Proposal: Aave Interest Rate Uber Model v0.1

Governance General
1

Hey Aave community! Thank you for opening.

I’m proposing a new interest rate model for Aave.

Aave Interest Rate Uber Model v0.1

Aave current interest rate model & problems:

1. Current interest rate model is relying too much on utilization of token pool:

    1. If the utilization rate of token pool goes down by more users returning their borrowed amount at the same time or more collateral deposited, and there will be super low rate(only regular amount of users will borrow, borrowing amount won’t change as much)
    1. If the utilization rate of token pool goes up by LPs withdrawing their liquidity, or more loans taken and there will be super high rate(people won’t borrow, and more borrowers will reterne their fund)
    1. So on

2. Stable coins and Gov tokens are using same interest rate model:

    1. Compared to stable coins Gov tokens use cases, borrowing volume, and borrowing intentions are different , so most of the time Gov tokens earns super low interest rate, and some time super high interest rate for short period of time(when there is gov voting events of that token)
    1. Because Gov tokens pools has low liquidity, so most of the time there will be super low interest rate, and super high interest rate for short period of time.

3. Aave and other token distribution problems:

    1. Currently Gov tokens are distributed to both side, and getting distributed to most of the token pools, why?
      1. Most of the token distribution are useless, whether distribute tokens or no, doesn’t make much differences
    1. Aave has to save those tokens for now, and distribute it to only necessary, and productive ways
      1. For example Stable coins are the root of Aave, in the future if stable coin pools makes lower rate, then distribute to Stable coins pools to attract more stable coin liquidity to Aave

4. Unproductive tokens are sharing profit with productive tokens

    1. Based on current model there won’t happen 100% utilization(or higher utilizaiton for long time)
      1. Because after utilization rate goes above 85%, the interest rate goes up drastically that borrower doesn’t want to borrow, and borrowers will return fund
      2. So there is always will be unproductive tokens sharing profit with productive tokens
    1. there will always will be big spread differences between the deposit rate and borrow rate

5. Different treatment for different LPs and borrowers(users):

    1. Based on current AAVE interest rate model users are equal
    1. There is no better treatment for attract larger LPs to inspire deposit large amount and deposit longer time or attract large Borrowers to take larger loans for longer time, and so on

6. There is no voice contribution from aTokens holders

    1. aToken holders can’t say anything about the Interest rate
    1. There is no connection between aToken holders and Aave Governance

7. Interest rate can’t be changed by voting

    1. Different tokens has different needs on different times, so there has to be different rates for different tokens at different time
    1. Currently every tokens has same interest rate model that can’t be changed

8. Stable rate:

    1. It makes complicated for average uses to understand
    1. Not many people use it(useless)

What is the Aave Interest Rate Uber Model?

  1. It’s a Interest rate model that inspired by Uber charging model
  2. It’s a interest rate model that can help Aave be more robust, and more productive
  3. By using changeable and predictable interest rate model(Aave interest Rate Uber Model), make Aave more flexible and compatible

How does Aave Interest Rate Uber Model going to Solve the problems?

1. Stable coin & Gas token Interest rate changes:

    1. By Changing to Aave Interest Rate Uber Model

      1. Interest Rate will more stable that won’t relies too much on utilization
      2. interest rate will be more predictable that won’t get too high, so borrowers can comfortable with taking loans without worrying about the rate gets too high
      3. Interest rate will be more acceptable that interest rate won’t get too low or too high that LPs and borrower both will be happy
      4. Will be possible to rich higher utilization rate that interest rate won’t gets too high even the pool rickes higher utilization rate, so borrowers will welling to take loans even the utilization rate is high
      5. Treat users differently that attract more people to use Aave
      6. Guarantees the LPs with base rate yield amount to attract more stable coin liquidity

2. Gov token Interest Rate Changes:

    1. By Changing to Aave Interest Rate Uber Model:
      1. Gov tokens will earn higher yield(deposit rate)

3. Gov Token distribution & Base rate insurance fund:

    1. Stop all the token distribution, and save it. In the future distribute to only Stable coin pools, when ever the any stable coins earning rate gets below the bases rate, then use that tokens to incentivize that pools to give guaranteed base rate to LPs
    1. Token distribution can’t be continued forever:
      1. Build a base rate insurance fund pool that will takes 0.1%(or X%) of all the fees on the platform(from pools that earning above the base rate), and use that funds to distribute to stable coin pools that earning below the base rate
        1. if any stable coin pools earning rate gets below the base rate, and Gov token distribution funds are already distributed, and then use the base rate insurance fund pool to incentivize the pools to make them earn base rate at minimum at any given time

  1. Productive token pool and Lazy token pool:(it’s optional, just suggestion, not necessary)

    It’s not done yet, I’m working on it.

    It’s kind works like Uniswap v3 Concentrated liquidity model(not similar, it’s kind of easy way to explain it to you guys)

    1. Nonproductive tokens will only earns base rate, and all other fees will be earned by productive tokens

      1. Create a Lazy pool for each tokens, that not utilized amount will be on Lazy pool that only earned base rate, and utilized tokens will earn as usual(higher)
      2. First come first serve that only productive tokens will earn higher yield
        1. And it also makes LPs on productive pool to deposit their tokens longer, and LPs on Lazy pool want to get in to productive pool to earn higher
      3. So on

5. VIP user treatment for LPs and borrowers:

    1. What is a VIP user?
      1. LPs that deposits above X amount and deposits longer than X days
      2. Borrowers Pays above X amount in fees to the protocol in X days
    1. VIP LPs:
      1. VIP LPs will be rewarded by 5% fees of not VIP LPs
      2. VIP LP bite pool: smaller and shorter LPs will pay 5% of their earned rewards to this VIP LP bite pool, and all the rewards will be payed to VIP LPs as extra bonus
    1. VIP borrowers:
      1. VIP borrowers will get 10% discount, if the interest rate above the discount rate
    1. Why VIP user Treatment?
      1. By incentivizing larger users(LPs & Borrowers) to make them feel special, and inspire them to use Aave even more.

