Proposal to improve Liqwid protocol

Temperature Check
1

I am writing to submit a draft proposal to improve the Liqwid Protocol.

Aim of changes:

  1. Increase TVL on Liqwid Protocol
  2. Efficiently allocate protocol resources with the aim of growing TVL incentivising long term supply and borrowing
  3. Prevent arbitrage and prevent valuable protocol resources (LQ rewards) from being wasted on temporary ADA borrowing
  4. Incentivise and onboard larger players to the protocol
  5. Reward risk taken by participants to increase bridged asset supply

Changes I recommend to improve the protocol:

  1. Reduce the maximum withdrawal rule in one go from 1/16 down to 1/3 of 90% available supply (ideally get rid of this rule).
  2. Increase the above from 90% of 1/16 drawdown to 100% so that suppliers can withdraw up to 100% of supply for any market.
  3. Rule for interest rate when utilization of an asset reaches 90-95% - higher interest rates.
  4. Rule for interest rate when utilization of an asset reaches 100% and above meaning that borrowing is no longer possible and borrowers are not repaying the debt causing borrowing to exceed 100% of supply by way of interest accruing – higher interest rates.
  5. Limit LQ rewards to 1:1
  6. Change LQ Rewards to reward assets based upon highest; 1. Utilisation % and 2. $USD value in assets supplied. Reward supply side.
  7. Incentivise bridged stablecoins supply more by way of paying; A. Minimum LQ APY reward (Minimum monthly/weekly LQ) rewards for stablecoins supplied and B. Minimum LQ APY reward (Minimum monthly/weekly LQ) rewards for bridged assets supplied.
  8. % of interest generated by the protocol should go towards an insurance fund/bug bounty program etc
  9. Remove LQ rewards for borrowing assets and repaying interest
  10. Increase minimum interest rate to match (Aada charges 25% of interest minimum for the fixed term loan)

Reasoning for above:

  1. Larger players cannot withdraw assets, for example 326k USDC available but maximum withdrawal in one transaction is only 5.7k USDC. This will put off larger players from bringing in volume to the platform. This is also additional risk for those who brought over bridged stablecoins, now they are stuck. Thus nobody is supplying it despite the higher supply rates current offered. Market has decided that the risk isn’t worth the reward.
    Larger participants are currently held hostage by this rule, as they cannot withdraw the funds they supplied! If 5.7k USDC is withdrawn by a participant then in the next withdrawal transaction they can only withdraw a maximum of 5.5k USDC then a maximum of 5.3k USDC in the following transaction. The max withdrawal amount decreases with every new withdrawal! A participant who supplied 1,000,000 USDC to the protocol would never be able to withdraw their funds without doing hundreds of withdrawal transactions. Imagine placing your money into a bank and the bank telling you that you can only withdraw a maximum of 5.5k at a time, then on the second withdrawal the max withdrawal drops down to 5.3k then the later one down to 5.1k and so on.
    In the above example where 326k USDC is available suppliers should be able to withdraw at least 1/3 of it in a single transaction at 100% face value instead of 90% face value.
  2. Higher interest rates once 90% utilisation is reached will bring in more supply as the current reward structure isn’t bringing in more supply to stables (USDC and USDT), rates are not rewarding enough, if rates were rewarding enough for risk taken then users would have supplied more USDC and USDT etc.
  3. Prevent interest repaid to LQ Rewards arbitrage e.g. currently we are facing this with the Optim bonds.
  4. LQ APY to reward most borrowed assets on Liqwid to promote the increased supply of those assets mostly borrowed by utilization AND USD value. Increased supply reduces borrowing costs thus increasing borrowing and protocol TVL.
  5. Insurance policy to protect LQ participants, a small percentage of all interest generated to go towards protection from loss, hacking and bug bounty program. Percentage of interest automatically should go to wallet(s) as an insurance policy offering assurance to the participants of the protocol that it can handle errors/hacks etc. Institutional or big money isn’t coming onboard primarily because of this reason.
  6. Efficient allocation of LQ assets to incentivise asset supply which increases TVL, bringing down borrowing costs (APR), which in turn increases borrowing and further supply into the protocol by borrowers who provide collateral to borrow. This cycle promotes actual TVL growth on the protocol.
6 Likes
2

Thanks @AliD for putting up this great proposal. I agree with some of your points. Below are my takes on these issues.

