Arbitrum LTIPP Incentive Matching

[AbdullahUmar]

Authors: @AbdullahUmar, @Juanbug (Uniswap Arbitrum Delegate Program (UADP))

TLDR:

• In April, the UADP submitted an application to participate in the Arbitrum LTIPP (long term incentive program pilot)
• We were able to receive 1,000,000 of ARB, the largest amount out of all other DEX applicants
• In our application, we stated that
  • “the UADP will request the Uni DAO to decide whether or not it wants to partially match this 1.0M ARB ask. The options the DAO has to vote on are four: $250k, $500k, $750k, and $1M. We cannot guarantee that the DAO will vote to match incentives, but we will make a best-effort attempt and report the results of the temperature check”
• We will be collaborating with Gauntlet and Merkl to distribute the incentives (cost breakdowns are below)
• To those ends, we are running this temperature check to see if the DAO is interested in matching the given LTIPP grant to some capacity

About the LTIPP Application

Uniswap did not apply to the first two rounds of Arbitrum incentives (STIP 1 & 2). This was for two reasons:

• Native DEXs and smaller protocols deserve a chance to make a name for themselves and perhaps offer unique protocols for trading
• We did not have a formal structure like the UADP to help facilitate an application

Since many native DEXs–and non-native ones–have already had the chance to apply for ARB incentives throughout the past year across multiple incentive distribution initiatives, the UADP decided that it’s now the right time for Uniswap to also partake in these programs, especially since other blue-chip protocols also applied for these incentives. Plus, the establishment of this committee has allowed Uniswap DAO to further mature its relationship with Arbitrum DAO. We’ve partaken in multiple discourses, participating in all of the votes taking place on Arbitrum since November 2023 (see Communication Thread here).

The LTIPP pilot program is a 3-month long incentive program, with 45M worth of ARB to be distributed among an elected group of protocols. A total of 174 applications were vetted by the LTIPP Advisors, with whom we interacted and obtained feedback during the month of March. One of the primary sticking points was the amount of capital that we initially requested, along with clarifications around some of the target KPIs behind our request.

After some deliberation, in line with many other projects, we ended up lowering our ARB ask from 2.5M to 1M, which increased our chances of being admitted into the program by the Advisors and ARB delegates. Uniswap was able to attain the largest amount of incentives from the cohort of DEX applicants:

Uniswap LTIPP Grant Structure and Execution

Below is a breakdown of how the LTIPP funds will be used:

• 900k ARB for incentives:
  • 882k ARB: The bulk of these funds will be used to incentivize liquidity providers on Uniswap using Gauntlet’s dynamic optimization engine.
  • 18k ARB: For Merkl to distribute the funds. Merkl charges based on a percentage of the incentives distributed.
• 85k ARB for Gauntlet: they will dedicate part of its Applied Research team, the same team currently managing the Uniswap/Arbitrum liquidity mining program, to this initiative.
• 15k ARB for UADP: These funds will be sent to the UADP multisig with the goal of making this meta-governance initiative a self-sustaining program. This will allow Uniswap DAO to maintain its voting participation in the Arbitrum DAO.

In order for us to accept this grant, we had to elect three signers onto a Gnosis Safe, and each member had to follow Arbitrum Foundation’s KYC process.

• Multisig Address: 0x1026D3D219098D7b1B0A180F7E557DEeA7DA82C1
• ⅔ Signers
  • @Juanbug (UADP), 0xB8Dcad009E533066F12e408075E10E3a30F1f15A
  • @AbdullahUmar (UADP), 0x3d0e30031b547737fFCf13c127350159A6C4ce17
  • Picodes (Merkl), 0x34Eb88EAD486A09CAcD8DaBe013682Dc5F1DC41D

All performance reporting will be conducted by Gauntlet, just as they have been providing analytics regarding their previous Arb-Uniswap incentive program here. Below are some basic metrics from the previous initiative.

Incentive Matching

It’s clear that Uniswap has a steeped history with Arbitrum. We are currently the dominant DEX by TVL and volume on the L2. After ETH L1, Arbitrum is where Uniswap has the largest stronghold. It is therefore important for delegates to potentially consider doubling down on this ecosystem.