6. aToken holder has to speak their opinions about interest rate of their tokens:

    1. With help of Aave Interest Rate DAO
      1. Use Forums to create discussion
      2. Use Snapshot to get votes from aToken holders to decide the base rate of their tokens
        1. As long as aToken holders aggrade rate, then there always will be liquidity no matter the rate will be, so it’s not a bad idea to atlas get their opinions about the interest rate, and propose to AAVE governance

7. Interest rate changes and updates:

  1. Based on the aToken & AAVE token voting results, we can update base rate of different tokens
  2. Whether now or one day in the future Aave eventually needs upgradable Interest rate function to make AAVE more proactive to different market(demand, Volume ) reaction

8. Canceling Stable Rate:

  1. After implementing Aave Interest Rate Uber model, no longer need for Stable rate
  2. Just cancel it

How does Aave interest Rate Uber Model work?

1. Stable coins Interest Rate Model Proposal:

    1. Base rate:

      1. Set a base rate(ex: 1.5%) that no matter the utilization of the token pool, anyone borrowers from the pool will be charged base rate at minimum
      2. We can use same base rate to every USD Stable coin or we can also use different Base rate to different stable coins

    1. Demand added rate:

      1. Set a 5% Demand added rate that will variance based on the utilization of the token pool utilization
      2. It’s similar to every stable coin that max rate of any pools equals (Base rate + 5% Demand added Rate)

    1. Discount rate:

      1. Only applicable to VIP user
      2. Set a limit that if any borrowers pays X amount or larger in interest, then the protocol will gives 10%(of fees) discount(so we can inspire users to borrow larger amount for longer time )
      3. The charging rate has to be above Discount rate to eligible to discount

    1. Token distribution & Base rate insurance fund token distribution:

      1. If any circumstances any stable coins earns below base rate, then there will be token distribution to guarantee LPs that they will always earn above the base rate
    1. More details are coming…………

2. Gov tokens interest Rate proposal:

    1. Gov tokens: AAVE, YFI, UNI, COMP, s on
      1. Gov token unlike stable coins users only borrower for short periods of time, and only borrow ones in a while, so borrowers doesn’t really cares about the interest rate amount
      2. So interest rate can be higher, and borrowers still borrow it
    1. Option 1: use Stable coin like interest rate model, and set higher base rate
    1. Option 2: set Fixed daily based interest rate that charge 0.1%/day(and minimum 0.1%(base rate))
    1. Option 3 : At normal times will lend as unusual, and at the gov proposal voting period will change 0.1% of the token value for per 24 hours

3. Gas token Interest Rate Proposal

  1. Gas tokens: BTC, ETH, AVAX, and so on
  2. It will work like Stable coin interest rate model, but with different base rate for each tokens.

4. Stable coin LP base rate guarantee:

    1. Stable coin LPs are the root of the protocol, so we have to guarantee the LPs that always will earns above the base rate
      1. Based on current borrowing rate there won’t happen earning lower than base rate, but it might happen in the future
        1. If utilization goes down grammatically, then will happen
      2. Guarantee mean is not unlimited
        1. There will be distribution limitation if there will be large amount of distribution needed
        2. If the Distribution amount is not as big, then Will be distributed immediately
    1. If the LPs earns lower than base rate, then AAVE will incentivizes with gov token to give LPs guaranteed base rate
    1. The liquidity at first comes from AAVE token Distribution program, in the future comes from Base Rate Insurance fund pool
    1. Main purpose of Stable coin base rate guarantee is to attract as much stable coin liquidity as possible, and promise them there always will be guaranteed yield, so don’t withdraw the liquidity even when low utilization happens

More details are coming:

Risks and Trade offs:

    1. Risks:
      1. Unperditibe user reactions at first implementation
      2. Not sure about the Gov token borrowers and LP reacts to new interest rate model
    1. Trade offs:
      1. I can’t think of any reasonable trade offs

Benefits:

  1. There will be more fair borrowing rate no matter the utilization rate of any pool
  2. Can reach higher utilization rate with acceptable interest rate
  3. Make higher interest on gov tokens
  4. Give LPs guaranteed Yield amount to attract more stable coin liquidities
  5. The interest rate won’t gets too low or won’t get too high so borrowers can borrow without too much hesitation
  6. Treat different users differently to inspire users to use Aave more
  7. With aToken voting(on Snapshot) & AAVE token voting(Governance Voting ), rates can be changed

Potential implementations & technicole parts:

1. Active Interest Rate Changing & time to time Interest Rate upgrade

    1. Option1: Based on current OpenZaplin proxy contract it’s a good idea to make one time Interest Rate upgrade
    1. Option 2: If Aave dev team aggress to upgrade the contract to EIP-2535 standard then Aave will have more actively changeable Interest Rate function
      1. We can change the interest rate any time we want(with gov voting), and make the Aave more competitive, and more Productive by actively changing the base rate of the different tokens based on different demand & time