Point 1
I agree, but be mindful that we are trading volume over speed here. There will definitely be future congestion and more people complaining transaction cannot go through.

This should be dynamic I guess. Perhaps now we can reduce it to say 8 instead of 16. But when the times comes where TVL has grown significantly; maybe the importance then would be to accommodate to as many users as possible at one time, instead of volume.

Point 2,3 and 4
I disagree. The 90% borrowing cap is there to provide buffer so that it may take some time for interest accrued (in case of non-repayance) to hit 100%. The cosmetics just don’t look good if utilisation is above 100% no? People may lose confidence to supply if they see users are not repaying and utilisation is >100%.

All in all, the protocol is still working fine as every debt is over collateralised. Sooner or later, it’s either more people will supply, the users repay their debt (and/or interest), the users add more collateral, or ultimately liquidators doing their job repaying the loans for a profit instead.

Point 5
I disagree. LQ rewards should be keep as it is. We are already slow in monthly emission, reducing it won’t help with the tokenomics. Currently, we are emitting an average of 350k LQ per month, i.e. ~1.7% of 21 mil LQ cap. If anything, we should even accelerate this so that we reach at least 50% emission the soonest possible. And then we can start talking about slowing it down.

I agree the tokenomics of LQ is bad, but the product is great. In short term, I propose let the emission rate as it is and let the market do its own thing. If price falls further, let it be. It will find its bottom soon (if not already).

Instead, I propose to have our own protocol owned liquidity (POL). We should utilise all of the protocol share of 10% of income in ADA and pair it with our idle 3.36 million LQ sitting on treasury. This should help to deepen our liquidity pool, stabilizing the price and incentivise whales to make transaction with lesser slippage.

Point 6, and 7
Yes, I agree on this. As I said above, I do not agree on reducing LQ emission rate, but rather I suppose we can improve on the distribution mechanism that is fairer; i.e. eliminate any easy wash borrowing options.

Point 8
No, I disagree. We already have and idle 3.36 million LQ sitting on treasury. We should utilise that instead. The protocol income should be used for POL.

Point 9
I suppose the 20k ADA borrowing incentive is being gamed by those supplying optim bonds. It is as good as wash borrowing.

Point 10
I think we should just keep min interest rate as it is. Maybe just need to tweak the post kink point interest rate. Users don’t seem to budge with the 50% rate. I suppose we need higher rates.

4 Likes
4

First off thanks for creating this temp check and for sparking engaging governance discussion both here and on Discord. I have reviewed all of your points across Discord and this forum post many of which I agree with to some degree. To begin my response I will say I agree with all of your points as part of Liqwid’s major long term goals in your “Aim of changes” section fully except 3 (Prevent arbitrage and prevent valuable protocol resources (LQ rewards) from being wasted on temporary ADA borrowing).

The reason is simple: Money markets are 2-sided marketplaces meaning they must effectively bootstrap both the supply and the demand side to achieve product market fit and long term sustainable revenues. This is not even specific to DeFi money markets or web3 this is how all 2-sided marketplaces across any vertical you can think of operate. This is also why Aave, Compound and the other largest DeFi money markets (e.g. marginfi on Solana) all launch with liquidity incentives significantly weighted towards borrowers (4 or 5:1 borrower to lender rewards split is normal). We did the opposite with Liqwid since all borrower rewards were paused after 2 months on mainnet and only restarted for ADA borrows in the last few weeks. Liqwid is the exception by targeting suppliers with LQ rewards only.

