI is the primary asset pair on Uniswap.

Part 1: The Uniswap Credit Facility
Before UNI can supplant ETH, it has to be much more accessible to liquidity providers. UNI currently has a limited presence on lending platforms. There isn’t much demand to borrow UNI right now, so it’s not a big deal. But for UNI to make a play as the base pair, it needs to be extremely cheap to borrow at scale.

Increasing UNI available on an external lending platform like Aave would be a step in the right direction. However, V4 hooks provide the foundation to build a much cleaner and more efficient solution.

The UNI Credit Facility (UCF) is a hook that allows LPs to borrow UNI for liquidity positions. It’s more capital efficient than Aave in all cases, ranging from 1x to infinitely more. It requires no capital up-front, as borrowers only need to collateralize the deficit in their position if the pair token loses value against UNI.

Basic points:

• Hook provides UNI required for a liquidity position as needed
• UCF only requires LP to collateralize the deficit relative to the initial borrow, though in practice there will need to be a small up-front min collateral deposit
• Full initial UNI borrow is restored to UCF when liquidity position is closed

There’s a scenario analysis with all the math at the bottom of this post. XYK math is tricky and some of these outputs won’t seem intuitive on reading them. There’s also a chance I got it wrong, but the larger points hold.

Basic example:

• Potential LP has $100 of SCOIN and needs $100 of UNI
• Via credit facility, LP borrows $100 of UNI to be immediately deposited with the SCOIN
• LP provides $20 in collateral as a buffer, as deficit needs to be backed by 120% LTV, even though there is no deficit at deposit time (XYK dictates that a token deficit only occurs if pair token loses value against UNI)
• SCOIN price drops by 25% relative to UNI (while UNI remains stable in $ terms), decreasing value of UNI side of position from $100 to $87
• LP now owes a debt of $13 but is still safe from liquidation with $20 of collateral (154% collat. ratio)
• LP pays down $13 of debt to close liquidity position and retrieve collateral

Uniswap Governance should seed the UCF with 100 million UNI and set the borrow rate for UNI to 0% for the foreseeable future.

Even at 0%, there are two ways that value accrues to token holders in this design. First, easy credit promotes UNI’s role in liquidity markets, increasing demand for the token. Second, the UCF should be able to take a configurable share of the UNI side of swap fees generated by the hook’s positions, thereby removing tokens from circulation over time.

In terms of dependencies, all that is required is a UNI oracle and an oracle for any collaterals. These can be configured safely within the Uniswap ecosystem. For the sake of simplicity, it might make sense to start with only ETH as collateral, even if it forgoes some of the reclaimed opportunity cost (staking yield).

The UCF might end up more like a B2B tool, rather than a retail lending platform. I won’t get too deep into the technicals, but it may be difficult to ensure the solvency of many positions in one hook, meaning it’ll probably need to be one position per hook/pool. That’s fine - the UCF as a capital efficient, non-dilutive alternative to liquidity mining could be a boon to the protocol and token.

Part 2: A lot of incentives
This part is more straightforward. Armed with this new credit infrastructure, Uniswap should run small incentives/co-incentives programs with hundreds and hundreds of projects. Finding and allocating to this many worthy and unworthy projects is hard and it might not be perfectly fair, but an ambitious bootstrap period would go a long way towards getting this new model off the ground.

As with all token incentives, there will definitely be some dumping. But the treasury is worth $2b and there are currently no reasonable ways to use nearly that much money. If this works – and UNI effectively becomes an index for the long tail of tokens – the treasury could one day be worth much much more. It’s a risky move for a conservative community, but there is a clear, attainable goal.

Presumably V4 (and the UCF) will be deployed on multiple chains, so these incentives would have a wide canvas to make an impact. Whoever ends up allocating these incentives - whether a DAO committee, the Foundation or the core team – would need to be accountable to the community while maintaining some latitude to do a difficult job.

It’s not going to be easy, but Uniswap may be the only project in the space with the gravity to dislodge ETH as the base pair.

Closing Thoughts

This is just the first version of this proposal. There are probably holes in the design I haven’t considered and potential additions that could improve the model. I have a bunch of other thoughts on the financial implications and derivatives of this model, but I think it’s better to open it up to community feedback for now.