2. Proxy upgradable contract

    1. Currently Aave using OpenZeppelin Proxy contract:
      1. aave-v3-core/contracts/dependencies/openzeppelin/upgradeability at master · aave/aave-v3-core · GitHub
      2. Not too flexible, difficult to upgrade, and takes too much work
      3. So on

    1. My recommendation to upgrade to EIP-2535 Standard (only recommended, but not necessary)

      1. EIP-2535: Diamonds, Multi-Facet Proxy
      2. EIP-2535 standard is better option that makes Aave smart contract better on fast changing Defi space
        1. Easy and flexible deployment to merge all the contracts super fast
        2. Add more contracts in the future that adds more functions to Aave protocol
      3. It’s good for overall Aave smart contracts
        1. Faster, and easy upgrade functions
        2. Saves on gas cost, no smart contract space limitation, and so on

3. Implementation Steps:

    1. If Aave dev team want me to write the code out, I can do it, but it might take some time
      1. I have to Fork the Aave, implement it and test it(test might take long time)
    1. If Aave dev team want to implement it:
      1. I think it’s better option, because Aave dev team know the codebase better than I do
    1. Steps to implement:
      1. First deploy it on test net and check out if this interest model works safely
      2. And implomet on Polygon or Avalanche to get real world feedback, and if works better than current version then implement it to all other chains and versions

4. So on

Road Map:

    1. Currently we got $5K USD grants from Aave Grants DAO (shout-out to https://twitter.com/AaveGrants and https://twitter.com/_alekslarsen)
    1. I’m currently working on more details of Aave Interest rate Uber Model
    1. I’m currently working on AaveRateDAO:
      1. It’s DAO that works like Grants DAO that constantly monitor, and use Forum & Snapshot to collect voice of aToken holders and make Aave Governance proposal to change Base Rate of different tokens, and so on
      2. More details are coming
    1. Purpose Snapshot for “Aave interest Rate Uber model” and “AaveRateDAO”
      1. I’ll make proposal on Snapshot to see if Aave token holders supports or not
    1. Purpose to Aave governance to implement the “Aave Interest Rate Uber model
      1. If the Snapshot resolute greate, then I’ll make proposal to the Aave Governance voting to implement it
    1. If get implemented, then propose Aave Governance and Aave Grants DAO to fund Aave Interest Rate DAO
      1. Aave Interest Rate DAO doesn’t need too much budget
    1. So on

Extra thoughts: (read only)

1. Gov tokens extra thoughts:

    1. Gov token original use cases & misuse case:
      1. Gov token LPs are using Gov tokens as collateral, doesn’t want earn yield on it(can’t make high enough yield on it anyway), just use it as collateral to take stable coin loan
      2. Aave made it available to borrow Gov tokens as similar to borrowing stable coins without permission of Gov token LPs(Gov token LPs has no choose but accept it)
    1. Governance attack risks & Aave good intention:
      1. Who will borrow the Gov token, and what are the use cases?
        1. People that have enough money to buy, but borrow it to use vote on Governance proposal
        2. I don’t think anyone don’t want to own that token and borrow the token to vote on governance of that protocol won’t as good intentions as token holders does
        3. If borrower makes bad vote, then the collateral provider ends of getting affected by bad decision that borrowers made, so it’s better for safety, and fair governance of DAOs, if gov tokens can’t be borrowed
      2. Aave by making unavailable to borrow Gov tokens:
        1. unintentionally protect and help DAO space in a huge way
      3. Trade offs:
        1. Gov tokens not making high enough yield anyways
        2. Yes some people use that Gov tokens for staking, Swap liquidity providing, and so on(they can buy instead of taking loans)
        3. Yes there is not enough liquidity on Aave to bring big risks to the voting process, doesn’t mean there won’t on in the future, and there’s still some level of risks
    1. Aave unintentional surprise:
      1. Why despite Gov tokens like UNI, COMP, MKR, and so on doesn’t make any profit, and why not more people use it to take stable coin loans?
        1. Because larger holders worries about Gov voting attack
        2. If Aave make it unavailable to take Gov token loan, then I think it might inspire more people to use their Gov Tokens to take Stable coin loan

1. Higher utilization & LP instant withdrawal problem:

    1. A or B withdrawal(works for stable coin only)
      1. If A token doesn’t have enough available liquidity to withdraw
      2. Ask LP if they willing to withdraw B token that has same value
      3. If the LP doesn’t want to withdraw B token, then Aave can swap(Curve) B token to A token, and give withdrawal to LP
      4. Already borrowed A token future earning interest rate will goes to B token pool
    1. Sleepy withdrawal: (works for any tokens)
      1. If there is not enough liquidity to withdraw, then LP will deposit their aTokens on Sleepy withdraw vault, and sleepy withdraw will slowly send the liquidity as soon as the borrowers returns the funds
      2. Borrowers has to return the funds every 6 months
        1. if not will charged higher APY
        2. Insure LPs can withdraw their liquidity, if they want(with help of Sleepy withdrawal)

2. First come first serve:

    1. If any token pool deposit rate gets to base rate, then any tokens that deposited after that can’t get yield on the deposit, but only use their tokens as the collateral to borrow
    1. When other withdraw their tokens from the pool, then not yielding tokens replace that token start earning yield
    1. Or if the deposit rate goes up by more borrowing, then not yielding tokens start earning yield
    1. Main purpose is give productive utilized tokens give guaranteed yield above the base rate

