ARC - Consolidate Aave V1, V2 & AMM Reserve Factors, Purchase CVX and Deploy to Earn Yield

title: Consolidate Aave V1, V2 & AMM Reserve Factors, Purchase CVX and Deploy to Earn Yield
author: Llama Community @MatthewGraham @Dydymoon
created: 03/01/2022

Simple Summary

The Aave V1, V2, and AMM market Reserve Factors (RF) on the Ethereum network hold a combined value of over $28.27M. The Llama Community proposes consolidating the RF into the V2 RF and deploying the capital with the intention of retaining TVL and earning additional yield. The yield generated comes with the added benefit of acquiring governance tokens of other DeFi protocols and increasing Aave’s metagovernance influence.

Abstract

The Llama Community recommends implementing several yield generating strategies spanning Balancer, Curve, and Convex and subsequently participating in the respective communities’ tokenomic constructs. Aave will earn yield in the form of swap fee income, aToken yield derived from Aave V2, and governance token incentives in the form of BAL, CRV, CVX, stkAAVE and LDO.

The first investment theme is deploying capital into the Balancer Boosted Pool and BTC Stable Pool to earn a vAPY of around 13.4% and 4.6% respectively. The Balancer Boosted Pool invests around 80% or more of the deposited liquidity into Aave V2 to earn yield whilst also generating swap fees, stkAAVE, and BAL rewards. The Stable Pool generates income via swap fees and BAL rewards, with the latter making up the majority of the incentives on both strategies.

The second investment theme is more complex. It involves acquiring both CVX and veCRV to participate in gauge weight voting on Curve in order to influence a portion of the CRV inflation schedules to be distributed to aToken pool Liquidity Providers (LPs). Aave will deploy stables and ETH from the RF to earn additional CRV and CVX rewards, which will be used to further boost the incentive allocation to the aToken pools.

The data presented here is accurate as of the end of December 2021 data.

Motivation

Part A - Consolidation

The legacy V1 market continues to generate mostly flash loan fee revenue and this has slowed down in recent times. There is only ~$219M of TVL remaining in the V1 market, and at the rate of which revenue is generated is expected to continue declining as more and more users move to V2 or V3 over time.

This proposal recommends consolidating the Aave V1, V2, and AMM market revenue into a single location before deploying the capital to earn additional yield. The V1 RF holds around ~$821K of assets at the time of writing, whereas the Ethereum V2 + AMM market holds $27.45M of assets. Consolidating the RF balance would increase the mainnet treasury to around $28.27M. A summary of the main assets held in the V1 RF are shown below:

Part B - Retain TVL & Increase Returns

Whilst holding aTokens is arguably one of the lowest risk yield opportunities in DeFi and something Llama Community encourages most, if not all, treasuries to hold. The Aave DAO is fortunate since it receives income in the form of aTokens. Thus, Llama recommends firstly making all unproductive stable coins productive and then optimizing further by earning additional yield and metagovernance influence.

The two main high level strategies proposed in this ARC share three common goals:

  • Retain TVL within Aave
  • Exceed aToken returns
  • Accumulate metagovernance influence

In addition to these three goals, the Convex / Curve strategies have the added benefit of improving the extrinsic use cases for aTokens. aToken holders will be able to deposit liquidity into the Curve aToken pools and then deposit the Liquidity Provider (LP) token into either Curve or Convex to earn additional yield.

The Compound pool on Curve has $174.5M of TVL relative to the Aave pool which has $55.3M of TVL.This is most likely due to the higher incentive APR that the Compound pool, 10.13% vAPY, relative to the Aave pool earning 7.58% vAPY on Convex.

Specification

Overview

Llama proposes diversifying $26,67M (94.36%) of the combined Ethereum RFs across three protocols: Convex, Curve and Balancer. The combined Aave V1, AMM, and V2 market RF totals $28.27M at the time of writing. A detailed breakdown of the consolidated constituents can be found on the Aave Treasury Finances by Llama Dune Dashboard and a high-level overview is shown in the figure below.

The table below shows the assets that will be deployed within this proposal. Note this is only 94.4% of the assets and 5.6% of the assets are to remain unchanged within the Aave V2 RF.**

The table above includes the following stablecoins, DAI, USDC, USDT, TUSD, BUSD and sUSD from the Aave V1 and V2 markets. The AMM markets holdings have not been included as they only total $119.3K at the time of writing.

