Remove CREAM from the Idle RAI aggregator. This would make the aggregator deploy capital only in Fuse Pool 9 and Aave v2.
Problem
The Reflexer community expressed their preference to focus RAI capital in protocols such as Fuse and Aave. Currently, the Idle aggregator tends to allocate too much capital in CREAM. Many RAI lenders mentioned that they are not comfortable with CREAM’s protocol risk, even hinting that they will stop using Idle RAI altogether if the situation isn’t improved.
Given that Idle governance just voted IDLE rewards for RAI alongside the current FLX rewards given to depositors, I believe we should strive to keep depositors happy and hopefully attract even more capital.
Solution
Remove CREAM from the RAI aggregator and only leave Aave v2 and Fuse Pool 9.
Benefits
Attract and retain people who prefer to lend their RAI only on specific protocols such as Aave v2 and Fuse.
Potential Obstacles
What are the risks of removing a lending protocol from an aggregator that’s in production?
Next Steps
If the community agrees to proceed with this initiative, I can post an on-chain proposal by 22nd of August.
Idle is working as expected (it’s optimizing the interest it receives, in RAI terms) but is in an usual “cartel” regime right now due to the FLX subsidy to RAI lenders. Basically, Idle is engaging in automated price-gouging of RAI borrowers because the borrow rate curves aren’t linear and because “lending” to Cream is an option although it is essentially just withholding inventory.
Right now lenders receive a higher rate on their RAI by lending either directly to Fuse or Aave but for the most part aren’t doing this because the FLX subsidy cancels out the increased rate on RAI.
The crux of the problem, in my opinion, is that the risk of keeping funds in Cream isn’t being properly accounted for. I think most people can agree that the chance of funds in Cream being lost is significantly higher than 0% (Cream has had several instances of bad debt already) and funds being held there are actually at higher risk of being lost than RAI which is simply unallocated.
Removing Cream is probably the easiest way to fix this. The other possibility would be to include a risk factor when allocating to Cream. (e.g. 1% chance that funds will be lost if deposited into Cream for 1 year).
You got the point, basically lending in CREAM now is equal to withholding inventory, and given that Idle is one of the largest lender for RAI, it’s basically done to push the lending rate (and the borrow rate as a consequence) of AAVE and Fuse higher
Currently RAI it’s included in the Best Yield strategy so risk is not considered in the algo as it’s in the Risk adjusted strategies but I agree that Cream historically had more issues than other protocols like Compound and Aave.
Your concerns are valid and I’m keen to hear the opinion of our community as well on this. @stefan a snapshot poll can be setup following these guidelines → [Guide] – Community Governance Process (We recently edited it to add support for stkIDLE holders, Basically 2 snapshots should be created one for IDLE holders and one stkIDLE holders)
Thanks @dr-stone for sharing here these concerns and possible mitigation.
For the idleRAI pool, Idle protocol is managing most of the lent assets, with the ability to further optimize the utilization ratio. The algo is now able to increase the overall market returns and all the lenders can benefit from it.
The mechanism that @ghost has described is true and the concern is definitely valid.
The issue is not Cream as a lending source, which has been approved following our Integration standard Requirements and has a risk profile similar to Fuse.
A future algo improvement might consist in enabling the holding of unlent funds if the returns can be higher; this would further reduce the risks while continuing to maximize the profits.
In the current circumstances, the proposal to remove Cream from idleRAI seems an actionable plan.