ARC - Aave V2 - Liquidity Mining Program (90 days at 30% reduced rate)

aip: TBA
title: Aave V2 - Liquidity Mining Program (90 days at 30% reduced rate)
status: Proposal
author: @MatthewGraham
created: 2021-10 25

Simple Summary

Liquidity mining incentives were introduced to Aave V2on the 26th April 2021 via AIP-16 and then voted to continue on the 24th August 2021 via AIP-32 for an additional 90 days. AIP-32 incentives are due to expire 22nd November 2021.

This ARC presents the community with the opportunity to continue offering stkAAVE incentives for an additional 90 days from when AIP-32 finishes, up to including the 20th February 2021. The incentives will be distributed at a 30% reduced rate, commencing the tapering of incentives on the Aave V2 market.

Stable coin / low vol assets receive lending and borrowing incentives, split 1:2, in favour of borrowers over lenders. Aave V2’s revenue is derived from interest paid by borrowers and borrowing demand drives lending yield through higher liquidity utilisation. High vol assets receive lending incentives, skewed towards assets more recently listed and communities that have static relationships with Aave, like Balancer and Chainlink.

Abstract

The intent is to recognise liquidity mining incentives played a large role in helping grow Aave V2 TVL, whilst being conscious that competitors continue to offer incentives APRs marginally exceeding what Aave is currently offering.

When incentives were reduced aggressively on the Polygon market, the daily revenue reduced significantly. There is a risk that by tapering liquidity mining incentives on Aave V2 too aggressively could trigger the same outcome as what was experienced on the Polygon market.

Based on the learnings from the Polygon market and reviewing the governance forum, this proposal reduces Liquidity Mining rewards by 30% over the next 90 days. For the following 90 days during 2022, the logic would be to continue further reducing incentives pending how the market reacts to the implementation of this proposal.

It is to the benefit of the protocol to incentivise borrows more than depositors. Conceptually, incentivizing borrows already increases deposit rates, so passive depositors still benefit significantly.

To encourage the depositing of high volatility assets, this proposal suggests incentivising the lending of newly listed assets and continuing to offer incentives on those that are strategically important for the Aave community. Offering incentives on newly listed assets is expected to kickstart the supply of liquidity.

Motivation

As a community member, liquidity mining has been great for bringing more users as stakeholders for the Aave Protocol and distributing the governance power to a wider more decentralised group of holders. Distributing stkAAVE means users also backstoppers of the protocol and automatically have skin in the game.

I would say that most liquidity that a liquidity protocol such as the Aave Protocol needs is stablecoins - incentivised as diversely as possible and also encouraging stablecoins that are aiming towards decentralization such as RAI, FEI and FRAX.

I would like to present to the community an opportunity to continue distributing stkAAVE at the same rate per day across the Aave V2 market with the intention of achieving the following:

  1. Grow Total Value Locked (TVL)
  2. Increase liquidity
  3. Attractive (low) borrow rates
  4. Increase the protocol income via growing the Reserve Factor
  5. Redistribute governance power towards users of the platform

As the value deposited within the Aave V2 market has grown from ~$5B to ~$21B and the $AAVE price has underperformed this growth, the incentive APR is lower now relative to the start of the prior incentive period.

On the 17th June 2021, the incentives on Polygon were reduced by 5x and revenue fell from $55.77K per day to $32.27K per day. This highlights the risk of reducing incentives to hastily. It is worth noting Compound is providing liquidity mining rewards similar to the existing 2,200 stkAAVE distribution level in terms of APR across key stable coin (low vol asset) markets.

Ref. Details: Phase 1 and Phase 2 Liquidity Mining on Polygon.

Specification

The below section outlines the proposed liquidity mining incentives to be applied from 22.11.2021 up to and including 20.02.2022.

Key changes from the previous liquidity mining campaign:

  • 30% reduction in incentives to 1,540 stkAAVE per day
  • New additions BUSD, FEI, FRAX, CRV and DPI have all been included
  • Removed UNI and GUSD incentives
  • Incentives are skewed in favour of borrowers over lenders for low vol assets
  • Borrowing incentives are not offered for borrowers of high vol assets
  • Distribution of stkAAVE is pro rata based upon deposit dollar value

The table below shows the intend stkAAVE distribution across the various Aave V2 listings. The overall number of stkAAVE distributor per day is the same and with a larger capital base, generating to a lower APR. The allocations are based loosely around the dollar value of deposits (lenders) and position users to borrow stables. The intent is to drive borrowing demand which will then lead to a higher lending APR. This is further encouraged by incentivising depositing of higher vol assets as collateral which will grow the TVL.

Original Proposal.


Note: TVL per token is as per the 10th November 2021.

Final Proposal


Note: TVL per token is as per the 21st October 2021.