The table below shows asset allocation per each strategy:

The balance within the Aave V2 RF, not deployed in one of the above strategies is shown as Other (Longtailed Assets) in the figure below. The percentages reflect the overall composition of the RF, noting that 94.4% of the consolidated RF is deployed into one of the strabled detailed in the above table.

The actual implementation of this proposal is likely to be spread across several AIPs. This is to reduce the implementation complexity and allow the several positions to be swapped / redeemed before being deployed into a bonding curve whilst we await the Chainlink CVX oracle being released and getting the Aave V2 RF whitelisted to deposit into the CRV locking contract to attain veCRV.

The subsequent sections provide further insight into each strategy and how they are to be funded.

Convex / Curve Strategy

Convex Allocation

The Convex (CVX) allocation is to be funded by the assets mentioned in the table below:

In order to fund the initial $2M purchase of CVX, all of the AMPL, BUSD and TUSD will be used ($842.2K) and the balance, $1,16M, will be funded with USDC. This is the initial purchase of CVX and with a unit price of $43.20 is equivalent to about 46,300 CVX tokens.

We expect to make additional CVX purchases in order to meet the minimum voting hurdle required to influence how Convex’s veCRV holding is deployed on Curve gauges. Based upon this snapshot and the Voting Rules and Information section of Convex’s documentation, we estimate around 70,000 CVX votes are needed to clear the hurdle rate. This is expected to cost a further $1.02M to what is proposed above and represents less than 1 months revenue. We intend to fund this through future revenue and have a sufficient buffer to the hurdle rate based upon voter participation on Convex’s snapshot.

The assets are to be converted to ETH and then used to purchase CVX via a Bonding Curve with a 50 bps premium that utilizes a Chainlink oracle price feed. Alternative methods can be discussed in the comments. However, we opt for a solution that can be implemented entirely through the AIP governance process and does not include manual trading.

AMPL is the most illiquid of the mentioned tokens and this position will need to be converted to ETH over a series of transactions. Trades are to be routed through an aggregator with MEV protection like CowSwap. The most liquid AMPL pools are on Uniswap V2 and Sushiswap. Wherever possible, aTokens will be swapped rather than redeemed in an attempt to retain TVL in the Aave V2 market.

The newly acquired CVX will be locked for 16 weeks with voting rights delegated to Llama in order to vote on both Aave and sAave gauges via Convex’s snapshot to increase the amount of CRV rewards received on the Curve pools. Voting occurs every 2 weeks via Snapshot and thus the need for a delegate.

The locked CVX also receives a portion of the protocol fee income and votes on how Convex’s veCRV are distributed across Curve’s gauges. Llama Community acting as the delegate will vote to boost CRV rewards on the Aave pool as a priority and sAave once sufficient voting influence has been attained.

4 Year Locked Curve (veCRV) Allocation

The aCRV tokens that have accumulated within Aave V2 are to be redeemed for CRV and then locked for 4 years on Curve to receive an equal amount of veCRV. This strategy is subject to Aave’s V2 RF being approved and whitelisted via Curve’s governance process.

The veCRV position will then be used to vote for CRV incentives to be allocated to the Aave and sAave pools via the respective gauges. Based on the CVX strategy discussed prior, these veCRV votes are likely to be allocated to the sAave gauge.

Subject to the timeline and attaining the whitelist permission to deposit into Curve’s veCRV contract, the alternative strategy is to convert CRV to cvxCRV, then stake to earn CRV, CVX and 3crv, pulse any airdrops that go to veCRV holders. This is expected to generate 47.35% vAPY, see snippet below.

The earned CRV will be converted to cvxCRV and this will compound over time. The earned CVX will be locked to increase Aave’s voting influence and the 3crv will be exchanged for CRV, then converted to cvxCRV to further compound the strategies returns over time.

Productive aToken Stable Strategy

The sAave and Aave strategies include depositing capital into a Curve pool and then staking the LP token on Convex to earn stkAAVE, CVX and CRV rewards. The two strategies are expected to generate around 4.07% vAPY and 6.19% vAPY before Aave’s CRV and CVX voting influences the CRV incentive derived yield.

The table below shows the allocation, how each strategy is funded, and which pool those funds are to be deposited into.

