Optimize rate curve for ETH & improve revenue generation

Summary

Aave has seen a massive surge in ETH borrowings since the addition of stETH, allowing users to earn ~4% returns on their staked ETH while still being utilized as collateral in Aave. The utilization of ETH has increased by 4-6x from 3-4% to 18% and is now critical to growing much further under current conditions (explained β€œwhy” below):

Description

ETH’s rate curves in all lending protocols have long been stagnant, considering there has been no massive demand for ETH borrowings. Therefore the ETH rate curves are not as optimized as stable coins rate curves. The current rate curve for ETH is not optimized for the protocol to receive the best revenue, with the inclusion of stETH collateral, the demand for borrowing of ETH could be much higher (~$5B) and more so with accompanying e-Modes, I believe the interest rates need to be optimized accordingly. Optimizing the interest rates for ETH will allow the protocol to get vastly more revenue, lenders will receive better returns, and possibly suck billions of $ worth of ETH into the AAVE protocol.

Staking amount vs staking rates:-

Current stETH rewards are ~4%. Staking rewards on ETH via Lido are still relatively high and stETH still has the capacity to accept another ~$4.5B of ETH src. This increase in stETH supply would only reduce the rewards by 0.3%, meaning there is at least 1.5M ETH (~$4-5B USD) that could get staked. For Aave, it represents $4-5B in borrowings of ETH (more significant than any borrowing market that currently exists across any token)

Problems with current ETH rates on Aave:

Before the stETH addition to Aave:

  • Utilization: 3-4%
  • Borrow rates: ~0.4-0.5%
  • Supply rates: <0.1%

After the stETH addition to Aave (as of today):

  • Utilization: 18%
  • Borrow rates: 2.22%
  • Supply rates: 0.3-0.4%

source for all the data.

Lets determine how high the rates could go with the current rate curve before borrowing ETH against stETH is no longer sensible. One way to calculate this would be to calculate when non-leveraged gains (holding stETH in the wallet) = leveraged gains.

Leveraged Gains Formula

Mini formula to calculate it:

  • Leveraged gains. The maximum leverage scenario here is at 75% (the liquidation limit of stETH, most optimized but no user will go that far).
  • 75% meaning 4x collateral & 3x debt. 4x stETH collateral & 3x ETH debt.
  • Gains = 4 * y (stETH rates) - 3 * z (borrowing rates) = 4y - 3z.
  • Formula: 4y - 3z.

Now for leveraging (borrowing ETH) to make sense, the gains from leveraging should be greater than what we earn from stETH hodling. Meaning, 4y - 3z > y.

If y is β€œ4”, z needs to be less than β€œ4”, at β€œ4”, the benefit of borrowing ETH is 0. I suggest the borrowing rate should be around β€œ3” for leveragers to have good benefits. This would require the rate modal to be optimized in a way where borrowers have enough incentives to borrow, and lenders have incentives for lending.

Considering the situation with the current modal at a 3% borrowing rate:

  • Utilization: 24.5% [very low utilization leading to less revenue generation]
  • Borrowing rate: 3.02%
  • Supplying rate: 0.66% [low incentives for new lenders to come in]

Considering new rates: kink at 3% borrow rate and 70% utilization.

  • Utilization: 70% [good revenue for governance]
  • Borrow rate: 3%
  • Supplying rate: 1.9% [decent for new lenders to come in or to get a discount for stable coin borrowing against ETH]

Conclusion:

Refactor the ETH borrowings rate as the borrowing demand and market conditions of ETH have changed. The improved rate curve will make it better for Borrowers, Lenders and increase the protocol revenue from ETH by at least 3x.

Update:

The proposal is live. Find it here.

15 Likes

100% in support of this proposal. This will drive capital efficiency up and bring more protocol revenue.

5 Likes

Great proposal @samyak .
The diversification/improvement of interest rate strategies per asset is definitely an under-explored field on the protocol and it is pretty clear that in the ETH case after the stETH listing, more important than ever.
Some aspects I would like to comment on:

  • As you said, the introduction of stETH as collateral in Aave is so big ($1.25B at the moment deposited), that is pretty clear that previous assumptions when defining the optimal point of utilization on ETH are not valid anymore. So I’m in full support of increasing that optimal point.
  • It is really important to still keep in mind the nature of ETH. Obviously, I think leverage strategies with stETH/ETH are and will be a big thing from now onwards, but still, there is a really meaningful amount of the ETH deposited used as collateral. So it is fundamental to model which is the optimal point taken that into account, to avoid a situation in which the utilization targeted is too high, causing a lack of availability of ETH on liquidations. I would say there are alternative ways of achieving that availability (at least temporary for liquidations), given the liquidity on the stETH/ETH pair on dexes, but still, really important to include in the analysis and improvement of the strategy.

PS. Aave v3 basically solves/optimizes some of these aspects, so I would say it is a priority to explore the listing of stETH after the Ethereum market is updated to v3.

2 Likes

On the second point, nice one. My suggested point was to target 70% utilization which leaves enough liquidity of ETH for withdrawals and also incentivizes new ETH holders to come in to get good returns on their ETH holdings. On the liquidation side of stETH, thanks to Curve & Balancer which provides the swap of ~$1B at 1-2% slippage. So I think we can count on having enough liquidity in the market in case of liquidations between stETH <> ETH. Src - 1inch

Lastly, if in edge case the price slippage of stETH is 1% then that leaves an opportunity for arbitrage to leverage against and earn crazy returns. Eg:- let’s say with e-mode users can go up to 20x with stETH <> ETH. Then with 1% slippage users have the opportunity to earn straight 20% on their asset in a single transaction as 1stETH is always backed by 1ETH. Considering this we can assume that stETH will never go too low and e-mode will help strongly in stabilizing the oracle price.

2 Likes

This proposal would definitely increase the supply on ETH for the Aave Protocol and bring potentially new users as well, 70% utilization rate seems also in line to give cushion to withdraw ETH and the pool would act as meta pool for ETH staking yield with withdrawals. Since large amount of stETH itself introduces some amount of wider scale of systemic risk (given its only 1 way staking atm), might be interesting also to revisit the RF of ETH and increase it to correspond the overall risk as well.

3 Likes

Pretty good argument, yes. In general, I think could be good to proceed with the proposed configuration, so if the OP thinks this change should be done promptly, I would suggest proceeding with the deployment of the new interest rate strategy with these improvement parameters.
Members of the Aave community (I include myself) can review that everything is correct before proposal submission.

1 Like

Hi All,

Over at Llama, we did some analysis on the affect of changing the ETH rates curves. Check out the new dashboard that has been created. We also added a feature that enables uses to change the parameters and see what affect it has on various pool conditions. Adding stETH as collateral was a big success in terms of revenue generation. This dashboard is intended to show the post affect of the stETH listing and the affect of AIP-68. :slight_smile:

5 Likes