One final point: implementing this token model would require a high level of support and coordination from/between many stakeholders. While the UCF would need to be approved by governance, UNI would still have no privileged role within the Uniswap protocol. This might be the first attempt at a “social” token model, meaning the token derives value from coordination, not protocol-enforced tokenomics. While UNI makes sense thematically as the new base asset for Uniswap LPs, PEPE would work almost as well.

Looking forward to hearing questions or feedback from the community. If there is enough interest and engagement, I’ll prepare a more formal proposal in the coming weeks.

Math

[mjc716]

To those asking, “couldn’t someone take a loan from the UCF and then dump within an atomic transaction before the liquidation mechanism could work?” –

This attack vector can be easily protected against with a pre/post swap hook. If a given swap will take the position beyond the collateralization threshold, the transaction will just revert. The liquidity would essentially be locked in one direction in this case.

[Jerome_Coinchange]

Very impressive innovative thinking using hooks when it just release less than a month ago.

Here are my first few thoughts on your post. Hope the reasonning process help in maturing the idea.

Part1
Although it seems much more efficient than using external lending platform (AAVE) the added complexity of managing such position could deter user. Here we agree that such feature would likely be targeted towards protocols and institutions (traders, MMs, yield providers). I don’t really get if this “borrowed UNI” is free to be used somewhere else (trade, deposit in lending pool) or if its sole purpose is to act as the counterparty of the collateral the user provided => making UNI the de facto asset to pair shitcoin with?

If the latter, then we have an interesting situation that greatly favor UNI when paired with long tail asset. By becoming exit liquidity, degen trader will need to buy UNI to exit their trade, pumping UNI price up (if trader do not sell it for ETH or USDC; more on that later). Furthermore UNI will benefit from perhaps the greatest arbitrage network that exist in DeFi, although not highly liquid, a flywheel could be created=> arbitrage increase inventory and trade, pushing volume, attracting TVL in, creating more arbitrage, etc…

There is also the question of management of such facility. The issue of insolvency of the UCF. How “liquidation” work in this UCF model?
Also UniswapDAO would likely need to deploy a set of pools with such hook model in place and develop smart contract logic to enable the accounting of the lent UNI. Then how can it prevent every Uniswap user to interact with such pool? whitelisting would not work as against Uniswap ethos. Hardcoded threshold for participation like, minimum collateral amount = 100ETH?

Part2
As you state, this part might be difficult and require some deep philosophical changes to the way the DAO manage the treasury. There is some work that Gauntlet has done to better understand how incentivization can work compared to the past but it has yet to be tested in practice against mercenary capital. I strongly believe that sound incentive via game theory make the best model. The capital efficiency benefit that the UCF brings should be enough. If more benefit is needed then the DAO can use the airdrop it receives to incentivize this product.

Final remark/thoughts
Overall great ideas which should be nurtured as V4 might change how DeFi work, so does UNI token model need to change with it.
Though some concerns remain in the two part, mainly:

• UNI selling pressure for ETH of USDC after being used as exit liquidity for SCOIN
• accrual of fee, reintegrating the value for UNI holders somehow

Here i believe that the set of pools that the DAO need to deploy need to work autonomously while being governance minimized.

Going further
Here we go further with the idea: the DAO should deploy an AMO fork to manage its UNI pairs alongside the UCF. This could also serve as mitigation to liquidation on top of a proper liquidation mechanism. R=The AMO could be created to take on arbitrage on major UNI pairs, leaving the other pairs to standalone arbitragers. Looking at FRAX’s resilience and deep liquidity is testament to the success of the model.

This idea probably needs a dedicated gov post but i think it would fit nicely with the UCF model you propose. I’ve been thinking about an AMO model lately to solve UNI treasury issue.

The benefit of such model for UNI liquidity is self explanatory and well known, we just gotta look at how Frax did it.

[mjc716]

Thanks Jerome, appreciate the thoughtful note.

A few responses:

• the borrowed UNI isn’t free to use anywhere but the liquidity pool
• we wouldn’t need to reinvent the wheel on liquidation. it would work similarly to other lending platforms
• LPs can only borrow UNI against prime collateral like ETH and maybe some LSDs
• this idea would require a big technical lift for sure; building the smart contracts would take time
• I’m not sure I understand the point around whitelists
• minimum collateral would probably need to be something like 10% of UNI borrow value
• I don’t think there’s any reason the UCF would result in net selling pressure for UNI

[Jerome_Coinchange]

Thanks for the clarification mjc716.

