Initial Discussion 2: Origination Fees

[defifrog]

Hello Aave world,

Intro

I write to you today in hopes to initiate the discussion around Aave’s revenue model. The goal of this thread is to discuss how we might set the framework for effectively:

1. Setting the fees generated by the Aave protocol (origination fee)
2. Utilizing the revenues generated by the fees (incentives for dev fund, LP rewards, etc.)

Aave Protocol Fees

For the sake of aligning all people to the current state, here is a direct quote from the Aave documents:

@Aave Referral Program

Aave is currently kickstarting governance. Origination Fees will be decided by the Aave community and are currently “paused” with a minimal low amount. You can still register for the referral program.

• Each loan currently incurs a 0.0000001% origination fee.
• 80% of this fee is used to buy LEND tokens from the market and burn them, which reduces the monetary supply of the LEND token.
• 20% of this fee is redistributed to the referrers.
• The referrers consist of the liquidity providers ( deposit() ) and the borrow providers ( borrow() ).
• 40% of this fee goes to the liquidity providers (effectively 8% of origination fee).
• 60% of this fee goes to the borrow providers (effectively 12% of origination fee).

Source: https://docs.aave.com/developers/integrating-aave/referral-program

From the official documentation above, it appears this is now our role as a community to put forth our best opinions on how to set fees. If I’m not mistaken, this initial 0.0000001% origination fee was set extremely low in order to bootstrap adoption. However, with $1.36B total locked value, as the 3rd highest DeFi protocol by value locked, and with the launch of $Aave Token and Governance, it seems fitting to kick off discussions now.

Pricing Strategy

The way I’d approach pricing strategy/valuations is 3-fold:

1. Cost Approach: What is the upkeep cost / maintenance for the Aave team? Do we have an understanding of their annual budgeting? How big of a team and how many years must they exist? The result of this annual budget costing analysis should be the revenue floor.
2. Value Approach: How do our users value our protocol? What exactly is our value proposition? I’d like to confirm my hypothesis with data, but I’d bet that most of our volume is based off users that want to keep exposure to an underlying asset and borrow stable-coins or wBTC. (personally, hoping for renBTC in the future ). Once these are borrowed, what do the users earn with them? Our fees must be low enough for the users to still be profitable after!
3. Market Approach: This is where I’ll some extra help and hope to get people’s thoughts. This is where we decide how we compete. The only thing I could find for MakerDAO was the 3.5% stability fee. Couldn’t find fee data for Compound. Aside from DeFi, with TradFi, the origination fee is between 0.5% and 1%.

Outstanding thoughts for me here:

• If anybody has Compound fee data, please share
• How much longer do we want to bootstrap adoption with low fee?
• What can we do to further adoption of Aave?
  • I’d suggest a native BTC-onramp integration where a user can take BTC, mint ERC-20 BTC, deposit that into Aave, all in one transaction. RenJS is an SDK that enables this. You can see Curve as an example)
• What does the Aave team think?
  • Don’t want to step on any toes here and would like their guidance as they likely have material, non-public information
• Finally, once we have an estimate based on these 3 approaches I outlined, we can triangulate an appropriate fee (maybe by a vote based on weighted average by vote?)

Usage of Revenues

I’ll begin this section with a comment from Stani in another thread:

@stani

I agree with @depressedape and @EmmanuelD that there should be incentives for LPs and stakers for providing safety for LPs, DeFi is about incentives in general. It’s more of a question on

1. the healthy balance between stakers and LPs
2. ensuring that there is incentive for LPs to stake their earned AAVE
3. ensure that LPs are interested in long-term sustainability for the protocol (that would be inline with the governance since short-term LPs would merely want faster gains)

I 100% agree with @stani that we must align the incentives with a healthy balance.

Here’s my breakdown:

1. $Aave stakers should receive a lion’s share of the rewards as their collateral supports the stability and security of the protocol. Currently, they (we) receive 0% of origination fees; the 400 Aave/day was set by the community in a prior thread, will end after the set period, and does not come from fees. This must change to incentivizing maintaining a threshold of Aave staked.
2. I believe liquidity providers should receive the next largest amount. Currently, they receive 8% of origination fees.
3. I’m not sure what “Borrow providers” are. This term is outside my applicable understanding of traditional finance. Currently, they receive 12% of origination fees. If this term means “Borrower”, I’d strongly recommend we don’t incentivize borrowers more than stakers or depositors. I’d actually suggest this becomes 0%. The borrower is the user with the most benefit, and it should be safe to assume they’ll be using their loan to make more money.
4. Last but certainly not least, we should discuss how the dev fund should be properly funded to continue the innovation that the Aave team ships

I hope this helps kick off the discussion, and I hope to hear what everybody thinks. Once we have come to a decent alignment on structure/framework, we can determine what to push forward as an official proposal. Thank you for taking the time to review my request for discussion.

