ARC: Reduce flash loan fees, introduce a fee for 1-tx borrows and add a flash loans whitelist

ARC Summary:

Reduce the Aave v2 flash loan fees to 0.02% (with 0.01% going to lenders and 0.01% to the protocol), introduce the same 0.02% fee to borrowings that are paid back within the same transaction (a workaround that many use to avoid the flash loan fees) and add a no fees whitelist for flash loan users.

ARC Rationale:

The flash loans market is growing more saturated with sources for flash loaning or flash minting with lower fees - dYdX has no flash loan fees, WETH10 has free flash minting, Maker has voted on a 0.05% starting fee for their flash mint module. With the market gradually trending towards lower and 0% fees, the Aave protocol is currently lagging behind and losing its competitiveness.

DeFi Saver is the largest Aave flash loans user and we have gladly used Aave flash loans by default for all FL-powered features rolled out since early 2020 because of the number of available assets and great available liquidity. However, with the introduction of our Recipe Creator that includes both Aave and dYdX FL integrations, we can see first hand that many users will gladly take the time re-create leverage management transactions using dYdX flash loans to avoid the 0.09% fee, even though leverage management options are available as single click options in our protocol specific dashboards.

In order for Aave to stay competitive in this growing market and maintain its position as a major flash loan provider, we would like to propose three changes:

  1. Reduce flash loan fees from 0.09% to 0.02% with 50% going to liquidity providers and 50% to the protocol.
  2. Introduce a 0.02% fee for all standard borrowings that are made and paid back within the same transaction, a process that is currently being used as a way to avoid the flash loan fees.
  3. Introduce a flash loan users whitelist controlled by Aave Governance where integrators could be voted in to utilize flash loans without any fees as a reward for promoting Aave and introducing other benefits to the Aave protocol and its users.

This change could have an additional positive impact by pushing any integrators to switch to using Aave v2 flash loans from v1, as we’ve seen that ~40% flash loan volume is still done on Aave v1.

If you would consider supporting this ARC and the potential proposal, please consider delegating proposition power to 0x446aD06C447b26D129C131E893f48b3a518a63c7 (DeFi Saver Deployer account).

Looking forward to hearing more thoughts from the Aave community.

6 Likes

At the weekly report * I have read that the utilization was growing by 273%, so if it’s growing why reducing the income?

If someone has data of competitors and we recognize that it stops growing we can reduce the fees, I see no reason why

1 Like

I 100% agree with this proposal. In v2 currently the protocol isn’t earning anything from flashloans (the whole 9bps fee goes to liquidity providers). In an environment where multiple entities are coming with very cheap or free flashloans, there must be the possibility for integrators to keep accessing aave liquidity as long as they provide real value to the ecosystem. Defi Saver is for sure one of these integrators, always with Aave from day 1 and providing great accessibility for the Aave protocol.
Lowering the fee will greatly increase flashloan volume, which will compensate the fee reduction.

4 Likes

This is a great proposal! Flashloans are a great tool that make the market more efficient.
With other protocols offering flashloans too and flashmint slowly being implemented, fees will get competitive. In Refinancing operations compounding 0.09% to debt, in one transaction is a significant cost.

1 Like

Very good proposal !

I really like the idea of a whitelist that allows free flash loans for integrators.

4 Likes

I was really hoping the thread would catch more eyes and cause more discussion with time or at least bring in more replies on whether most agree or not, but so far it seems it hasn’t caught nearly enough attention it deserves - so let me please try to kick-start this talk once again.

Let me start with replies to comments posted so far:

Looking at flash loans volume at a weekly level isn’t very relevant, imho. Major flash loan volume happens in bursts, accompanying highly volatile market times (leverage adjustment, self-liquidations and liquidations), major farm rotations (refinancing) or major exploits (flash loans being applied to abuse open vulnerabilities).

This chart from AaveWatch shows this pretty well:
image

Both great points. Thanks for joining the discussion, Emilio!

I very much agree this should be the goal to strive for here, yes.

Exactly - this specifically is a HUGE issue right now. For example, we had a humongous Maker Vault owner reach out to us to ask about using our tools for refinancing. The protocol fee would’ve been multiple hundreds of thousands in their case, as the flash loaned amount would be their full debt amount. They of course decided against moving forward with the process. This is needlessly lost business for the Aave protocol, in my opinion.

This specific point is something I would assume will cause most controversy, so we would agree to have it left out if the rest of the proposed changes depend on it. However, if the community agrees, we’d love to see it included. It would allow us and other similar integrators to utilize Aave flash loans without any second thoughts, exclusively, wherever applicable.

For anyone seeing this thread - please post a comment sharing whether you agree or disagree and join the discussion.

1 Like

I agree with everything but the whitelist. We don’t REALLY want to give preferential treatment to certain users in the protocol, do we? We have other ways to compensate value added. The protocol should not discriminate on fees whatsoever.

1 Like

Appreciate the response, pakim.

The whitelist is definitely not something we’d consider required for proceeding with the proposal.