3. Aave Interest Rate DAO: (not necessary, but helpful)

    1. What is Aave Interest Rate DAO & how does it work?
      1. It works like Grants DAO that main purpose is focused on interest rate, that works to help Aave improve Interest rate model to be more productive, and better over time
      2. connect aToken holders opinion about interest rate to Aave governance
        1. With aToken holders Snapshot voting results
      3. Operate:
        1. Aave Interest Rate DAO Forum
        2. Aave Interest Rate DAO Snapshot
        3. Manage: Multisig wallet, Website, Social media accounts
    1. Work:
      1. Aave Interest Rate DAO Forum:
        1. Discussion:
          1. Anyone can discuss about interest rate of any token pools on Aave
        2. Community updates
          1. Operator will writes monthly report to update all about interest rate of each token pools
      2. Aave Interest Rate DAO Snapshot:
        1. aToken Voice:
          1. aToken holders can vote on Snapshot to express their thoughts to AAVE holders, and Aave DAO will vote to change the interest rate as they want
        2. Voting tokens:
          1. any aTokens
        3. Vote to change Based rate
          1. based on the time, utilization, Crypto Market, Risk, and so on
      3. Social:
        1. Setup Website:
          1. All about Aave Interest Rate
        2. Setup all kinds of Social media account on different platforms to inform more users:
          1. Github, Twitter, Discord, Medium, and so on
      4. Multisig Wallet:
        1. Store the funds to operate the DAO
        2. Accept AAVE vote delegation, and vote on Aave Governance proposals

Disclaimer:
1. It’s not a finish version yet, I only finished 70% of what I have in mind;
2. It’s not the final version yet(can be improved), if the Aave community supports me to work on it, then I’ll continue work on it.

Quick note:

  • I wrote all of this almost 4 month ago, and submitted to Aave Grants program, and few days ago I received Aave Grants confirmation & they required me to post it on the forum, so I made this post, that is y it’s kind of long, hard to understand, and messy
  • I’m going to work on the next version, and I will post on the forum. the next version will be more understandable, and shorter
  • If you guys disagree with any of my thoughts just comment bellow, if you don’t understand it, wait for next version post
    If anyone have better solution or want to join the team, just DM me on Twitter

Thank you for reading it, if you have any question just ask me on on Forum or DM me on Twitter

I hope you guys like it. if you guys like it, give a like and comment below.

Thanks!

Project Twitter: https://twitter.com/AaveRateDAO

Author Twitter: https://twitter.com/samuchiha_eth

5 Likes
2

Hi @SamUchiha , nice research work, congrats.

Some details, before entering in further conversation.

  1. You mention how the system as it is right is un-optimal because of the “spread” existing due to not reaching 100% utilization. This is true, but at the same time is the desired mechanic. In practice what it means is that if the utilisation is let’s say 95% on stablecoins (above the optimal point), and borrowers don’t repay together with depositors not depositing more, there is market equilibrium.
    In this model, if the rate would stay overtime at 100%, usually it means that the curve of the interest rate strategy is not correctly defined: borrowers are willing to pay more interest.
    Of course this could be improved, but just to explain the rationale.
  2. In reference to your point 7) of the current model section. It is actually possible to change the rate, just by updating the interest rate strategy contract. Also, not all the tokens have the same interest rate strategy, it depends on the asset itself.
  3. Another fundamental aspect that I don’t know if you are considering is what is actually the main point of the interest rate for non-stablecoins. These assets are mainly used as collateral, so as you correctly point out, the borrowing demand is not high; in some cases, people simply don’t borrow them.
    The interest rate based on utilization, in this case, acts also as a protection for proper liquidations. The optimal point on the curve is always lower than the one of stablecoins, and the interest rate spikes really aggressively, reaching usually hundreds of % APY on 100% utilization and close. This way, if somebody would borrow a big % of the available liquidity of the asset, or if somebody would withdraw a big % of that same available liquidity, the pressure of borrowers to repay would be really high in order to get liquidity in again, to have available for liquidations.
    The “bad” scenario would be if for example user A deposits $1m in collateral of asset X and borrows against it. Then user B borrows $900k of X with some other collateral. Then price of X goes down in a way that A gets liquidated, but there is only $100k of X for the liquidator to “take”, so it is technically not possible.
  4. Concerning your new model, from what I understand, it introduces an important component of fixed periods of time or similar, in order to calculate the discounts you described depending on being above amount and time borrowing. From my experience, I can already tell you that combining the current model with that model based on periods is a daunting implementation, not sure if even possible without completely rebuilding Aave, and maybe not even then in an environment with the constraints of Ethereum.
  5. You mention about the lack of voting power for aToken holders. As I was deeply involved on that decision I can explain a bit more detailed the rationale.
    In general, we can divide AAVE (and its flavors like stkAAVE or stkABPT) and aTokens between “slow” and “fast” assets. What this means is that the profile of holders of AAVE is generally more looking to a longer period than aTokens, as these last usually are just a mechanism to an end (provide liquidity to yield, have collateral to borrow, etc). Another aspect of the fast/slow dichotomy is that enter/exit on the asset has completely different speeds and constraints. Enter in the AAVE asset means usually buying it from the market or borrowing OTC, with the limitations of cost and availability on both cases. Enter in aTokens is as simple as depositing in Aave.
    The previous aspects lead to a situation where a potential bad actor could have a way easier path to manipulate the Aave governance if aTokens would have a meaningful influence on the voting/proposition strategy.
    In addition, by participating in systems like the Safety Module, AAVE holders are taking a risk supporting the protocol (with a reward of course), which strengthens the position on why they decide over the Aave pools.
    That being said, I generally tend to agree that could be interesting to have aToken holders having some kind of participation in governance, even if minoritary. But it should be in all cases really well thought, tested, and progressively ramped up, to avoid bad surprises.
  6. Concerning the implementation, especially your points about proxies, at the moment the current model used on Aave is a pretty proven standard, and even if sometimes not the most optimal, in general, I don’t agree that it is so burdensome to deal with upgrades. Also, given where proxies are used, it is not really an option to change at the moment.
  7. Could you elaborate more on 3.2.1 of the section of your proposed new model?