Source: Dune Analytics (@whale_hunter)

Uniswap DAO and Arbitrum DAO have also conducted themselves symbiotically ever since Uniswap launched on the L2 in August, 2021. Below are some contributions that Uniswap has made to the Arbitrum ecosystem in the past year alone:

• Gauntlet has shipped out a Dynamic Incentive Optimization for Uniswap V3
  • Live dashboard
  • Uniswap/Arbitrum liquidity mining mid-point retrospective
• Set up a Uniswap-Arbitrum Working Group
• Funded a ~$2m Uniswap-Arbitrum Grants Program
• Created an active metagovernance group (Uniswap-Arbitrum Delegate Program)

Now that Uniswap has attained 1M ARB from Arbitrum, we are fulfilling our end of the bargain to see if Uniswap would like to reciprocate to some level:

As mentioned, it is up to Uniswap delegates to decide whether or not matching is favorable. Doing so would signify to the Arbitrum community our continued support, and in the UADP’s opinion, allow for further collaboration and grants in the future. If any funds are approved, they will follow a similar distribution structure and execution as above and be grouped together in Gauntlet’s incentives allocation.

Snapshot Options:

• $250k
• $500k
• $750k
• $1M
• Do Not Fund

[blockchainedu]

In support of matching incentives for the equivalent amount.

Arbitrum’s largest DEX is Uniswap ($316m) and Arbitrum is Uniswap’s second largest deployment. As @AbdullahUmar mentioned, it’s good to double down on a thriving ecosystem, especially since success on one chain benefits UNI holders across all chains.

LTIPP is $45M worth so Uniswap should continue applying for larger amounts from future Arbitrum programs to continue growing. This incentive from Uniswap’s side will help to show mutual commitment.

If Arbitrum says that future funding should only go to newer, smaller projects, then a key benefit of Uniswap is that it directly helps new projects by supporting liquidity for their governance tokens. Future incentives could be allocated towards those pairs if it’s more aligned.

[Doo_StableLab]

Curious to hear how so far the incentive that Uniswap has provided impacted growth of such chains. For example, I am aware Uniswap on Base is growing rapidly but how much is it can be contributed to the incentive program?

[WintermuteGovernance]

We are supportive of Uniswap contributing to additional incentives, however, we are unsure of the size given there isn’t a lot of information on which pools these incentives will be pushed to. So we think it would be quite valuable if Gauntlet would provide some thoughts on this prior to the Snapshot vote.

For example, we’d be supportive of a more aggressive incentive package if there are clear areas of liquidity in which Uniswap is lacking in comparison to Arb competitors.

[eek637]

Cross-linking from the arb forum here - gauntlet set out pretty clear KPIs there based on their experience running the program uni governance funded last summer.

[eek637]

I do think a retro of the non-gauntlet-run incentive program is a good idea. But I’d note that those pools are chosen less methodically than Gauntlet is suggesting here. The more comparable results to gauge this program’s potential (in my mind) are those from the 7 month program they ran with the ARB airdrop from last summer. They talk about those results in the second section of the STIPP post here.

[kfx]

Re the matching options:

@AbdullahUmar

Snapshot Options:

• $250k
• $500k
• $750k
• $1M
• Do Not Fund

I think there are ways how to approach the allocation.

The first is the purely utility-function-maximizing perspective. From my impressions on the liquidity mining program results, most of the incentivized pool’s TVL look quite similar to this (pic from Gauntlet’s dashboard):

Incentivized LPs are perhaps less forgetful that we hoped for, and don’t keep their liquidity in the pool after the incentives end. Sure, the LM programs have been quite efficient in attracting and retaining Uniswap’s market share for the specifically selected niche pools where it previously had almost none, but they have been less efficient for the short-tail asset pools. (Again, these are just my impressions, and I’m curious to see if there will be a final report of the Gauntlet’s LM program that confirms them.)

The second way is to approach if from a game-theory perspective, and just from the perspective of being a good actor in the ecosystem. Arbitrum has been great for Uniswap and DeFi in general. Here I see a much stronger motivation to contribute to the LP incentives, and try to match the Arbitrum’s contribution.

In conclusion, I think it’s a good move to match the amounts, but I want to be clear about my reasoning.

[Userisky]

Great job on securing additional funds.