Productive ETH Strategy

After transferring the ETH from Aave V1 RF to Aave V2 RF, the awETH is to be redeemed for ETH and the combined ETH deposited into the Curve stETH pool. The LP token steCRV is to then be deposited and staked in Convex to earn CRV, CVX and LDO incentives. The expected yield is around 6.65% vAPY.

No vlCVX or veCRV voting rights will be used to attract CRV incentives on this pool as the pool does not contain any aTokens.

Balancer Strategy

Balancer Boosted Pool Strategy

Balancer has recently launched the first Boosted Pool on Wednesday 16th December and liquidity mining incentives have already commenced generating around 13.4% vAPY at the time of writing. Using the Stable Pool as a guide, we expect Balancer to grow the Boosted Pool with the use of Liquidity Mining (LM) and even replace the Stable Pool in time.

The Balancer Boosted Pools earn swap fee income, LM incentives, and a new type of yield derived from unproductive capital being deposited into Aave V2 to earn yield. Llama expects the yield of the Boosted pool to exceed the Stable Pool in order to entice users to move across. Llama expects LM incentives to be the mechanism for achieving an elevated return while the pool is in its bootstrapping phase.

The strategy for Aave is to deposit an equal amount of DAI, USDC and USDT into the Boosted Pool. All BAL rewards are to be retained to accumulate governance influence.

Productive BTC Strategy

After transferring the wBTC from Aave V1 RF to Aave V2 RF, the awBTC is to be redeemed for wBTC and the combined wBTC deposited into the Balancer BTC Stable pool. The LP token staBAL3-BTC is to be held in the Aave V2 RF and the expected yield is around 4.60% vAPY. The BTC Stable pool receives around 4.07% vAPR in BAL rewards.

Conclusion

The implementation of the above proposal is expected to be performed in parts as it involves deploying capital into six different yield generating strategies, Aave being whitelisted of Curve and swapping existing assets for CVX. This proposal includes the Aave DAO’s first allocations to another governance token and the first deployment of community’s capital to earn additional yield. The increased APY on the two Curve aToken liquidity pools is expected to improve the extrinsic use case of the most commonly held aToken as well as increase the CRV and CVX rewards Aave DAO receives for being an LP in the two pools.

Copyright

Copyright and related rights waived via CC0.

16 Likes

Hey,
Glad to see that Aave’s governance finally wants to generate yield on the accumulated treasury and enter the Curve Wars.
For the most part, I agree with the proposal. However, there are several changes that I would like to propose.

1. Strategy Overview

I think the $CVX strategy would be an excellent decision for AAVE governance but I think this strategy would be conservative for the protocol. If Aave really wants to get into the Curve Wars I think they need to buy a minimum of $5 million worth $CVX.

I propose this new allocation:
image

2. Convex Allocation

Below is the new allocation for the purchase of $CVX:
image

Buying $5 million worth $CVX at the beginning would solve this problem. I think waiting another month would just be a waste of time for Aave. If we have the 70,000 $CVX required at the beginning, we can have more voting power on the Curve governance. This would increase the yield on Curve pools, and therefore the TVL of these pools, which would increase the amount of stablecoins deposited on Aave. This could reduce the average borrowing rate on Aave for stablecoins and thus encourage more stablecoin borrowing. This would bring in more revenue for the protocol in the medium term.
We would lose 1 month of yield on $3 million but we would not have to pay a premium on the purchase of new $CVX and we would also generate more revenue for the protocol but we would also generate revenue from $CVX locks (4.61% vAPR). All of this could make up for some of the lost yield.

In addition, owning $CVX not only allows one to have voting power over Curve’s governance but also over other governance bodies such as Frax Finance where Convex already owns 2.5% of the protocol’s voting power. In the future Convex also plans to deploy on other protocols and other blockchains.

Owning $CVX would therefore allow Aave to have voting power in many protocols. In the future, we can make a partnership with the Frax governance to launch a Curve aFRAX+Aave pool that we could incentivize by voting with the $CVX owned by DAO and not spend a budget to incentivize this pool. (it’s just an exemple with Frax but we can make this with others protocols who their token on Aave)

3. Productive aToken Stable Strategy

The new allocation if we buy 5 millions worth $CVX:
image

4. Balancer Boosted Pool Strategy

The new strategy with the new allocation for Aave to deposit Stablecoins in the Boosted Pool:

image

Other ideas:

One idea could also be to set up a recurring purchase of $CVX. I was thinking of $250k worth $CVX per month for 3 months to start and then that could be renewed for 3 months if it works well. This would ensure that the protocol keeps a decent voting power in Curve’s governance.