Response to your questions and comment:

1- so here this means that Alice has 10ETH and 10,000 SCOIN worth 5ETH. She wants to LP those SCOIN. she deposit 10ETH as collateral in the UCF pool and borrow 1,800 UNI worth 5 ETH. She can atomically depsosit the 10,000 SCOIN + 1,800UNI in the pool SCOIN/UNI. she can’t sell the borrowed UNI, only deposit it as LP with another token.
Is this correct?

5- whitelist or any “barrier” would be to prevent “non-eligible” users to use the UCF. As you state correctly, the pools might have a hard time to manage accounting of thousands of little positions. Hence requiring some “barrier” like minimum collateral amount to be 50 or 100 ETH, to make sure only sizeable wallet with aligned interest would borrow UNI and LP with it.

7- I guess you are right. The UCF itself does not lead to UNI selling pressure even indirectly. Please correct me if wrong in the reasonning: UCF goal is to make UNI the exit liquidity of SCOINs. Like in my example above, degen holder of SCOIN has made some unrealized gains. He sell SCOIN and buy UNI from the pool, leading to UNI price going up. Since he does not care about UNI as it can’t be used for much in DeFi he decide to sell it for ETH or USDC/t. Theoretically, price movement should be zero for UNI since it would net each other out through arbitrage if any discrepancy arise.

What do you think about the AMO fork which the DAO would own, building POL and get the LP fees which it could do anything with it. The AMO contract would need to be reworked heavily to accomodate for UniV3 position management. Thats something that Gamma strategies, XtokenTerminal or Coinchange could help with as its the core of their business models. This would litteraly propel UNI token as worthy of investement provided the LP fees accrued by the AMO are used to benefit UNI holders in some ways.

[mjc716]

1 - Yes, correct
5 - No, no reason to restrict access with a whitelist. But there would likely need to be a contract-level restriction of 1 position per pool/hook. Deploying new pools/hooks is a lot cheaper and easier in v4, so this is not a major obstacle
7 - I’m a little rusty on Frax-style AMOs, but in general I think we want to minimize governance here. When you start incorporating POL and liquidity strategies, there are a lot more moving parts to manage. The UCF could conceivably be deployed with zero governance requirements. AMOs are gov-heavy

[kaibakker]

On 1) Why does Alice need 10 ETH? This does not solve the problem around ETHs increasingly higher opportunity cost. Could we solve this without depth?

I would propose a simple 50/50 spilt on UNI pairs were the faucet deposits the needed UNI tokens. Below I wrote down a very dirty example hook:

via beforeModifyPosition hook

• On Creation faucet deposits equal in UNI. This is used to create an mirrors position on the current price.
• On Removal the the assets in the position is 50/50 split between owner and faucet. With check to prevent abuse of the faucet (short time liquitity could be blocked)

Now risks and rewards are shared between UNI LPs and UNI token holders. In this case you probably would want whitelisting on high volume tokens.

Might miss something obvious, we should make sure we learn from past liquidity programs on bancor and other tokens.

[Figu3]

Hey Michael, thank you for the vivifying proposal.
I think it’s fair to say that the fee switch is a failed narrative with very little additional network effects, Uniswap can and should strive to do better with its tokenomics.

v4 is an extremely good point to kickstart new discussions as it is a beautiful piece of immutable and permisionless software. It’s good to set an ambitious goal such as making UNI the default paired token in crypto. There is definitely an opportunity, as ETH cannot be the gas token, the internet bond and the default pair token in a durable fashion.

However I do think there are several economic and social risks in the suggested approach. So here’s my 2 cents:

1. There are litterally dozens of lending protocols, building one, natively in a hook, is a monumental task that brings in low value vs using an existing codebase / brand, so why not explore what the market has to offer? I’m pretty sure you can find a protocol willing to share their infrastrcuture for little to no fees to Uniswap.