Previous discussions/important links:

1. Initial discussion: AAVE reserve emission for safety and ecosystem incentives
2. Thank you for providing me the source for Aave fees @PabloCandela

[Zer0dot]

Great time to start the discussion!

I definitely think we’re building one of the most incredible innovations in finance here.

On the note of origination fees, don’t quote me but I believe Compound charges no origination fees of any kind- and I think that’s not a bad idea.

First off, should we have a noticeable origination fee, we’d be locked into “second choice” territory in regards to Compound assuming similar rates. This is a big deal, because though Aave and Compound can (and should) be in a symbiotic relationship, Aave must cement its place in the standard borrowing niche.

Another point I’d like to make is that staking in the security module provides safety for liquidity providers, not borrowers. Borrowers already pay liquidity providers through interest, I fundamentally believe the course of action to take would be to redirect a small portion of this interest generated towards stakers.

In other words, Aave remains competitive, maintains its edge in having multiple markets (see and signal your vote on my UNI-V2 liquidity token market proposal here) and simultaneously rewards those who bear the heaviest risk.

It is definitely worth noting that the security module is seriously one of Aave’s biggest selling points. Especially as more traditional finance users get onboarded on to Aave. On the broader financial spectrum, receiving more than the average 0.05% APY on what’s essentially US dollars with in-built insurance against black swans is aavesome!

In Summary

For the following two reasons, I am leaning against implementing a loan origination fee in favor of an interest-redirection model:

1. Compound, the biggest lending protocol in DeFi charges no origination fee (seemingly).
2. The safety module benefits depositing more than borrowing.

Side note, regardless of the origination fee, governance can and should enact a Dev fund and a “treasury.” I totally agree with you there.

Let me know what you think about my input. I’m not a financial expert (yet), but I just wanted to voice my opinions. Thanks again for starting this discussion!

P.S. On flash loans, I’m not sure if it’s a good idea to redirect a portion of the fee towards safety module stakers, since flash loans are essentially (almost) risk-free and only require liquidity to function. Could be a good idea to think about this too. Flash loan fees are one of Aave’s most innovative value propositions, and their fees are an important incentive for liquidity providers

[defifrog]

@Zer0dot

Borrowers already pay liquidity providers through interest, I fundamentally believe the course of action to take would be to redirect a small portion of this interest generated towards stakers.

Thanks for the response, @Zer0dot! - I agree. Aave is pretty awesome.

Interesting idea. Whether we choose an origination fee % or cut a % from stakers, the fee will effectively come from the same pocket. (Please correct me if I’m wrong). On the Aave application, a borrower is a staker no matter what, right? They have to provide tokens to borrow other tokens.

I can’t speak to the implementation complexity from a technical standpoint, but from a business standpoint, we basically just have to consider which revenue stream we want to open and determine how we’d want to position ourselves in the market and brand it.

Comparing the Two

Origination Fee ($) = Principal * Origination fee (%)

• Pro: Instant fee on each loan that occurs
• Pro: More occurrences of fees than taking a continuous cut of staking interest (Each time you go in and out of a loan, it would generate a fee as opposed to the just one continuous fee on one principal)
• Cons: Compound has no origination fees
• Cons: Fee is not Continuous

Fee from the staked tokens = Staked Principal * (e)^(% interest*T)… assuming continuous compounding… New variable being T=time:

• Pro: A larger principal because staked value will always be > borrowed value. This does not necessarily more fee revenue because the fee is over time with no fixed cost
• Pro: Revenues could become a bit more continuous/stable and a function of locked value rather than transaction volume
• Pro: Attract more borrowing by branding with no origination fee loans
• Cons: Fee is not Continuous
• Thought: We’d basically play market maker and win within the spread

I like your ideas, but I’m open to hearing if the overall sentiment is to keep origination fees at 0. I have 2 additional thoughts now:

1. I haven’t read any docs on the fee structure for flash loans. I’d be interested to add that to the above comparison
2. I also just remembered the stable APY fee that can be opted-in for. Does anybody know where that additional revenue goes?