Now to address your recommended changes:

  1. Reducing the number of action UTXOs per market is a tradeoff as the lower the amount the higher chances for contention or users needing to wait until after upcoming batching cycle (though lower UTXOs allows users to borrow or withdraw larger amounts per action). If there was only 1 action UTXO this means only 1 user per action per batching cycle could complete actions in the market. This would be infeasible in our current setup with no support for Tx queuing. The core team has tested a reduction to 8 action UTXOs for some markets on preview. With Tx queues in the future number of action UTXOs could likely be reduced further to 3-4 UTXOs.

  2. The borrow cap is a Liqwid v2 risk control feature implemented to protect suppliers during increased market utilization (similar to the scenario you and other USDC and DAI suppliers are experiencing currently). As you can see if there was no borrow cap (Liqwid v1) there would be no ability to cap how much of a pool could be borrowed, this is how Aave v2 and Liqwid v1 operate. Here is a real example that played out in Aave v2 USDC market as a result of having no borrow caps (TLDR: Aave v2 USDC market hit 99.86% utilization and suppliers had to wait on borrowers to repay down. As a result of this event Aave implemented borrow caps as a key risk control feature in their v3 smart contract and Aave governance passed proposals to increase the interest rate curve past the kink for the USDC (something I support) to encourage borrow repayments and additional supply.

I would recommend to governance to adjust the 2nd slope of the interest rate curve (post kink point) so the Borrow APY at 100% utilization becomes the rate at 90% (perhaps even higher) and keep the 90% borrow cap (again this is a risk control feature to protect suppliers from a market going to 99.9% utilization).

  1. I think a standard rule for managing markets that reach some defined percentage utilization is a good idea.

I would recommend the rule should be implemented at 75% utilization for stables and not 90% since all stablecoins have the 90% borrow cap applied (this means it should be 65% for ADA and SHEN since they have 80% borrow caps). Higher rates yes but as far as the specific rule (e.g. increase the post-kink point slope multiplier by 10% every 7 days utilization for a market remains at or above the rule’s target utilization). Also think the rule could be more effective if it’s followed by some adjustment to the LQ rewards for the market.

  1. Utilization will never reach above the market’s borrow cap. The point above applies here fully.

  2. The initial governance vote to allocate the 20k LQ to ADA borrowers per month did not define the multiplier. Also in traditional liquidity incentives there is no multiplier at all: it is purely based on the USD value of repaid interest compared to the USD value of the incentive alloc. (this means with low usage users could earn multipliers of 5-10x or more and with great usage a multiplier as low as 1-3x or even less). Because of how our rewards engineer built the borrow incentives we are able to throttle the total LQ emitted by using multipliers, this means not the full 20k LQ allocation is being distributed (currently only 5-7k LQ per month of it is). Again want to point out this level of granular control on incentives is not how most liquidity protocol’s incentives programs work. The 3x multiplier has already been reduced to 1.5x and future changes to the rewards multiplier will be voted on via governance.

  3. This is imho an all around bad idea. Rewarding based on utilization % does not help lenders or the protocol in a meaningful way just take the current state of USDC, USDT and DAI markets: all have very high utilizations but very low monthly interest payment volumes in comparison to ADA, DJED and iUSD which have far lower utilizations but far greater monthly interest payments. Rewarding based on USD value of assets supplied will see majority of the rewards go to the ADA market, I am not sure that’s the most strategic way to grow the TVL and revenue of the protocol.

I recommend no changes to the current LQ rewards model. Suppliers in markets with the most USD value of interest repayments now earn the most LQ rewards as they should: these liquidity providers are depositing in markets with strong borrow demand and borrowers actively paying down interest. Borrowers repaying interest all earn the same 1.5x on the USD value of their repaid interest. I would recommend the formation of a subDAO team that periodically reviews market data (USD value of interest repaid, utilization %, USD value of each asset being used as collateral across markets) to update the rewards model based on latest data within some defined upper and lower bounds similar to how the Minswap Yield Farms updates their tiers/yields based on trading volume every few weeks.