Math: % of value deposited x multiplier then scaled to provide LM Allocation that totals 100%. Link here.

All gas costs incurred implementing the AIP that follows on from this ARC is to be reimbursed by Aave in the form of $AAVE.

Rationale

Below is a list of the key considerations made when designing the third iteration of Liquidity Mining for the Aave V2 market:

  • Compound continues to offer liquidity mining incentives that marginally exceed those on Aave V2 marketplace. Thus, the risk is that if incentives are reduced aggressively capital migrates to another platform offering more compelling incentives.
  • When the Polygon market aggressively reduced incentives, TVL and therefore revenue dropped significantly. There is a risk that similar behaviour could be experienced on mainnet.
  • Borrowing is the key activity on Aave. Only borrowers pay a fee to use Aave’s liquidity and Aave’s portion of this fee accumulates in the Reserve Factor. Thus, borrowers technically provide more value to the protocol by:
    • Increasing depositor interest
    • Increasing reserve factor earnings
    • Incentivizing borrows more than deposits reduces the maximum earnings from a “leverage loop”
  • Aave V2 TVL is heavily composed of three main stable coins (low vol assets) USDC, DAI and USDT. In effort to support the largest revenue generating pools, incentives are skewed towards low vol assets.
  • In an attempt to grow TVL and capture the less centralised low vol asset markets, incentives for lending and borrowing assets like RAI and FEI have been introduced and are treated in a similar fashion to existing more centralised low vol assets like USDC. Note, neither RAI nor FEI are listed on any other Tier 1 money market protocol.

Capital Efficiency and Risk Considerations

  • We did consider introducing a high health factor >2 which reduces the appeal of gamification and recursive lending. However, the implementation was ruled out as too much work and increased centralisation of the incentive scheme.
  • Recursive behavior does not significantly increase risk as it is typically looped within a single asset. The Aave V2 market is designed to accommodate high levels of borrowing and the risk parameters set in place ensure the protocol is in a healthy state. The risk within the protocol is not connected to liquidity mining, but to the overall proper risk parameter management.

Economics

  • Aave V2 markets daily Reserve Revenue is averaging around $170K per day. Daily aToken interest is around $3K per day.
  • Liquidity Mining is estimated to cost $492,800 per day assuming 1,540 stkAAVE at $320 per token. Down from $704,000 per day assuming 2,200 stkAAVE at $320 per token.
  • Ecosystem Reserve is worth around $700M with an AAVE price of $320.
  • 90 days at 1,540 stkAAVE at $320 per token is $44,352,000 or around 6.33% of the Ecosystem Reserve.
  • The Net spend is $29,052,000 per day after taking into consideration the daily Reserve Revenue and daily liquidity mining expenses. This does not take into consideration the 550 AAVE/day Safety Module incentives
  • Daily Reserve Revenue is continually growing over time, any Reserve Revenue moving average trends higher from the middle of May onwards when the price of ETH dropped 60%.

Further details on the daily Reserve Revenue and Economic Reserve valuation can be found on the Aave Treasury Dashboard created by the Llama Community.

Copyright

Copyright and related rights waived via CC0.

10 Likes

@MatthewGraham thanks for the detailed post.

A few questions off the bat - about distribution and the data you included.

What are these values for $XYZ and $ABC? Seems like the numbers may be missing.

Second you acknowledge the worth in incentivizing borrowing:

We agree with this hypothesis however are not quite sure it is reflected in the proposed distribution table. New assets introduced (BUSD, FEI, FRAX, CRV and DP) lean towards borrowers yet the total program is skewed heavily towards lenders - the split is 11 for borrowers, 20 for lenders.

This distribution is nearly 2:1 in favor of lenders. ETH for example is 100% split for lenders, and accounts for 36% of the liquidity mining allocation. Does this seem fair? There are assets in that list with more volatility.

This breakdown also seems to offer commentary on the future of the market. By shifting the LM split towards borrowers for stables, borrowers shoulder the risk of bullish price action. It is a bet that prices go up. What if the opposite happens? Shouldn’t borrowers benefit from a lower cost of collateral?

Is the distribution able to be variable based on stAAVE prices? To better achieve cost efficiency it seems advantageous inject more liquidity via stAAVE while prices are depressed.

We think this proposal needs more work but has good intentions.

5 Likes

The proposal favours low vol assets like stable coins relative to high vol assets like DPI, there is a 1.5x multiple applied in the linked excel spreadsheet. For high vol assets there are only incentives for lenders, thus encouraging users to deposit a high vol asset to earn incentives therefore growing AUM. Once deposited, users are then incentivised to borrow stables. The goal is to attract high vol asset TVL and then users have the ability to draw a loan as they please.