I support the UNI matching incentive, but LP incentives have historically not been a good use of funds after the initial bootstrapping period.

It would be much more interesting if the ARB funds and UNI incentive matching were used for fee sponsoring experiments, such as paying for Uniswap users’ gas fees for specific pools via alternative frontends like @Oku. If the goal is to attract sticky LPs, focus on improving the swapper experience and activity.

The success of Uniswap on Base has a lot to do with user experience rather than LP incentives, in my opinion.

My vote for incentive matching would be:

250k - 500k if used for LP incentives

1 million if used for fee sponsoring experiments and the resulting data

[alicecorsini]

Receiving ARB incentives is great news and I am in favor of doubling-down with UNI incentives.

While Arbitrum is the largest non-Ethereum market for Uniswap in terms of TVL and trading volumes, it is lagging behind Base in terms of fees generation. Considering the large amount of incentives and the ongoing discussion on the fee-switch, this seems an interesting challenge.

While ARB incentives (as per the LTIPP application) will be focused on optimizing TVL gained per incentive spend and volume gained per incentive spend, it would be interesting for the UNI portion to be dedicated at improving fee-generation from the Arbitrum ecosystem. Or, to see how the ARB + UNI incentives impact fee-generation.

[kfx]

I think this is very important point. From a technical perspective, Merkl distributes the incentives based on three factors:

1. Fees earned
2. Active liquidity for asset A
3. Active liquidity for asset B

The weights of these three components is determined by the distributor, and as a result can be controlled by the DAO. The question is how much each component should be weighted? Here it stops being a purely technical discussion, as different stakeholders will have different preferences.

Fees are determined by and proportional to swap volume, so one of the stated goals of the Uniswap’s grant (“Volume gained per incentive spend”) is already optimizing for them, unless you meant something else.

The interesting aspect is that paradoxically, directly incentivizing fee generation is probably the worst option for LPs. Compare two extreme cases:
a) 100% weight on fees
b) 0% weight on fees, 50% on asset A liquidity, 50% on asset B liquidity.

The first option incentivizes rapid reallocation of capital, creating a very competitive environment where active LPs create a “race to the bottom”. Gains are expected to be minimal for most LPs due to inter-LP competition, operational costs, and impermanent loss. The second option incentivizes all in-range LPs equally, meaning that full-range positions probably get the best returns, as they have the least IL. In the real world, a case somewhere between these extremes probably is the best, but the selection of weights is actually a hard problem.

Also I encourage everyone to read Gauntlet’s blogpost on why liquidity mining was selected to be incentivized in the first place, and what are the alternatives.

[alicecorsini]

You are right that targeting swap volume also inherently targets fees. But, Uniswap’s fee-generation on Base is significantly higher than the one on Arbitrum despite trading volumes being lower.

In May, trading volumes on Arbitrum and Base were $11.59B and $5.24B respectively. Despite Arbitrum’s volumes being more than double, Base generated $9.37M in fees while Arbitrum only generated $7.49M.

It would be interesting to understand if incentives can play a role in improving Arbitrum’s performance. However, that might be out of scope here.

[kfx]

Now I get what you meant!

I’m assuming that Base’s higher fees are mostly driven by exogenous factors, not by difference in LP behavior, as Base has a much more active memecoin trading, which are more likely to be traded in higher fee-tier pools. I doubt that it’s possible to recreate such a memecoin ecosystem or Arbitrum, but liquidity incentives could attract higher fees per unit of volume in two ways:
a) by incentivizing longer-tail pools
b) by incentivizing higher fee-tier pools for short-tails assets

From these options, the second looks better from the LP perspective, as it’s lower risk.

While I know that Gauntlet has done lot of work on selecting the pairs to incentivize, I’m not sure they have equally thoroughly investigated the tradeoffs of incentivizing different fee tiers. Nudging them to prefer higher fee-tier pools sounds be something the DAO could do, but seems like it would conflict with the volume maximization objective. So I agree with you that it could be out of scope for this program, unfortunately. What do others think?

[Juanbug]

https://snapshot.org/#/uniswapgovernance.eth/proposal/0x82c77c3a10bc17ce65c3fa2fd553d224e00ccaa012ea5c203024fa3695acb7d0

Snapshot Vote is live!