@Zer0dot

The safety module benefits depositing more than borrowing.

Doesn’t it benefit both people, though? Because the depositor/borrower is the same person? (Outside of flash loan)

[ecent]

Ideally the compensation for staking goes up but it does’t necessarily have to come from origination fees.

As @Zer0dot correctly noted we want to compare favourable to other platforms and while there are many factors to compare, fees are certainly a significant one.

Adding a reward to staking proportionally to the amount of time a stake is in place might solve that without touching origination fees. The longer you stake, the more you receive above the base yield. In theory would incentivise keeping the stake in place, doesn’t touch the fees associated with borrowing from the platform and could help in addressing some of the current complaints amongst holders about the staking structure.

The incentive could come either in the form of AAVE or as a bigger portion of the eventual BAL / ETH rewards.

[Dydymoon]

Hey everyone, really good discussions !

I totally agree with you both on a needed tresaury, yearn did it and it worked pretty well, and i think we could do the same, somewhere like 500k - 1M.
This tresaury would be used for maintenance of the protocol, or help in case of a short fall event.

However, i think we can add an origination fees, even if compound doesn’t, i mean, few people are still using compound, and even less when V2 will be there …
Moreover, the referral program could be stopped (considering this is for marketing and everyone knows AAVE right now), in order to leave the DAO with 100% of the fees.

Regarding the fees, there is flash loans fees, fees on the borrow for the different markets (Aave, UniV1) and future potentials markets such as (TokeSet, Realt, Univ2, Aavegotchi and more …)
I don’t know how to define correct rate for each market in order to do some calculations, so if someone can help on this, would be amazing !

In my opinion, Aave has the ability to stay attractive and competitive despite a raise of the fees, because the available markets on aave are and will only be offered by aave.

So, to get the tresaury filled, instead of using part of the ecosystem reserve, we could use fees from borrows and flash loans. This money would be swapped to stablecoins and start compounding interests, and fees would be added until the decided amount is reached.

Once it’s good, this could create two additional revenues for stakers :

• the tresaury would always generate interests, and some % could be distributed to stakers ( let’s say 50%) The other half would keep compounding to increase tresaury and interests for stakers
• All the fees generated could be paid in the currency borrowed, and automatically used to buy AAVE tokens on the market and distribute it to stakers

This would increase the scarcity for aave token, and leave the ecosystem reserve fully available for the DAO, especially considering this reserve will not last forever, and so, we bet on the fact that revenues that comes from the protocol will be higher than incentive distributed right now.
So, in order to play safe, this reserve should last a couple years.

I did a quick estimation of how could be distributed this reserve over the next 6 years (could be less if the protocole generate enough revenues by then)
This takes into account rewards for stakers, rewards for LP, and a compounding bonus that could incentivise people to compound aave received and stake it all into the Safety module ( that means more security for the safety module, less dump of the rewards for the price and less aave available on the markets, so higher price)

You may ask me why stakers get more than LP ? Because stakers are taking a bigger risk by protecting the protocole ( slashing ) compared to LP.
Moreover, the liquidity pool is a 80/20, so not much impermanent loss, and even if there is some, that should easily be covered by AAVE rewards + BAL rewards + fees anyway. Lp could also get compounding bonus if they stake the AAVE rewards.

This example of distribution is decreasing with time, and would be with a total of 2.5M AAVE, leaving 0.5M aave free.

For those 0.5M AAVE remaining, we got some ideas with @EmmanuelD :

• Universal airdrop like UNI to recompense users that used aave before migration or else.
• Airdrop for stakers
• Burn total
• Additional insurance in case of short fall event

So to resume and answer @defifrog on some points, i believe we wont need to keep boostrap adoption with low fees once there is V2, especially if L2 available.
The protocol would get a compounding tresaury, LP would get AAVE+ BAL+ Pool fees, Stakers would get AAVE + Fees from the protocol (once tresaury is complete) in form of AAVE + Part of interets generated by the tresaury.

I also like the idea of RenJs implementation a lot.

I hope this can help even if there is more than fees in this proposition made with @EmmanuelD . Thank you for taking time to review it and it can’t wait to get your opinions on these points.