  1. All markets already have a 100 LQ minimum for suppliers per month. I agree the protocol needs to find more ways to incentivize bridged stable supply but I do not think having wild LQ reward APRs is the solution, real yields need to increase and bridged stable borrowers must repay loans.

I recommend instead of trying to divide stables down into bridged vs. Cardano native we should just go the full way and completely remove the “stablecoin target APR” and treat them the same as all other markets (meaning suppliers earn an LQ rewards APR based on the USD value of interest repaid in the market they are supplying to). Some multiplier can also be applied individually per market (e.g. 3x for bridged stables, 2x for Cardano stables, 0.5-1x for ADA) to throttle the rewards even further.

  1. Liqwid Safety Pool is LQ held in the staking rewards multisig and is used during shortfall events and is the only implementation of an “insurance fund” in the Liqwid protocol. Setting % of interest generated by the protocol aside for this does not move the needle at all with a protocol that has $50m in total deposits. Currently there is no bug bounty program and any future bug bounty program will have specific USD amounts approved via governance vote and will utilize LQ from the Liqwid DAO treasury.

  2. This is a bad idea and have already explained the reasoning for this above. Liqwid is a 2 sided marketplace and demand side must be bootstrapped to generate revenue, this is exactly why all of the largest DeFi protocols start with borrow incentives 4-5x larger than supplier incentives.

  3. Liqwid loans are perpetual not fixed term so calculating what 25% of the interest will be an an open term loan is impossible.

4 Likes
5

i agree fully here

I also fully agree with this point.

Id like to see the 20k borrower rewards being available for all markets proportional to interest repaid. This way suppliers and borrowers earn the most LQ from markets with the most real yield generated.

4 Likes
6

I agree with itzDanny.

I was about to pay all my debt and withdraw my ADA but seeing that the APY on LQ rises to 8%… I have left them and am adding ADA. Stable coins will arrive when there is a good supply to maintain (USDM, USDC) and also the ADA reaches a price (0.70-$1) at which many begin to sell ADA or short it. DC I ask you to please think about it several more times. Getting it right, sometimes it’s a matter of not modifying anything and waiting for the right conditions

1 Like
7

As Danny wrote, the next step should be applying the borrowers rewards to all markets proportional to interest repaid.

2 Likes
8

I agree with the point of #1 to reduce from 16 to 8.

The points to increase the interest rate, I do not agree with. For the following.

  • Seems to be favoring suppliers over borrowers.
  • The supply has increased 25% in 10 days. Give this another 1-2 months and it will stabilize
  • The number of borrowers even at a rate of 45% has increased
  • The number of suppliers has increased
  • If nothing changes and no new suppliers of borrowers enter, liquidation will occur.

is there some advanced analytics that can be shared on the number of suppliers, borrowers, and total supply over the past 3 months?

From where I am sitting I am seeing an inflow of $50k-$100k the past 10-14 days.

Borrowers should not be ‘punished’ with Higher APRs than what was advertised to them when they entered the market.

1 Like
9

Changes I recommend to improve the protocol:
11. Dynamic LQ Rewards Emissions;

On the ‘backend’ the protocol should have multiple ‘LQ multipliers’ to represent different points of strategy and risk. A hybrid and dynamic system which distributes LQ to achieve goals.