I think a lot of your ideas are really worth to develope forward, so let’s continue the discussion.

3 Likes
3

Thanks for the reply.

1.1. Reply:

  • First of all, spread is not caused by desire mechanic, but caused by low utilization, and low utilization is caused by high interest rate(when the utilization reaches over 75%, there will be high rate)(when the utilization goes up that users borrow from Compound instead) there is enough demand to reach 100% utilization, as long as the interstate is right(EX: currently USDC liquidity on Aave $4.655 Billion, and total loan taken(Aave $2.785 Billion)(Compound $1.519 Billion)(not calculating so on lending protocols across chains) there is at less $4.304 Billion USDC demand only on two platforms) I don’t see any demand problem here? All I see is current interest rate indirectly causing low utilization & big spread between borrow rate and deposit rate
  • You guys can check out, almost all of stablecoins most of the time at max will only reaches around 75% utilization rate!(why is the utilization rate always below 75%?) Why is that? I think because there is no reason to take loans on as high rate, and it might gets even higher(scaring the borrowers)
  • By implementing new model we can’t make users to borrow more, but by offering stable & reasonable interest rate, new model will indirectly inspire users to take loan on Aave(results higher utilization rate)
  • I’m more focusing on how to reach higher utilization first, and then think about market equilibrium. If you have read “Higher utilization & LP instant withdrawal problem” there is 100+ ways to create market equilibrium, but it’s super hard to reach 95%+ utilization
  • If the utilization stay on 100% not mean is interest rate strategy is not correctly defined, but absolutely opposite(when utilization reaches 100% there will be higher enough interest rate that reasonable for borrowers to take for short period of time & LPs will be happy to earn stable higher rate), (if the pools are earning higher enough yield, there is always will be more liquidity deposited to lower the utilization)(we can also implement a time limit for borrower that takes loan after utilization reaches 90%, and so on)
  • Borrowers won’t pay too high interest rate, but will pay fair enough higher interest rate that can be acceptable for a while(weeks).

1.2. Reply:

  • Sure, I know it’s possible to change the rate, have you guys changed in the past? Did it make any differences, no matter how you guys change it, it doesn’t make much differences
  • By same interest rate strategy, I meant all of them relies too much on utilization based interest rate model

1.3. Reply:

  • By non-stablecoins it can be Gov tokens(AAVE, MKR, UNI, so on), and Gas tokens(WBTC, ETH, AVAX, MATIC, so on)
  • For Gov tokens, I offered 3 different options, and one more extra opinion on “Extra thoughts”; for Gov tokens and low liquidity collateral tokens just cancel all the utilization model, and offer daily stable rate(because A, B, C, D reasons)
  • For Gas tokes, we can use Stable coins like interest rate model, because there is always will be high enough liquidity, and I thinking Gas tokens in the future will compete against to stable coin
  • Gas token like MATIC, AVAX are makes as high as 9% ~ 11%APY by staking it
  • I’m assuming overtime there will be more use case for Gas tokens in the future, and biggest reason of gas tokens used as collateral is because there is not higher stable coin liquidity to match the demand, and stable coin can earn higher rates than gas tokens
  • Over time more people adaptes blockchain, then there will be more stable coin liquidity comes in, and over time stable coin rate will goes down slowly, that people will use stable coins as collateral to borrow gas tokens
  • We can implement first come first serve method(as I mentioned on “Productive token pool & Lazy token pool), so utilized tokens will earn high rates, not utilized tokens will wait to get in to productive tokens pool to earn high rate in the future

1.4. Reply:

  • We can make changes to that. Make it easier to make it implomentbel by changing the rules
  • Basic idea is: 10%(or 5%) of all the fees can go to a pool, and only distribute to VIP users, and so on
  • This is also one of the reason why I’m pushing you guys to implement EIP-2535, and doesn’t need to be completely rebuild the whole thing, go check out the EIP-2535, current contracts will be stored on library, and can be used immediately, and then update the interest rate model at anytime

1.5. Reply:

  • I only mentioned little bit about aTokens voice contribution, and in the future I’ll share more details about how aToken holders contributes their voice without interrupting Aave fair governance
  • aTokens advantages over Aave tokens: 1. aToken holders are the root of the protocol, if any aToken holders disagrees with interest rate then that is not a good sign ; 2. Let aToken holders diode the interest rate, and if even earns lower, there no blaming to Aave governance; 3. Make aToken holder feel important to Aave governance; 4. And so on
  • aToken holder voice indirect contribution with help Aave Interest Rate DAO & limitations: aToken holders only contribute their voice to interest rate(interest rate related) issues of their own aTokens, any aTokens won’t have any voting rights on Aave governance, Aave Interest rate DAO will operate aToken Forum & aToken Snapshot, and update all the activities on the Aave Governance & get voting results on Snapshot
  • Anyone can delegate AAVE voting rights to Aave Interest rate DAO multisig wallet or Aave governance decide to delegate a limited amount of voting power to Aave Interest rate DAO multisig wallet
  • About interest rate update, Aave Interest rate DAO actively makes proposal to Aave governance(based on aToken holder Snapshot voting result), and proactively vote on all other governance proposals that indirectly affects the interest rate of any aTokens
  • Basically aTokens will have a representative, and the representative won’t have any interest on attacking Aave governance, the representative will mostly consider the aTokens holders decision, but makes rational voting design to protect both sides, the voting rights and voting size will given by AAVE token holders(or governance)