The proposal didn’t include CRV as AIP 5 was executed back in 2020. The proposal shows favour to more newly listed assets in attempt to grow AUM that way.

The yield on ETH is greater than the AIP-32. I am open to reducing this and reallocating this stkAAVE across other assets. A great thing about forums like this is we can fine tune proposal together :slight_smile:

I am not sure I follow the logic here. With incentives skewed towards borrowers, the cost of capital for the borrower is lowered by the incentives. If I guess as to what I you could be saying is if we incentivise lending, borrowers will come ? If incentives were to be turned off, we want sufficient borrowing to create a yield that will continues to attract TVL. For the yield to be sustainable, there needs to be good utilisation of the capital within the pool.

Ideally, IMHO with respect to low vol assets, we could in future perhaps only incentivise borrowing and see the interest rates generated are sufficient to attract new capital. That would be pretty cool dynamic to emerge.

3 Likes

What’s the rationale behind removing rewards for GUSD and UNI? Also, you say that BUSD have been included, but the distribution chart shows zero rewards for this asset. The overall intentions stated in this proposal seem good, but the proposed incentives look an awful like just arbitrarily picking winners and losers to me. Which would be fine if there was a more formalized process for governance to have an input on these parameters, but this doesn’t seem to be the case.

I also agree with @fig that even though one of the goals seems to be incentivizing borrowing, the proposed distribution doesn’t seem to match this intention. For example, why would AMPL have a 100% weight towards lenders?

Hi @pakim249

I will be updating the incentive distribution to capture the feedback from the discussion here. Let me respond to some of the points raised and we can continue the dicussion.

  • GUSD removal was more around the lack of growth. This was also flagged back in August, link here.
  • UNI and CRV were removed because they are old high vol listings and the incentives favour more recent high vol assets.
  • BUSD is currently not incentivised. It was meant to be included, there is a mistake in my table.
  • I did not include incentives for borrowing high vol assets. Conceptually I think of high vol and low assets differently, low vol assets have borrowing incentives, high vol assets do not.

To date, the following amendments are being considered:

  • Reducing the allocation to ETH, freeing up stkAAVE to be allocated across the other assets.
  • Include BUSD incentives in the table.

Please do provide feedback. I shared this post with the intention of discussing the topic and including the feedback. :slight_smile:

2 Likes

If we are to reduce the allocation to ETH, wouldn’t it make sense to apply this same logic to WBTC? They are essentially used in AAVE for the same purposes. In general, I think assets that are used largely as collateral for borrowing don’t need a large incentive, and it would make sense to reduce the multiplier equally for any asset that is simply used as collateral and has been incentivized for a while.

am i reading right that the proposal aims to allocate 36% of the total to ETH suppliers?

if yes, this would greatly reduce the efficiency of the LM program. What aave needs to grow the most are stablecoins. Incentivizing the supply side of a collateral asset would essentially mean giving away AAVE for free.
I agree that if the emission is further reduced, the incentives should be skewed towards borrowing as this optimizes interest rates for borrowers and suppliers while increasing the average protocol income. But out of 1540 stkAAVE/day, i would allocate maybe 95%+ to stablecoins. That’s where the borrowing demand is.

3 Likes

Hi @Emilio

We flagged the ETH yield and the wBTC yield as being too high. These need to be scaled back.
Thank you for suggesting +95% to stable coins that is a great place to start the conversation.

1 Like

great @MatthewGraham - i would also considering giving rewards to CRV, greatly skewed towards CRVborrowing. Acquiring CRV might be a great way to gain governance power in Curve which could help redirecting more stablecoins toward aave. Might also help with future treasury management strategies where the CRV could be used to eventually boost the aave treasury CRV rewards.

1 Like

LM program is an efficient but expansive growth tool.

The slow decay of the LM programs allows for more long-term sustainability and is a welcomed proposal.

I join the other comments on the protocol’s need for stablecoins being the protocol’s main asset borrowed and main source of fees collected for the protocol.

Maybe this quarter is also the opportunity to open the Bonding debate as an additional source of treasury collection for the protocol, Aave protocol being decentralized a treasury long term reaching a critical mass would generate enough interest to support grants, risk, and dev DAOs providing long term sustainability even when the ecosystem reserve will eventually run out of funds.

Slow down of LM program might create leeway for the small start of a bonding program.
If bonding has some community support another thread specific can be created.