LQ Multipliers:
*LQ multiplier for ‘cool markets’ (borrower) - promote borrowing
*LQ multiplier for ‘overheated’ markets (supply) - promote supply to reduce rates and promote further borrowing
*LQ multiplier for stablecoin supply
*LQ multiplier for bridged asset supply (reward risk taken by suppliers to promote supply and participation)
*LQ multiplier for other asset supply
*LQ multiplier for asset borrow

At any point the borrower or supplier for any market could benefit from multiple LQ multipliers simultaneously e.g. ADA market is ‘cold’ and not enough borrowing so borrowers should benefit from the normal LQ borrower multiplier and the LQ multiplier for cold markets.
Example for overheated markets; USDC utilisation and interest rate is high, despite this new supply isn’t coming in. So USDC suppliers to benefit from;
*LQ multiplier for ‘overheated’ markets (supply) - promote supply to reduce rates and promote further borrowing
*LQ multiplier for stablecoin supply
*LQ multiplier for bridged asset supply

When a market is ‘cool/cold’ and utilisation is below say 10-20% then the ‘cooled’ market LQ multiplier for borrowers should kick in to heat up the market to incentive borrowing. When the market is ‘overheated’ and say utilisation is above 70-80% then the overheated lq multiplier should kick in for suppliers to incentivise supply.

  1. Analytics page for Liqwid protocol and analytics for wallet supply/borrowing.
    Line charts showing;
    A. Liqwid Protocol analytics and
    B. User Wallet analytics
    Data should show at least; supply/borrowing amounts, supply/borrow rates, utilisation % per asset.
    @DC1

  2. Implement third ‘slope’ of interest rate to be applied for utilisation above 75-80% for any asset across the platform ideally capping around 365% max APR at 100% utilisation.

  3. Implement minimum interest ‘fee’ for borrowing between 0.1-1% of the amount borrowed.

  4. Remove the borrowing limit on assets to enable up to 98-100% utilisation, this coupled with the ‘3rd’ slope for interest rates above will make the market truly free, just like Aave and others. If borrowers want to utilise 98-100% of the market they should be free to do so.

Remove the max utilisation to enable 100% borrowing and introduce the 3rd slope (interest rate) at the same time. If the market wants to borrow 100% of utilisation they should be free to do so. True free market capitalism should be at play on the Liqwid Protocol. No more intervention by Liqwid! Let the market do it’s thing. 3rd slope kicks in around 75-80% utilisation with a max APR of 365% (or more) at 100% utilisation.

*If Borrowers want to borrow 100% of utilisation to leverage position they are free to do so.
*If Suppliers want to withdraw supply they should be free to do so.
*APR and APY rates to reflect real market dynamics and reward risk appropriately.
*If you are a supplier and cannot fully withdraw due to market utilisation you should be rewarded ‘fairly’ for having to wait.
*If a borrower can make 20% gains in a week by borrowing stablecoins and leveraging it to buy more ADA (while ADA is pumping) they can afford to pay max 1% APR per day interest rate to suppliers for 100% utilisation.

Set the market free.

2 Likes
10

@DC1, I appreciate you taking the time to read through all of my messages and considering them. Likewise, I have thoroughly reviewed all of your messages on both Discord and here. Here are my responses to your reply above:

  1. I agree that the majority of LQ rewards should be allocated to the Borrow side since borrowers are the primary customers, and they naturally drive interest rates, which incentivizes suppliers to provide liquidity. However, I believe the LQ emissions system should be dynamic and algorithmic, as outlined in Point 11 of my “Changes I recommend to improve the protocol” above. These emissions can be automated for efficiency.

  2. The reduction of the number of action UTXOs to 3-4 in the future, down from 16, is a significant improvement that should be implemented promptly.

  3. Regarding the incident where the Aave v2 USDC market hit 99.86% utilization, the Liqwid Protocol should not intervene in the free market. Borrowers should have the freedom to borrow up to 100%. However, when utilization exceeds 75-80%, a higher interest rate should kick in, potentially reaching up to 365% APR to reward suppliers for waiting. This incentivizes more supply without excessive intervention by the Liqwid Protocol.

  4. Introducing a third interest rate “slope” with a maximum of 365% could address theoretical issues where borrowing exceeds 75-80% and repayment is delayed. This aspect needs further discussion.

  5. Multiple LQ multipliers representing different segments, strategies, and risks of the Protocol, as suggested in Point 11, would be beneficial for the system.