1.6. Reply:

  • About OpenZeppelin proxy contract, I exaggerated a little bit, I love the way that how OpenZeppelin made their Proxy contract, with current proxy method, Aave better suitable for upgrade once in a while, and more reactive to the historic performance of the protocol
  • I’m pretty convinced that in not too long future, when Blockchain users grow, then there will be even more competitions, and banks will offer more centralized Coinbase like lending mechanism, and they can more actively change the interest rate to compete with Aave like Defi protocols
  • It doesn’t have to be centralized lending platforms, but other money market protocol that uses EIP-2535 standard to actively change the interest rate to actively competes against to Aave rates
  • In order for Aave to compete against to other protocols, Aave eventually immigrate to the EIP-2535 or like standard to make it easier to proactively change the interest rate based on market data & events
  • Based on the current less competitive market, you guys think it doesn’t make much sense to change the proxy right now, I’m pretty sure in the future Aave needs fast & more interest rate changes(to stay competitive). You guys might also think Aave is too big to make big changes to contract, but potential size Aave will be in the future, it’ll never big to improve it as early as possible(EIP-2535 just makes Aave more easier to make changes, and more compatible to more future implementation, and as you considering the contract memory size limitation(EIP-2535 also can help with it) )
  • I’m just suggesting, not necessary for now if you guys don’t want to do it.

1.7. Reply:

  • Reply: not sure about what you are referring to? if it’s about “Gov Token distribution & Base rate insurance fund” I made some new changes to it , you can check it out again

Extra Note:

Based on what Aave is currently experiencing, I might exaggerating the situation, but it’s always helpful to think ahead to prepare and improve

I’m predicting in not too longer future, there will be more gov regulation come to Defi, and Defi space will eventually flooded with stable coin & and all kinds of liquidities, and there will be scenario like current centralized organizations(Bank), that Aave will face super low utilization rate(especially with stable coins)

For current Defi space there is not enough competitions, and not enough liquidity to reach the high demand, so there is high interest rate, and current utilization based interest rate works well at this moment(Aave community taking interest rate model as granted), so there is not any big interest rate discussion, but on centralized organizations like Banks, lending companies, repo line, bonds, so on money markets the interest rate is the most important topic

I’m pretty sure based on liquidity amount grows, over time all the Aave like money market protocols will eventually works super hard to improve their interest rate model(that is why I’m proposing to change the Proxy contract to EIP-2535 & I’m going to purpose Aave Interest rate dao)

As you guys might feel, I’m a super pessimistic, long term thinking guy that looks at bigger picture to make Aave ready for all kinds of small and big probabilities

Aave can stay as it is now, and work well as it does right now. I’m just proposing all kinds outcomes to make Aave interest rate as perfect as possibly can to make Aave more competitive and compatible

The interest rate that I’m proposing can be improved, and I’m pretty convinced that Aave in order to be more efficient, eventually change the interest rate & Proxy upgratible contract now or some time in the future

Thank you @eboado for your support. Don’t forget to drop a like, and I’m happy to continue this discussion.

1 Like
5
  • I’d say it is more complex than that. There is important leverage of stable coins on the market, which causes always a “spread” created by LTV/liquidation threshold. For example, it could be important to understand how much % of the capital is in that situation.
  • I tend to agree that usually thinking on the liquidity available for withdrawals is secondary compared with optimization of interest rate, but that given a situation of non-100% utilization and enough absolute liquidity available.
  • I think the willingness of borrowers to pay high interest for time X is quite uncertain. It is clear that those dynamics happen in special cases (e.g. some farm appears and people start borrowing and paying 50% because the farm gives 100%, or some “discrete” event happens, like a snapshot for airdrop), but to assume your last point, it is mandatory to analyze the profile of borrowers based on borrowing period.
  • I think you are going too far in your assumption. From what I remember, I think the community decides to change some interest rate curves in the past, and pretty sure there was some effect. I can’t remember right now exact changes, but they are visible on-chain, and maybe @Alex_BertoG has some extra information.
  • It would be interesting to model how it would affect your proposal of applying to gas tokens same rate as stablecoins. I’m still a bit skeptical about it because my feeling is that it would affect liquidations, but I can easily be wrong.
  • Agree with the rate going down over time with adoption, but I think we are not really on the point of using stable coins to borrow gas tokens to pay for gas. Probably the only case of that at the moment is shorting the gas tokens.
  • As a remark, I don’t really think there is anybody depositing for the yield on borrowing on gas tokens, just for usage as collateral (apart from liquidity mining). Maybe now with an asset like stETH, but probably neither.
  • I’m not an expert on the details of the diamond pattern, but I’m fairly sure that or the entry point of interaction with Aave would need to be changed, or the current implementations under the proxy. So I don’t really think it is an option, also considering the upcoming V3. But if you have some precise idea, please share.
  • I think it is something interesting to experiment with, so not much to comment on. However, I think that there should not be any middle entity involved between aTokens → Aave gov decision, at least not by design. Then if aToken holders want to delegate on somebody, sounds legit to me.
  • I’m still a bit confused about the high correlation that you see between frequent changes and EIP-2535. If the target is to be able to change interest rates parameters by just calling a function, it is perfectly doable with the current architecture (with some implementation changes), could you please explain how you see it?
  • I think I get it now, but I think the token distribution to participants in Aave has a different objective than just “compensate” rates. At least at the moment, the rationale of the community doing it is about distributing governance participation for people involved in the market. Basically doing kind of the same as you propose with giving aToken holders a voice on governance, by giving them stkAAVE distribution.