1 Like

Nice work as usual @MatthewGraham . Some points from my side:

  • Is it possible to have a graphic showing the evolution of capital (in terms of asset-units instead of dollar) during the duration of the previous (current) 3 months of liquidity mining? It’s pretty important to understand how is the growth, specially on the new assets added to the program during this cycle.
  • Even if I agree that FEI is definitely a good candidate to start incentivizing a bit (close community and good growth), there is already an incentive program on their side with TRIBE tokens, for the borrowing side, so first that should be evaluated on how it affects.
  • I would say that BAL incentives could be a bit higher, but of course reduced of the current allocation, as the growth hasn’t been high.
  • I don’t think BUSD should be incentivized at the moment. Since the asset listing, the growth is really small, so in my opinion any incentivization will just create completely artificial growth (if anything).
  • I would keep AMPL out of the program for now. The dynamics on that reserve are pretty complex, so it is even difficult to model the outcome of incentivization.
  • In general, I agree with previous points (like @Emilio 's) that maybe ETH and WBTC should be lowered a bit. Not really aggressively, but on the current proposal they are disproportional compared with USD-pegged assets.
2 Likes

I don’t think FEI’s incentive program should play any role in whether AAVE incentivizes an asset or not. If incentives are lowered because of another protocol’s incentive program, that would send a negative signal to any protocol that would want to help incentivize lending and borrowing on AAVE. If anything, we should be giving more incentives to other protocols that are willing to put their own skin in the game to help the AAVE market thrive.

I agree with this. Though the fees generated are high on this market, I’m not entirely sure that this market is actually even good for AAVE considering the fact that every time there’s a positive rebase, a liquidity crisis happens.

I share this view. If we want to grow TVL, co-incentives help Aave achieve this. Co-incentives are relationship building and the proposal above favours those communities that have worked with Aave, ie: Balancer and Chainlink.

Noted and I will incorporate this when I update the rewards distribution table. :)

1 Like

Would like to see greater allocation to the ampl market.

Ampl is currently providing more revenue than Aave’s largest stablecoin market (USDC), despite having a market that is 200x smaller. It is in the best interests of the Aave protocol to incentivize Ampl lenders so that the market may grow, and revenue can increase further.

I propose that ampl LM allocation % be increased to 5%. While this may be quite high considering the market size, it would help increase the size of the ampl market on aave, and dramatically increase protocol revenue.

2 Likes

@MatthewGraham i think we should wrap this up asap to have a proposal ready to be snapshotted and proposed onchain within this week - otherwise we might go long on the actual onchain proposal

1 Like

Hi @Emilio

I have update the incentive distribution to capture the feedback from the discussion here. Amendments include:

  • ETH and wBTC reduced allocation by 80%
  • Non ETH & wBTC high vol assets reduced by 25% to allocate more towards stables
  • BUSD was included
  • AMPL was removed
  • Updated the Value Deposited values as of 12th November

The overall allocation split is around 80/20 Low/High vol assets. The main high vol assets being ETH & wBTC.

I expect the ETH yield to be marginally higher than it is now, 0.32%, but there will be no borrowing incentives. The wBTC yield falls a fair bit though from around 0.7% to 0.2%.

Overall, ETH & wBTC allocations needed to fall heavily to allocate more aggressively towards stables. The 95% mentioned above was not attained, that would mean

The newly proposed liquidity mining rewards are shown below. I will create a snapshot for next week, most likely starting Wednesday 17th November. I will use the new table in the snapshot.

Please do provide feedback :slight_smile:

@MatthewGraham thank you for your effort

a couple of points;

  • FEI cannot currently be rewarded as there are TRIBE rewards in progress; Aave V2 does not support multi token rewards - check here for more details https://twitter.com/The3D_/status/1458828361296400395

  • I would put, if the community agrees, some incentives on CRV borrowing; there is good demand for borrowing CRV right now and the Aave protocol should be interested in acquiring more CRV;

1 Like

Full support to add incentives on FEI, as their participation on the ecosystem is considerable, not only liquidity-wise, but with big involvement on technical topics too. My point was mainly in the line of @Emilio 's , mainly a technical consideration. If there is no problem, +1 to incentivise.

@MatthewGraham is there is any rationale for the incentives on BUSD ? Being a stable-asset with long maturity, from my perspective the volume on Aave is really small at the moment, considering its +$13b market cap.

BUSD has seen limited usage on Aave. However, this is an opportunity to catalyze liquidity and utilization.

BUSD is a top 5 stablecoin with a ~$13 billion marketcap and $5 billion+ in 24-hour volume.

@Emilio - The above quote is definitely true. We were working to create an AIP that enable dual incentives, but have since pivoted away from this. As a result, stkAAVE incentives on FEI have been removed.

@pakim249 - Yes sir. @Anjan-ParaFi has articulated this in his comment. Do note, it is 0.09% of incentives.

After fielding several requests for CRV, I have added applied lending incentives for CRV.

The newly revised incentive distribution is shown below.

Screenshot 2021-11-16 at 21.28.26

I also updated the ARC to include gas cost reimbursement.