  6. Reiterating the importance of implementing the third interest rate “slope” and referring to Point 11 for further details.

  7. High utilization triggering the third slope should be sufficient to incentivize further supply. If the market is not bringing in additional supply to stablecoins, it indicates that the current rewards do not outweigh the risks associated with bridged stablecoins/assets.

  8. Learning from the mistakes of other platforms like FTX and Celsius, which relied heavily on their native tokens (CEL and FTT), we should allocate a small percentage of all profits towards an insurance pool starting today to mitigate potential losses in case of a major hack or drain. The LQ tokens will be worthless in the event of a major hack or loss. What you rely on will be worthless.

  9. I now concur with the points stated above.

  10. While we cannot predict the exact borrowing behavior, implementing a minimum interest ‘fee’ from the moment borrowing occurs, such as a 0.1-1% interest fee, ensures that debt begins accruing immediately. This approach requires two calculators for interest: one for APR and another for the minimum interest fee.
    Example; borrower borrows 10,000 USDC they incur 0.1% minimum interest ‘fee’ so the debt starts at 10,010 USDC or 1% minimum interest ‘fee’ so the debt starts at 10,100 USDC. Then interest APR is added to the above figure(s). Lenfi imposes a minimum 25% interest for the fixed term loans even if the loan is repaid within 1 minute after borrowing. We too should implement a minimum interest ‘fee’ amount on the borrowings. 0.10-1% is very reasonable and fair.

1 Like
11

Point 2 we can reduce number of action UTXOs to 3-4 for all stablecoin markets as part of this governance proposal.

Point 3 I think you are still not acknowledging that it’s far worse for Liqwid markets to be at 99.86% in terms of liquidity for suppliers to exit. This is precisely why we support caps. I also hard disagree we need a 3rd slope, no lending protocol has 3 slopes they simply adjust the post kink slope 2 as required.

What you describe in point 4 is how Liqwid works today just the APR is not so steep that it gets to 365%, though I fully agree with your point the borrow APR needs to be higher to encourage new supply and loan repayments. Accelerating steepness to the post kink slope 2 can be tested encouraging one with APR of ~365% at 100% utilization. A simple rule that could be implemented at the interest curve level for when markets have utilization averaging within 15% of their borrow cap for 10 consecutive days is to set the borrow cap utilization (90% for stablecoins) borrow APR to the APR it would reach at 100% utilization. This “pseudo-slope 3” would need to be applied via Liqwid Admin Multisig once the criteria described above is met.

Point 5 we already do have multiple parameters (mostly individual multipliers) we can utilize to properly tune both borrow and supplier rewards. This process needs to be further fleshed out but is in the works. The goal is to implement a process similar to the Minswap Farm Tier bi-weekly rebalancing.

Point 8 I disagree and think the LQ Safety Pool should be used here. Your statement “we should allocate a small percentage of all profits towards an insurance pool starting today to mitigate potential losses in case of a major hack or drain.” misses the point that a small percent of monthly protocol revenue going towards an insurance pool does not move the needle for a protocol with ~$50M in deposits. The LQ Safety Pool as described in the Whitepaper and utilized recently should always be the primary insurance for the protocol. This also means a focus on liquidity for LQ needs to begin. I completely disagree with your last sentence, you don’t know what LQ price will be at in that situation (no one does) you are just purely speculating here.

Point 10 Liqwid already has a minimum interest parameter currently set to 0.08% for all markets. Internally we have considered adjusting this figure. I could see reasons for raising min interest to as high as 1% or keeping as is. Would like to see more compelling analysis here.

3 Likes
12

Point 2: Agreed, reducing the number of action UTXOs will indeed facilitate smoother movements for larger players. Lowering the barriers and friction for asset withdrawal encourages greater willingness among whales to supply liquidity.