2.1. I like quite a lot the UCF initiative but it carries several risk, the first one being what was already explained when Bancor tried it (ie: UNI price becomes dependant on the price of its largest counter LPs). This means you absolutely need to reach critical mass fast in order to avoid a death spiral.
2.2 Additionally, you’re moving all the shitcoin risk onto UNI.
For example: an ill-intentioned abuser can borrow UNI, pair it with his token, dump until he pushes the LP to liquidation, and then, you’re stuck with the shitcoin. The only way to solve this is making the system permisionned, which should be a no-go.

3. It’s almost bizarre to design such system and think the Uniswap Foundation should seed the UCF. There are 750,000,000 circulating UNI begging for utility, just create a vault where UNI holders can deposit their UNI to be lent out, in exchange they can get a combination of lending fees + swap fees.

4. You make a really strong point with the fact that UNI is likely to be the only protocol to have enough fire power in terms of incentives to actually succeed in becoming a default trading asset. I’m however a bit taken aback by the thought of seeing a manual process to direct these incentives, why not build an automated and permisionless system? If the rules are set correctly, abuses can be heavily minimized. As well as legal responsibility and centralization risks.

I hope this will contribute to the success of the project,

Cheers,

Figue

[mjc716]

Thanks for this Figue. My work has never been called vivifying before.

Might be helpful to clarify a few points here:

• I think we could draw liberally from other lending protocols, but UCF would need to be a custom implementation. Hooks are really interesting here because they enable a sort of semi-undercollateralized lending in a way that doesn’t exist anywhere else in defi (to my knowledge), so the code would need to be somewhat original
• I’m not sure where the risk of death spiral you’re suggesting comes from. I mentioned the BNT comparison because there are certainly similarities to BNT V1, but not really to BNT V2, which had some structural issues that could’ve led to a death spiral.
• The UCF design proposed does not allow for the abuse you mentioned. LPs can only collateralize their positions with ETH. The shitcoin side doesn’t play into the lending side directly. Check out the math at the bottom of the post for more info here. But at a high level, pool liquidity will be locked in one direction if a pool’s collateral reaches the 120% threshold
• I agree 100% that the UCF should be open to anyone that wants to deposit UNI for borrowing. Good call on that. I also think that the UCF should be bootstrapped with 100m UNI from Governance (not the Foundation) initially to ensure the scale required for this initiative
• Would love to hear any suggestions on how to automate incentives across hundreds of projects in a non-manipulatable way! I definitely agree that automation/permissionlessness is preferable to a manual process. I just didn’t have a good answer for how to approach it that way.

[LvisWang]

This is awesome!

The proposal can truly add value to the tokens and further ads in governonce!

If you need any data support or data tracking when doing an experiment, please let me know! Hit me up on tg @LvisWang I’m in the core team of SixdegreeLab and we would love to help!

Some background information around Sixdegree

• Founded in summer 2022, Sixdegree is a team composed of data scientists and crypto researchers with experience working with 20+ protocols such as Lens, ParaSwap, Gnosis Safe, Y-combinator among others as well as publishing 60+ Dune dashboards and being ranked top 4 team on Dune leaderboard
• Sixdegree services have expanded from one-off Dune dashboard to full data-driven growth services including Dune dashboard, community talks, financial reporting, airdrop strategy and DeFi liquidity proposal among others

[Figu3]

Gauges is probably the way to go

[MattOnChain]

The notion that the fee switch is dead is incorrect.

Poolkey is the result of the following variables (token0, token1, fee tier, tick spacing, and hook).

The DAO contract that controls the fee switch can read all poolkeys and return a protocol fee for specific cases.

EG the DAO can globally make the fee switch apply to all exisitng and future pools. Or to pools that fulfill certain requirements eg token0 = x and token1 = 1. For instance all ETH/USDC pools. Or whatever else the DAO may choose.

The fee switch is alive and well my friends.

[mjc716]

Thanks Matt. I’m glad to be corrected on the DAO having some ability to enforce the fee switch. That enforcement part of the code does not seem to be open source yet, so we’ll have to get more information before settling the discussion.

Even in the model you put forward, there are pretty simple ways to circumvent the fee like wrapper contracts, migrating liquidity to a different base token, etc. This is true in V2 and V3 as well, though I don’t doubt that the additional complexity give the DAO some synthetic earning power.