Fully agree with your last thoughts. The interest rate model is frequently seen as a fact around, but it would be a big mistake to just sleep on it and not trying to improve it. Personally, I believe these kind of developments are the strongest best long-term on a protocol like Aave.

1 Like
6

2.1 Reply

  • Sure , my point is we can make the spread smaller by improving utilization to the best possible levels.
  • About LTV/liquidation threshold: on the market there is enough stable coin liquidity ,so we can leverage other low utilized stable coin pools to temporarily share the liquidation risks, so don’t have to worry about LTV/liquidation threshold(I’m not explaining it clear enough, but we can solve LTV/liquidation threshold problem easily)
  • About available liquidity to withdraw: we can leverage other low utilized stable coinst, or implomet “Sleepy pool” set a time limit for borrowers if any pool utilization rate goes above 95% borrowers has to return their liquidity at last days of that month and then reborrow after returning the liquidity, and so on(I think it’s not a hard problems to solve)
  • with new model we completely ignore all the discrete event borrowers(despite small numbers of users willing to pay high amount for short peariate of time, over all they won’t contribute as high amount in fees), not every users earns that high amount at the same time, and by those borrowers raising the rate, all other borrowers gets affected
  • I do really get it, what do you mean by “mandatory to analyze the profile of borrowers based on borrowing period?”
  • It’s hard to give 10%+ borrowing rate for long time, but it’s not that hard to give 5%+ borrowing rate for a while(borrowers welling to take 5% borrow rate for a while)
  • Problem we’re facing: how to make the interest rate acceptable, and predictable; How to attract more borrowers to grow the utilization of stable coins?

2.2 Reply

  • Well, i’m not saying current interest rate model is absolutely useless, and new model is 100X better, but new model can absolutely helpful, that will solve a lot of problems of current model currently have and might face in the future;
  • Current model works well with utilization between 50% ~ 75%(will it stay forever liket this? An: NO); New Model will works well with utilization 0% ~ 100%(worthwhile to implement)
  • That’s a great news, I would love to know more about past interest rate updates & discussions

2.3 Reply

  • I’m not saying same rate, but same interest rate model with different Base rate, dement added rate, Discount rate, and so on
  • It takes pages of words to explain it, so wait for next version update
  • I’m thinking nometter the collateral is(stable coin, Gov tokens, NFT, Stock tokens, so on), based on use cases of Gas tokens grow, the utilization rate of Gas tokens will grow & will be higher than stable coin over time
  • Cases of borrowing Gas tokens & Gas token use cases: (0)use it to pay for Gas (1)use it to ear staking yield (2)Flash Loan (3)under collateralized loans(at some point will be solved) (4) Gas token based fractionalized stable coin (5) credit line between other Defi protocols to issue(options, futures, leverage tokens, and so on) (6) between other Defi and Aave liquidity line(EX: users provide their liquidity to Aave instead of Uniswap, and much more) (7) 0 collateral loan(I’m working on one(kind of longer term flash loan, that has some limitations & more functionality)) (8) tokenize the stocks of centralized organizations, and they use their stock tokens to borrow ETH (9) most of the NFT are valued in ETH, if Aave offers NFT as collateral to borrow ETH, then demand for ETH will grow drastically(i’m working on a solution, i’ll share in the future) (10) so on; (I just made it shorter, if you interested we can talk about it on twitter or zoom)
  • Aave should stop all the LM to Gas tokens(useless, waste of money, so on), instead offer higher Grants prize for developers to introduce better solutions to implement NFT as collateral to borrow ETH like Gas tokens solutions from Aave

2.4 Reply

  • It needs long explanation, so I just don’t comment on it for now
  • I was overthinking a lit bit, now I realize Aave doesn’t need it for now, lets focus on in front of us first
  • New interest rate model also works well with current proxy too, let’s forget about EIP-2535 for now

2.5 Reply

  • Well, Aave Interest rate DAO is a experimental thing, and at this point it’s the best available option to at less make some kind of differences to move forward, and it’s better than nothing
  • It can be canceled or upgraded over time(there is nothing to lose just by experimenting it first)
  • About aToken holders delegating to somebody:
    • First of all there won’t be too much aToken holders cares about Aave governance any time soon(only higher amount holders will care(and they most likely also holds AAVE token too))
    • There is more than one types of aTokens, does every aToken holders need to find somebody that holds highest biggest amount of same aToken as them, who does they find, and how does they trust him(it’s not realistic)
    • aToken holders are not a consistent, so it’s hard for them to unite each other to make a long term decisions(it’s not realistic)
    • If they have AAVE token they can still delegate to Aave Interest rate DAO multisig wallet too(just saying)
  • Why I’m saying delegate voting power to Aave Interest rate DAO with Aave governance decision:
    • Aave Governance proactively take the responsibility to make some kind of action to connect aToken holders to Aave governance
    • If Aave governance organize a aToken representative, aToken holders will be happy to follow, if not i don’t see there will be any aToken contribution at any time soon
    • Aave governance can decide the voting power of the aToken representative at anytime, and check it actions at anytime, and cancel it at anytime
    • So on
  • Aave interest rate DAO helpful parts & Risky parts:
    • Helpful parts: (1) operate a Forum to constantly inform & update aToken holders about Aave Governance (interest rate related)activities, and create discounstion between different aToken holders(because I think aToken holders only has to care about interest rate related issues only) (2) Help Aave Governance informe the opinions of aToken holders about interstate updates of any tokens
    • If Aave interest rate DAO doesn’t do it, then who will do it? aToken holders? Or Aave Governance? Somebody has to do it.
    • Risk: I don’t see any risk, if Aave governance worries about voting power delegation, then Aave governance can only delegate very small amount that by voting Aave interest rate DAO only express opinion of aToken holders, if you guys thing the middle man will miss represent, then Aave governance can vote on their own way, and check if representative of Aave interest rate DAO miss represented, and remove the delegated voting power at anytime
  • I know there is better way to solve it, but it takes more work, and at this time it’s not available, so Aave interest rate DAO is best available and easiest option