Point 3: 2nd slope is good enough, 99.86% utilisation with a higher second slope resolves the concerns I highlighted. High utilisation with 1/16 UTXOs as the withdrawal cap was the issue. Reducing UTXOs down in point 2 (above) and introducing a higher slope in point 4 (below) resolves the concerns I highlighted. Set the market free, let borrowers and suppliers do as they wish.

The increased incentive provided by the higher second slope, both rewards and penalises during extreme market conditions and fills a critical gap in the protocol. This encourages suppliers to provide more liquidity (due to significantly higher APY), thereby reducing rates for borrowers (APR), while also penalising borrowers who don’t repay when utilisation is extreme, thereby reducing borrowing and allowing suppliers to withdraw. Points 2 and 4 need to be implemented together to achieve equilibrium in extreme market conditions.
*Borrowers receive more liquidity when rates are higher due to supply incentives.
*Suppliers receive more liquidity to withdraw their assets from other suppliers and borrowers who repay due to incentives.

Point 4: AAVE v3 2nd slope tops out at 300%, they don’t have a 10 day consecutive rule, I advise against the 10 day rule because the 2nd slope rate won’t kick in and borrowers are still left without access to funds. Also the 10 day rule can be hacked by a borrower who borrows and then repays to break the 10 consecutive day rule you set allowing them to stay above the utilisation rate without triggering the kink/2nd slope. A large borrower or a group of them could repay the debt for 1 minute then reborrow to restart the 10 day consecutive rule. I agree with the 365% max APR at 100% utilisation.

Point 5: I will look into the Minswap Farm Tier, I also proposed a dynamic algorithmic system rewarding cold markets to incentivise borrowing and rewarding overheated markets to reduce borrowing costs APR (A higher 2nd slope could resolve this issue without the need for providing further LQ rewards to suppliers, allowing a higher % of LQ reward allocation to borrowers). The LQ rewards distribution system should be automated/algorithmic, this will make the organisation ‘leaner’ allowing it to free up human resources to other tasks speeding up development in other areas.

Point 8: This part is a sensitive area, I understand that you founded the protocol. Please note that I am writing this part with the best of intentions at heart.
We have proof from Celsius and FTX, what happened to their native tokens once their systems collapsed or bankruptcy was announced?
Here is a question; In the event of a major wallet drain or hack the LQ tokens’ value will suffer their fate, how do you then intend to compensate/repay users in this situation?
What is your backup plan if you cannot repay/cover the losses in LQ tokens?
The LQ token is not invincible.
We must learn from past experiences and ensure adequate safeguards are in place.
Both Celsius and FTX suffered from the approach of “you don’t know what LQ price will be at in that situation” this really isn’t a defence for not having an alternative plan in the event of a major adverse incident.
We must have an insurance policy in place to safeguard the interests of all stakeholders of the protocol, this includes but not limited to; Users, suppliers/borrowers, staff and LQ holders.

Point 10: AADA charges 25% minimum interest as the fee for a fixed term loan (if you took a 100 day loan and repaid it back in 1 minute you would incur interest worth 25 days of the loan term), banks charge a fee for lending, AAVE is charging 0.09% one-time fee plus network fees for flash loans.
Are you able to provide how much new borrowing is happening every month on the protocol for the past 3 months in USD terms? If $500,000 in loans are created every month with a 1% minimum fee then this brings in an additional $5,000 to suppliers and the LQ stackers. A fee of 0.2-1% is very reasonable for a perpetual term loan.
This could also be a dataset to add into the protocol analytics page.

3 Likes
13

Changes I recommend to improve the protocol:

  1. Enable institutional and organisational supply/borrow.
    Organisations, exchanges (centralised and decentralised) should be able to borrow assets with collateral, to enable their users to trade with them. Reduce LTV for borrowers who enable leveraged trading unless we have the liquidation systems in place to handle mass liquidations when the markets are volatile.

@DC1

1 Like
14

After discussions both here and on the Discord governance forum we agree on the following set of changes and will submit these 2 separate proposals, one for LQ rewards changes and one for protocol parameter changes.