I would also like to add that buying CVX to vote for Curve pools during gauges weight allows to increase the TVL of Aave in an indirect way, this could also be a way to reduce the budget of liquidity mining in $AAVE currently in place. Another way that could be more efficient for the protocol could be to create Curve V1 or V2 pools with the aTokens we want to incentivize and use bribes on Votium. But this is only a medium term idea for Aave.

13 Likes

Figured that you all would want a smooth brain to big-up this proposal in the comments.

Big-up this proposal!

1 Like

Hey @MatthewGraham - interesting proposal. Would like to focus on this part first as it is something I have experience with and pioneered elsewhere:

I agree this should be a common goal of most mature DAOs. How do you plan on controlling the day-to-day voting needs of the DAO? Will this be through community votes or an internal committee?

2nd - I would like to point out 94.4% is a massive amount of the treasury to place in other smart contracts and incur counter-party risk. More than 2/3 (67.43%) of this strategy will be in Convex alone.

While this is a mature, user-tested protocol, losing $17.9mm to a single protocol exploit would be devastating. Have you considered taking out insurance on these positions?

Third - let’s focus on the Curve steCRV strategy. Would you opt for the stETH Curve pool instead, sacrificing a bit of yield for more LDO and early participation in their governance?

This can continue with the goal of meta-governance efforts, diversifying exposure, and create a clear path to liquid staking down the road.

1 Like

First off, love seeing the treasury put to use, and looking forward to seeing the Polygon and Avalanche holdings activated as well.

Originally had similar concerns to @fig did above on the insurance side, however examining the on chain options, Nexus doesn’t have the capacity to insure deployment of this scale.

While this is a large amount, it is relatively small compared to the total treasury holdings, if you include the roughly $500M in Aave tokens the treasury holds, this is a much smaller total percentage.

I would love to see Aave DAO take a stance similar to how BadgerDAO has used their Convex/CRV holdings (only voting for the appropriate gauges). Without some sort of committee, like IndexCoop uses, full DAO votes for every proposal can become very difficult.

3 Likes

Thanks for this great feedback !

I fully agree with you, 2M$ is a conservative amount, and Aave definitely needs more than that.
We thought about starting with a lower amount, to increase the chances to get this voted, and to show results in the coming weeks.

However, I’d vote for your new split with buying 5M$ worth of $CVX too.
The recurring buy is indeed very important, even if the protocol will farm some with the strategies deployed.

Llama also thought about starting a bribe program in StkAave to incentivize $veCRV and $vlCVX holders to vote on Aave pools too, increasing the possible yield in CRV.

The goal here is to look at the improvements on the gauges after the first CVX buy and CRV lock, then, if it’s not enough to improve the yield on Aave pools, Llama will make a proposal about a bribing program.

3 Likes

This is super cool for AAVE markets and AAVE holders in general.

Also since Polygon market generates about 3M in weekly revenue for the DAO - it would be a good step to start the conversation for Polygon market too.

Because Curve gauges/Convex will be going multi chain so its in AAVE community’s interest to prepare for that.

Some comments about this proposal, mainly following with the ideas presented by @MatthewGraham and @JeanBrasse .
First of all, fully agree that it is necessary to increase the performance of the Aave treasury. Holding aTokens is already a really good strategy, but the composition of those on top creates interesting feedback loops.
My points that I want to highlight:

  1. If a strategy involving acquiring assets like CVX is implemented, considering the amounts discussed ($2m-$5m), is the optimal option to market-buy on DEX?
  2. Was there any target APR considered in general @MatthewGraham? Do you have some numbers on how much total APR would the strategy generate across all assets and allocations per platform?