2.6 Reply

  • I was thinking fast speed of crypto space, until the market grows bigger enough to set a automated interest rate model, it would be better to constantly monitor the market feedback to manually upgrade the base rates to make the Aave more affiliation to market reactions
  • And I don’t think 100% automating the interest rate is good option, we can set utilization based small dynamic changing function, but base rate and other rates has to be manually upgradable based on all kind of market data & aToken holder voting decision

2.7 Reply

  • Aave community think LM as: (1)fair token distribution to the community (2)incentivizing LPs to move their assets from v1 to v2 (3)incentivize LPs to deposit their tokens to Aave (4)lower the borrowing rate by incentivizing the borrowers (5) most of the LPs are also AAVE token holders, they vote to pass LM, so they can get as much as they can from Aave Treasury
    • Fair token distribution to the community:I fully agree with distributing tokens to the community(It’s a good thing in theory, but at this point it’s kind of a way for big aToken holder that has big AAVE voting power making excuses to earn Aave treasury amount); let’s be real here is it really a fair token distribution to community? It’s kind of giving money to LPs to stack their liquidity to the protocols, because I don’t think those LPs will hold their Aave rowordes(they will immediately sells it(it hearts the price of AAVE constantly)), and wasting Aave treasury liquidity;
    • Incentivizing LPs to move their assets from v1 to v2: can be done with other ways(chose lucky LPs that moved v1 to v2 make big airdrop or so on)
    • Incentivizing LPs to deposit their tokens to Aave: all LPs cares about the deposit rate, LM is not sustainable way to give better deposit rate(instead let’s spend all those funds to give ads, fund more developers, so on) use those funds to grow the utilization, then there will be higher deposit rate, and there will be more liquidity
    • Lower the borrowing rate by incentivizing the borrowers: well it’s kind of similar situation as incentivizing the LPs, lets focus on how to attract more borrowers, as long as there’s enough demand, Aave doesn’t have to worry about the borrower rate, if there’s not enough demand(there won’t be enough funds to lower the borrow rate forever)
    • AAVE holding as LPs: I’m pretty sure that most of LPs also holes AAVE, and I’m predicting those LPs voted to start the LM, so I suggest AAVE token holders to vote to stop all the LM, and use those fund to improve Aave not rob Aave treasury & community
    • Better way to operate LM: I’m pretty sure in the future there will be stable coins 15% ~ 30% utilization for long time, if that happens, then Aave should give LM to Stable coins only grow the deposit rate to base rate time to time
    • Token distribution is super easy, can be distribute at any time in the future, i strongly suggest Aave Governance to save and grow the treasury liquidity amount, because based on the grows of the Aave protocol the expenses will also grow together, and I suggest focus on more developer incentivization side to grow the Aave in real affective ways, allocate 10X current amount to Grants DAO
    • Funds currently using on LM instead should be used on: (1) 10X the Grants DAO(run more hackathon that focus on utilization grows of stable coins) (2) give Ads to attract more users (3) work event harder to make institutional connection to Aave(Aave ARC) (4) so on ways to incentivization developer to improve utilization of Aave
    • Yes, Aave treasury has enough liquidity to do all at the same time, but i’m just suggesting we have better more effective ways to solve those problems, so I’m really against so called LM
    • Yes, I’m little bit off topic there, but I’ll explain it in next version, why I’m talking about how to use funds and so on(LOL)

About working on Interest rate model:

  • I think the reasons that why nobody is working on interest rate model are: (1)Current model seems to woking well with utilization between 50% ~75%, and based on liquidity shortage on Defi space, users willing to pay higher rates, and Aave having higher(50% ~ 75%) utilization for most of time, so nobody think interest rate is a problem at the moment (2)Nobody thinking ahead to assume there will be more stable coin liquidities coming, and eventually there will be 15% ~ 25% stable coin utilization for long time in the future (3 )And nobody recognizing the indirect negative impacts of current model, that causing big spread between deposit rate & borrow rate(I mean we can make it much smaller by only improving interest rate model) (4)if the Aave team doesn’t do it others won’t do it (4)And nobody has 10x better solution, so nobody standing out like me(ignoring the so obvious problem) (5) work on Interest rate supposed to be one of the most important to put more work on , and it’s super hard to constantly improve;

Extra thoughts:

  • MORE Ways to grow utilization & TVL:
    • Improve liquidation prosprocess & threshold(Make Aave more competitive agins to Makerdao & and all other money market protocols): grow the liquidation rate(I have a solution for this, I will post it in the future)
    • Connecting Aave liquidity to more financial protocols like DYDX, Options protocols, and so on
    • Make Aave as liquidity source of other Defi protocols:
    • So on

Quick note

  • I made a lot of changes to v0.1(current version), and not long future I’ll post v0.2(it’ll be much better than all the mess on v0.1)