Proposal to improve Liqwid protocol: Protocol parameters

Updates:

  1. A 4x reduction in number of action UTXOs from16 to 4 for all markets except ADA, DJED, SHEN and iUSD which will be reduced to 8.
  2. Interest rate parameter updates. Update the utilMultiplier to 8% and the
    utilMultiplierJump to 500. The utilMultiplier change results in a 10.2% Borrow APR at the 65% utilization target (kinkPoint), current Borrow APR at the 65% utilization target is 7.71%. The utilMultiplierJump change results in a 188% Borrow APR at 100% utilization, current Borrow APR at 100% utilization is 91.81% so the proposed update results in a just over 2x the current rates at 100% utilization. The baseRate will remain unchanged at 5%.
  3. Introduce a 1% loan origination fee on all new loans with the mainnet launch of Aqueduct and Tap (with the fee accruing 50% to LQ stakers and 50% to the DAO treasury).

Reasoning:

  1. Reducing the number of action UTXOs allows users to borrow and withdraw larger percentages of assets from the pool per Tx. Currently all markets have 16 action UTXOs meaning each batching cycle users can complete 16 of each action per market (e.g. 16 supplies, 16 borrows, 16 repayments). The problem is in smaller markets that get high utilization rates the max amount that can be borrowed/withdrawn is divided by 16 forcing larger holders to complete multiple borrow or withdraw transactions. Reducing this down to 4 for majority of markets means users can now borrow or withdraw up to 25% of the market’s liquidity (as opposed to current 6.25% max).
  2. At sustained +80% utilization in the bridged stablecoin markets we are still not seeing borrowers repay loans. DAI is currently 84% utilization, USDC is 82% and USDT is 80%. The interest rates < 10% away from the 90% borrow cap still not incentivizing loan repayments points to the requirement of higher rates to incentivize new supplies and loan repayments.
    Screen Shot 2024-02-17 at 4.20.39 AM
    Screen Shot 2024-02-17 at 4.20.39 AM1920×1155 117 KB
  3. Currently Liqwid charges no fee to create CDPs while other DeFi protocols charge CDP origination fees of 1-2.5% and most banks charge 1-10% loan origination fee on personal loans.

*baseRate will remain unchanged.

Proposal to improve Liqwid protocol: Rewards parameters

Updates

  1. Remove the stablecoin target for stablecoin suppliers and individually calculate LQ rewards for stablecoin suppliers using the same method currently in place for all non-stable markets (based on interest repayment amount)
  2. Extend the 20k LQ per month borrow rewards for ADA borrowers to include all borrowers in every market (LQ is allocated to borrowers pro-rata based on interest repaid).
  3. Create an LQ rewards subDAO which reviews market data bi-weekly (interest accrued, interest repaid, utilization, TVL) and use it for tuning the 1. lender multipliers (currently all set to 1x) and 2. borrower-lender split of the 120k LQ rewards total monthly amount, (e.g. updating from 100k-20k LQ split for lenders-borrowers to 90k-30k LQ). The subDAO would not be able to issue more than the 120k LQ monthly max, the only two actions the LQ rewards subDAO control are tuning multipliers and rebalancing the split of LQ rewards.

Reasoning

  1. Stablecoins are varying too much in interest repayments to treat them identically and this will only grow as upcoming Cardano native stables are launched on Liqwid.
  2. Allow lenders in all markets to access subsidized borrowers.
  3. Similar to how the Minswap Farm tiers rebalancing process works based on defined metrics (e.g. 7d USD trading volume for each pool) are used to update the split of MIN rewards across pools, the LQ rewards subDAO team would analyze the latest data on market utilization, interest repayment and accruals and use this to update the lender multiplier and the split of lender-borrow monthly LQ incentives (currently 100k LQ lender rewards to 20k LQ borrower rewards per month).
3 Likes
15
2 Likes