[AbdullahUmar]

Thank you to all those who partook in this discussion.

Below are responses to some of the replies:

@Doo_StableLab

I am aware Uniswap on Base is growing rapidly but how much is it can be contributed to the incentive program?

There will be a review conducted on the efficacy of the incentive programs that the DAO’s running. The process for this and who will conduct it is still tbd. Most of these programs are either in the middle of being administered or have yet to start. The Base incentives were deployed on April 25 and will conclude on July 25.

Pools	Link
wETH/USDC 0.05%	0xd0b53d9277642d899df5c87a3966a349a798f224
cbETH/wETH 0.05%	0x10648ba41b8565907cfa1496765fa4d95390aa0d
USDC/USDT 0.01%	0xD56da2B74bA826f19015E6B7Dd9Dae1903E85DA1
wETH/USDT 0.05%	0xd92E0767473D1E3FF11Ac036f2b1DB90aD0aE55F

You can see the TVL and volume fluctuations in the incentivized pools above. Generally, the market has been more choppy since we applied the incentives–that’s why the trend looks generally towards the downside. The hope is that during drawdowns the severity of TVL bleed and volume decline is less due to incentives and that incentives actually lead to sticky TVL.

@WintermuteGovernance

unsure of the size given there isn’t a lot of information on which pools these incentives will be pushed to. So we think it would be quite valuable if Gauntlet would provide some thoughts on this prior to the Snapshot vote.

The UADP is working closely with Gauntlet to implement and report on Uniswap’s LTIPP allotment, although all the analysis and pool selection will be done by Gauntlet. The Uniswap-led incentive packages give the Accountability Committee jurisdiction to select which pools to deploy incentives to–these pools are selected with relatively loose criteria. If the DAO wishes to match the LTIPP allotment, we will be increasing the amount of incentives that Gauntlet will have at their disposal. The Accountability Committee will not be controlling the specific allocation of these incentives. In other words, if the DAO votes in $500k UNI as a match, then we’d just bundle that in with the existing 1M ARB, and Gauntlet will put that to work.

How does the Accountability Committee’s deployment of incentives differ from Gauntlet? The AC aims to incentivize blue chip pools to simply establish sticky liquidity and capture volume on commonly traded pairs like USDC/USDT or wETH/USDC. Many of these pools are low in liquidity to begin with due to being on a more long-tail or newer EVM deployment. Arbitrum is of course not in this position. The duration of the campaigns and gauges selected for each pool also matter.

As @kfx mentioned, Merkl has three gauges:

@kfx

• Fees earned
• Active liquidity for asset A
• Active liquidity for asset B

Gauntlet tends to follow a 98/1/1 breakdown (with 98 going to fees and the other 2 equally given to the other parameters). For the incentive programs being managed by the Accountability Committee, slightly different gauges have been selected, with both 60/20/20 and 40/30/30 structures having been implemented. The duration of Gauntlet incentives are also 2 weeks. The AC, however, usually distributes on a 3-month basis. This means that Gauntlet is involved in active management, constantly looking at data points from each previous campaign, factoring in current market conditions, and consequently selecting the best pools to deploy incentives to.

In a 98/1/1 model, 98% of rewards are given based on the fees generated from trades in the liquidity pool, encouraging LPs to move their funds to pools with higher trading volumes to maximize their fee-based rewards. This behavior leads to intense competition among LPs for high-fee opportunities, which can result in rebalancing costs and impermanent loss. But the 98/1/1 setting encourages LPs to allocate liquidity to where it is most efficient around the current tick, making it effective at driving volume growth. A flywheel is created here as well since more liquidity leads to better execution and therefore higher volume, which in turn increases liquidity. So to @alicecorsini’s commnet, this setup already favors fee generation.

In contrast, a 40/30/30 setup allocates 40% of rewards based on fees and 30% each on the liquidity provided for assets A and B. This approach reduces the need for rebalancing, encouraging LPs to maintain their positions longer across the two tokens in a pool. This setup can be beneficial for drawing more passive, sticky TVL. A drawback to this setup is that lazier LPs are often drawn to this incentive model–they simply allocate across the pool’s full range and are not significantly improving price execution per unit of liquidity provided.