It was just one idea for a way to drive certain trusted integrators to stick to Aave flash loans no matter what. For example, even with the fees reduced to 0.02% from 0.09%, there’s no reason why we wouldn’t allow our users to have dYdX flash loans utilised by default where applicable (e.g. for ETH, DAI and USDC), so in this sense the whitelist would allow Aave flash loans volume to remain higher or at least more stable, due to constant use within certain apps, which could be an important metric for the protocol.

I can understand your point about the protocol not discriminating against anyone when it comes to fees, though perhaps also consider that currently the same fee is paid to the protocol by a person using Aave flash loans to perform an exploit and by a different person using Aave flash loans to better manage their leveraged positions in DeFi Saver or similar apps.

Again, just to highlight this once more, the whitelist is definitely not something we’d want to insist on including in the proposal if the Aave community disagrees.

By the way, if convenient, please see my post that now showed up above yours, as it was previously temporarily hidden by a spam filter.

1 Like

Thanks for the the thoughtful reply.

If we are concerned about flash loan volume, then can I suggest a more radical solution? Why not simply lower the flash loan fee to zero? According to the weekly governance reports, flash loan fees make up roughly about 1% of Aave’s revenue, which is many times lower than it used to be. As the ecosystem matures and offerings are diversified, this revenue will only decrease. This is not a source of revenue that Aave should be concerned about. However, Aave does have a chance to remain the market leader in flash loans and strengthen its image by simply removing this flash loan fee. Oh, and of course we wouldn’t need a whitelist if flash loans were simply free for everyone!

EDIT: Decided I have more to say. I also disagree with the 0.02% fee you propose for flash borrowers. The fact that we can do everything in one transaction from smart wallet applications is a modern miracle that creates a more even playing field for ordinary users. The entire ecosystem is becoming vastly more efficient, as can be evidenced by Uniswap and Curve’s protocol improvements. In my opinion, Aave must keep up with this efficiency, if not strive to be a market leader in that efficiency.

1 Like

Thank you Nikola for putting in the time and effort to keep/make the Aave flash loans competitive.

I do agree with the Pakim, why not reduce the fees to 0 where users/protocols have (almost) free alternatives? If the loan fee is reduced to 0.02%, it would still make more (financial) sense for a protocol to use dYdX.

However, as far as I know dYdX only supports ETH, USDC & DAI. So perhaps this is where Aave can charge fees, as it offers a unique service. So perhaps charge 0.02% (or more) on those assets that other protocols don’t support? This would make it a no-brainer for protocols & users to use Aave for flash loans.

1 Like

Flashloans are a technology that I find incredible. However, I am disappointed that this ARC did not go to a governance vote and remained at the discussion stage. So I have to re-launch this discussion.

I looked at the weekly reports for the last 3 weeks (at the time of publication W27, W28, W29. Available here: Aave Weekly: Protocol Performance & Governance Update ). The flashloans have generated an average of $102,636 over the last 3. Knowing that the protocol revenue averaged $11,300,127 over the last 3 weeks, after a quick calculation we realize that flashloans represent about 0.91% of the protocol revenue.
This is a totally negligible revenue in my opinion.

So I agree with @pakim249 and @tobias , we should reduce the flashloans fees to 0% to stay competitive against Dydx, Cream or Unilend, whether on Ethereum, Polygon, and soon Arbitrum. However I don’t agree with this whitelist logic, and a 0% fee drop would settle this disagreement.

AAVE being currently the largest DEFI protocol by TVL, this fee reduction would allow any user to borrow for free (for the time of a transaction only) the highest possible amount of tokens (within the limit of the number of tokens available for borrowing), any protocol, and with a diversity of available tokens.

I’m going to digress slightly from the basic topic but I think this is important too.

We also know that the AAVE Pro Market (now its official name is AAVE Arc) should be available soon. In a logic of allowing all players to take advantage of the DEFI, it would be interesting to allow players present on this market to make flashloans. However, I am still not sure if we should lower the flashloans fees to 0% on this market from the beginning. As this market is regulated, we would also have to check if this would be possible from a legal point of view.

2 Likes

It doesn’t seem like there’s much resistance to the idea of removing flash loan fees. I’ll create a snapshot vote and then we can see if we can push it into an AIP.

3 Likes

I agree with @pakim249

Ultimately, there is no reason that flashloans should have a fee attached, flashloans bring no risk or inconvenience to the lenders, so why should borrowers have to pay for it? Especially when other protocols offer the same service free of charge.

One of the key strengths of defi is it’s removal of rent seeking middlemen, removing the fee for flashloans would be in line with these values.

1 Like

I tend to disagree with this view. There is still risk involved when capital leaves your contract, although it so small that is negligible. Despite that, people using liquidity for a profit should pay the liquidity providers. The fee can be reduced for sure, and we can add to the protocol a way to whitelist certain entities to access free flashloans (like @nikola_j proposed) but i wouldn’t make them free. Free flashloans also further facilitate exploits against vulnerable protocols.