  1. Kind of agree with the previous, it is a decent technical risk to allocate quite a lot of capital in only 1 platform. Considering TVL over time, I assume is pretty market proven, but would be important to show to the community the security procedures done on the Convex side.
  2. In a more fundamental layer, the investment on Convex pools has deeper implications, specially following a strategy to lock-to-boost yield. First, there is the market risk of acquiring CVX itself; yes, it allows to use $2-5m to then increase returns on a way bigger amount (stablecoins/ETH deposited on the pool), but the dependency on the platform becomes quite hard.
    As mentioned before, it is kind of useless to not buy-and-lock “enough” to start participating on the boost dynamics of yield on Convex. But at the same time, the investment and dependency on CVX/Convex only grows over time, so in 1 year we could find ourselves in a situation where the treasury holds $10m of locked CVX, and actually is not so profitable anymore to have the rest of the assets deposited on the Convex pools. In that case, we just hold an asset (CVX) which only value is potential price appreciation.
    In general, I’m personally skeptical about dynamics like Convex’s, specially when trying to involve the Aave dynamics, because even if I see the game-theoretical points of it, the fundamentals are pretty weak in my opinion, compared with more basic layers like Curve.
    We should also ask ourselves if it is the way to go for Aave to start participating on such “game”, involving dynamics like systematic governance bribing, which even if I understand the advantages of, find quite uncomfortable to normalise.
  3. Concerning my specific expertise, how is the plan to implement this features? I understand you mention @MatthewGraham that will be done in parts, but I would like to know who exactly will execute the implementation, security review of the contracts and related details.
2 Likes

Thanks to Llama community for this thoughtful proposal. By way of introduction, I’m leading the governance team at Avantgarde Finance; we’re the lead maintainers of Enzyme. We generally agree with the direction you’ve taken here and think Aave’s treasury would benefit from more active management, as we’ve written before. The plan you propose here creates a virtuous flywheel for the treasury and for Aave’s users; it increases liquidity for aTokens, it increases aToken supply, and it increases Aave’s inter-protocol governance stature, all while earning a fairly safe return. So at a high level, we agree with the allocations as designed in this post. We do, however, have a couple of concerns.

  1. As @eboado points out, there is a heavy concentration around Convex. This is not necessarily a bad thing in the short to medium term, but it does commit Aave to one protocol out of many that are working to monetize Curve’s governance power. Just gaming things out, it it conceivable that the yield from staking Curve LP tokens into Convex pools erodes over the course of the next year to the point where better options are available. The treasury would still be holding a significant position in the CVX underlying, and the unwind of that would be both very slow and very public (imagine a temperature check called ”Should the treasury sell its position in CVX”) . The nature of asset management for protocol DAOs such as Aave at the moment means that there will be large trades that are telegraphed well in advance; execution and performance will suffer as the market front-runs any planned rebalances. With all of that said, we don’t take issue with the allocation to Convex, but rather just want to raise the question of how to manage it on a medium to longer time horizon.
  2. The operational aspects of this proposal are… daunting. Do you have a process mapped out at the transaction level, along with noting which of those transactions require governance votes?
  3. As it stands, the governance rights that will accrue to the treasury (primarily in CVX) are to be delegated to Llama Community. In general, delegation makes sense. In our opinion inter-protocol governance actions should not be the subject of protocol-wide votes and should be handled by a smaller committee or an individual actor incentivized to vote with the protocol’s interest in mind. However, we’d like to see some more structure around this delegation. Some things to consider here include term limits, reporting requirements, compensation, et cetera.

An Enzyme vault is a simple solution to these concerns. At a high level, the steps would be as follows:

  • The RF is aggregated to one address.
  • Aave governance deploys an Enzyme vault into which only the RF could deposit.
  • Using Enzyme’s policy suite, we configure risk management controls that only allow the manager of the vault to interact with a pre-defined set of external protocols, hold a pre-defined set of assets, and accrue a pre-defined amount of trade slippage. These controls will be clearly visible to stakeholders and are enforced at the smart contract level.
  • The RF seeds the vault, and Llama are appointed vault manager.

This solves the first concern by giving Llama the flexibility to manage the portfolio without telegraphing every move to the wider crypto community and the second by granting permission for Llama to do the set of things it’s proposing to do here without going through the DAO voting process for each transaction

The third concern is a little more involved, and likely still involves building out some framework around the vault’s management. E.g. after a certain term, an election is held in Aave governance to re-elect Llama as the vault’s manager, or to transfer manager responsibilities to a new candidate.

Again, we are directionally supportive of this proposal, and think that its investment themes strike a responsible balance while also providing non-financial benefits to the Aave community. We’re excited to continue this conversation.

3 Likes

Just following up here to see if you had any further thoughts @MatthewGraham @Dydymoon ?