Also, the previous program that Gauntlet ran focused largely on capturing market share from competing DEXs on Arbitrum. Today, Uniswap has secured itself as the de facto DEX on the L2, so the goals can change slightly. As Gauntlet points out:

“​​For this program, our methodology begins with a heuristic targeting underrepresented pools on Arbitrum (e.g., disproportionately low volume vs. other chains), aiming to increase overall TVL on the chain, differing from our previous approach that focused on capturing market share from competing DEXs. Once the pools are identified and incentives initialized, we will transition to a model-driven approach for subsequent allocations. The 30 pools you see in the first batch of recs were ones we identified as underserved and having filtered out some tokens manually after additional DD.

Unlike the previous Arbitrum LM campaign, which relied on a simulation-based framework to identify pools with the greatest boosts to price execution under different liquidity scenarios, this program employs a predictive model that reallocates a fixed budget towards pools with the highest response to incentives, ensuring dynamic and efficient allocation. This method maximizes TVL growth by directing resources to pools with the most significant growth per dollar of incentives. We’ll also consider adding and removing pools as the program goes on if some pools are unresponsive to incentives - this is something we’ve observed in the past and quickly respond to maintain incentive spend efficiency.”

The Gauntlet team has completed their first analysis on pools, having selected 30 of them for the first two-week distribution of 150k ARB: https://arbiscan.io/tx/0x4ed72c0d11f7a12b83cc4f521999216b5ae723a02c1aa3ea69073e44f966887c

@kfx

Base has a much more active memecoin trading, which are more likely to be traded in higher fee-tier pools. I doubt that it’s possible to recreate such a memecoin ecosystem or Arbitrum

As for this comment, Gauntlet will optimally select which pools and their respective fee tiers to target. With the above batch, for example, ~5% of the incentives will be given to the PEPE-wETH 1% pool.

[Tane]

Thank you @AbdullahUmar for the proposal and we are glad to see the commitment made in the original Arbitrum LTIPP application has been addressed in an appropriate way in the Uniswap governance.

Regarding the amount of the matching, from our perspective, Uniswap has already a large position on Arbitrum One, thus the use of additional incentives would be limited while Gauntlet’s approach to incentivizing inactive LPs seems appropriate. We also looked into the Base case as a success case as we understand that the incentive implemented by the DAO contributed to a significant increase of its fee revenue, volume and TVL on Base, and the amount used was approximately $500k ($492k to be exact.)

Based on the above, we believe the amount to be used for the matching should be $500k and we would vote for $250k, $750k and $1M in this order.

[Userisky]

Why are the 500k votes not being shown in the poll?

[Juanbug]

It’s ranked choice voting, so while the way they display it looks a little confusing, the vote should finalize in the best “total preference” selection.

[Userisky]

The ranked choice examples in the link shown %'s going to all the choices at some proportion based on voting weight.

The current poll is showing 0% for three options and only accounting for the top 2 to equal 100%.

[Juanbug]

“In the second step if the first-choice candidate doesn’t get over 50% of the total votes the choice with the fewest number one votes is eliminated . Voters who had chosen the defeated choice as number one now have their number two choice counted as their number one choice.”

Basically some 500k and 1m first preference votes are being tallied, but since they are selected least, after the first step, they are dropped from the second step.

[kaereste]

The following reflects the views of L2BEAT’s governance team, composed of @kaereste and @Sinkas, and it’s based on the combined research, fact-checking, and ideation of the two.

We’ll vote for the proposal and opt to match the ARB received by LTIPP with $750,000 worth of UNI, going for an almost 1:1 matching in USD terms.

As outlined in the report provided by Gauntlet, the impact previous incentives had was extraordinary, and therefore we believe it’s worth doubling down on it. While it’s true that Uniswap has cemented itself in Arbitrum, we should seek to remain competitive, especially when other DEXs will also be offering incentives.

One thing we’d like to see the matching funds used for, however, is to either incentivize more pools or extend the timeline of the overall incentive distribution. If we are to match, we believe we shouldn’t simply double the amount of incentives distributed in the same pools over the same period.

Having said that, we understand that Gauntlet will be responsible for utilizing the funds in the way that makes the most sense, and we’ll trust their judgment in the matter.