This still doesn’t address the core point of the proposal, which is that the fee switch is extractive and a tax on the protocol. The goal should be to have a token that adds value, not subtracts from it.

EDIT: the fee logic controller contract will be writeable and settable by governance

[benhoneill]

I would hate to see this thread lost to inactivity and fall by the wayside. The value proposition for the $UNI token is primarily two fold:

1. “governance” (in quotes because the value of that function is yet to be proven in the market across all assets), and
2. the theoretical activation of a fee switch in the future, which is not yet guaranteed

While this concept needs to continue to be refined, it is in the communities best interest (IMO) to fund and support continued research into tokeneconomic design model changes based on the technology available in v4. Ideally, @mjc716 is not the only one that is working on this, nor is this the only idea considered.

I would formally propose the community discuss funding continued research into tokeneconomic designs to be presented, discussed, and, if any were to gain sufficient support, voted upon for implementation during the transition to v4.

Two potential options to fund this would be:

1. Take an XPrize approach whereby the winning “submission” is receives an outsized reward in order to attract the best minds in the industry to partake. This could be a flat reward (7+ figures of $UNI) or, to better align incentives, a % of captured upside shared from the protocol.
2. Set up a committee (akin to the Bridge Committee) to study the issue and propose options to the community.

In the spirit of decentralization, transparency, and openness, I’d love to see the community explore the appropriate structure for option #1 where everyone can partake, propose, analyze and iterate on the best solutions for the future for $UNI tokeneconomics.

May a rising tide lift all boats.

[AndyForman]

I like this. We should have a UNI token model that is aligned with the current incentives of the protocol AND is tenable for users (ie. not pay taxes). Believe with the right financial engineering / mechanism design this can be accomplished.

It’s worth over-spending on a short-term competition if it can lead to a sustainable model for the protocol.

The process to select a winner and implement seem important here. There is naturally some discretion in that process but something to think through as this continues.

[Jerome_Coinchange]

+1 on the gauges idea @Figu3 .
+1 on the fact that already circulating UNI should be put to contribution to this new model instead of relying 100% on UF funding it, @mjc716 .

A mix of automation and manual bootstraping seems to be the best combo. The UF seed the UCF while allowing external participant to “stake” $UNI to the UCF. The UCF hook by taking a portion of the $UNI side of the seeded pool is able to generate revenue, which it can then redirect to $UNI stakers and/or LP in of the seeded pools. In this case i’m not sure if we can call that “fee switch” but would be akin to what @MattOnChain described.

I’ve created a table with different variable to test out the hypothesis (see link at bottom). This is first draft, so many hypothesis are made on the numbers…

The TLDR is that using a gauge mechanism or simply rewarding stakers without token emission (unlike CRV) is quite difficult to achieve reasonnable incentives. We are talking about about less than 1% in ROI on lent $UNI for the UCF and max 0.5% for $UNI stakers (with UF seeding 65% of the UCF).
As an example with 100M $UNI lent with 65M from UF, it could theoretically create 856M in TVL for Uniswap. At an average 20% utilisation rate we have 171.2M in volume, of which 1% gets captured as fee and distributed to LP and UCF with different % ratio.
Below are the different models with different fee weighting: 80/20 in favor of LP all the way to 50/50 between LP and UCF.

The UCF revenue can then be allocated however Uniswap governance please, be it via gauge or simply rewarding $UNI stakers in the UCF to receive rewards from 0.23% to 0.57% per annum depending on the model.

This tax on the LP revenue could be considered the interest rate that LP need to pay to the UCF. The benefit of it is that it is not directly money out of the LPs pocket.

Staked $UNI in the UCF should still hold governance rights. If the DAO wantsto implement gauge voting, then a “veUNI” token should be created which would still hold 1:1 DAO governance vote but would need to also be able to hold more votes for the gauge voting based on the amount of time their UNI is staked in the UCF.

Here is the link to the excel sheet where all the calculation have been made.
Here is a screenshot of how the models scale as we increase the $UNI lent.

[zeb]

good to see Uniswap holders are still discussing tokenomics for Uniswap.
yes it is the market leader, no it should not take that position for granted.
veTokens without revenue share imo will just lead to bleeding your token.