[drew]

Interesting points on other opportunities to direct fees toward AAVE stakers. I’m curious if an official proposal has been put forth on where fees from Flash Loans will be directed? Are they currently being used somewhere?

I’d also advocate for finding alternative ways to collect revenue for stakers rather than through an origination fee. Though Aave has been quite successfully up till now there will be increasing competitive pressure from e.g. Compound, even if they are lagging behind now. There should be as little friction as possible from people wanting to borrow and deposit funds on the platform.

[Zer0dot]

Thanks for your thought-out reply @defifrog! Here’s what I’ve got to say:

@defifrog

Pro: More occurrences of fees than taking a continuous cut of staking interest (Each time you go in and out of a loan, it would generate a fee as opposed to the just one continuous fee on one principal)

I disagree with this being a benefit of implementing an origination fee. I don’t think there’s a material difference between having a fee be continuous or split, disregarding returns.

As for this pro when it comes to interest redirection:

@defifrog

Pro: Revenues could become a bit more continuous/stable and a function of locked value rather than transaction volume

I do want to point out that interest earned is proportional to the amount borrowed, thus, revenues are tied to the borrowed amount. Though this doesn’t have to do with transaction volume necessarily, as more gets borrowed, more fees can get collected in both models.

@defifrog

@Zer0dot

The safety module benefits depositing more than borrowing.

Doesn’t it benefit both people, though? Because the depositor/borrower is the same person? (Outside of flash loan)

Absolutely! A borrower is always a depositor, but a depositor is not necessarily a borrower. My thought process is angled at not making depositor-borrowers pay for the insurance of pure depositors. You know what I mean? In other words, by redirecting a portion of all interest, we’re only taking from the insured side of the equation.

Overall, I think we’re on the same page! There’s a lot of discussion to be had, good things take time after all.

@drew

Interesting points on other opportunities to direct fees toward AAVE stakers. I’m curious if an official proposal has been put forth on where fees from Flash Loans will be directed? Are they currently being used somewhere?

Hey Drew! Welcome to the governance forum!

To answer your question, as it stands right now the flash loan fees are 100% directed to depositors in the appropriate currency.

[stani]

Thanks @defifrog for opening the discussion on fees and also for others for contributing. Good insights from @Zer0dot and @Dydymoon.

I think the main treasury currently is the Aave Reserve that holds 3 mm AAVE tokens that can be used to incentivise various actions, such as there is now the 400 AAVE per day distribution to boostrap the Safety Module (currently there is around 2 700 stakers with 1,3 MM AAVE, quite nice in terms of decentralization: https://etherscan.io/address/0x4da27a545c0c5b758a6ba100e3a049001de870f5). The Safety Module indeed is a really good advantage within the Aave Protocol as it allows risk transfer from depositors to protocol governance and thus helps adoption.

Fees are something that could be added pretty much anytime if the governance decides so, for example could be during or right after the v2 launch that is coming quite soon. In terms of the fees, I think the fees should not decrease the ability to move capital since it would reduce the amount of innovation that other devs would build on top of the Aave Protocol. Instead, the fees should encourage traction such as changing loan position by swapping collateral, loan currency, interest rate and so on.

The issue with origination fee is that once you pay it, you would not want to change your position again since you would have to pay it once more. Instead if a borrower could freely change position on weekly/monthly basis depending on the market rates that would create more monetization events and allows to move capital freely.

The v2 has quite many ways to collect fees, for example part of the interest could go to stakers, each collateral swap, loan currency and interest rate swap or part of the liquidations. Even if you close your loan position by paying with your collateral that could be a place for monetization for protocol fees.

Fees going to mainly for stakers makes sense since the stakers are the ones that are taking the active risk of the protocol and LPs are receiving interest rates, but also could in the future receive Ecosystem Incentives (EI) for providing liquidity in form of AAVE. LPs should be encouraged to stake the AAVE into SM so that both stakers and LPs incentives are aligned towards safety.

The dev fund is an interesting idea especially to get more community contributors involved building the protocol and new features on top. Additionally, from that fund or separate fund there could be a decentralized grants program as well. The main treasury can be the Aave Reserve that collects all fees and allocates to places however the governance decides (stakers, dev fund, grant fund etc). It’s all up to the governance and also additional ways to embrace adoption via fee distribution is highly encouraged. Quite often the focus has been on devs building. What about non-devs and also community members who invent new features or bring new ideas in front of the community? How could the community reward everyone that brings special effort? Some thoughts.