Summary
We propose to modify Notional’s external lending logic to direct interest earned on unutilized supply funds to the protocol itself instead of variable rate lenders on Notional.
This will make Notional significantly more profitable without impacting the attractiveness of lending on the protocol due to the protocol’s ability to generate yield for lenders via borrowing.
Background
In the Notional V3 introductory blog post, we described the Prime Money Market as a variable rate lending market where lenders would earn interest from borrowers on Notional and external protocols at the same time.
We call this activity of lending unutilized supply funds on external money markets external lending.
External lending is not live on Notional V3 today. Unutilized supply funds are held in underlying on the Notional proxy contract. We propose to activate external lending but to retain the interest that it generates for the Notional protocol instead of directing it to variable rate lenders.
Motivation
We believe this is the best course of action for the following reasons:
-
It makes Notional significantly more profitable. Increased gross revenues and margins will help Notional grow and generate value. As of now, the changes we suggest would increase Notional V3’s current revenue rate by 84%.
-
It improves UX for users. Returning yield to lenders makes variable rate lending and borrowing on Notional very complicated and it becomes difficult to judge what variable rates could or will be in the future. This change makes variable rate lending and borrowing on Notional simple.
-
It is easy to implement. The code changes required are minor and don’t touch the core protocol.
To be clear, there are downsides to this change. Most notably, this could result in a less compelling proposition for lenders and LPs. Specifically, yields will be lower for lenders and LPs during times when utilization of the variable rate lending markets is below kink2. This means that providing a healthy yield for lenders will rely much more on generating borrow demand instead of using external lending as a fallback.
One of the reasons we are in favor of this change is that we think we will be able to drive borrowing demand. We can see empirically that Notional V3 on Arbitrum has been able to generate significant borrowing demand. This indicates that the protocol can reliably generate a baseline level of borrowing demand which removes the primary risk that external lending was originally meant to protect against.
Furthermore, if we can’t generate borrowing demand then we’ve already lost and making up the yield via external lending won’t save us. If we can’t generate borrowing demand, we won’t be able to offer higher yields than older and more established lending protocols. If we can’t offer higher yields than simpler protocols, our value prop is invalidated.
So on balance, we believe that this is the right decision for the future of Notional.
Proposal Details
Specifically, we propose to lend unutilized assets in all currencies up to the second kink of Notional’s variable rate utilization curve. For clarity, the second kink is the point in the utilization curve where the lending and borrowing rates start to spike. For most currencies on Notional this currently occurs at 70% utilization, but for some it occurs at 80% utilization.
So for example, if the USDC lending market were 50% utilized with 100 USDC in total supply and the second kink at 80% utilization, this proposal would imply lending 30 USDC externally and keeping 20 USDC on the Notional proxy contract.
The purpose of the second kink is to ensure redeemability for lenders as well as to ensure that Notional holds enough underlying cash to be able to promptly perform liquidations in the event of a decline in collateral values. By only lending on external markets up to the second kink we do not compromise on either of these points and maintain a conservative risk posture.
External lending would be subject to a few constraints:
Eligible external lending markets
We will initially use Aave V3 as the exclusive external lending market. Aave V3 is the most battle-tested and least risky variable option available. It has the added advantage of supporting almost all of the assets that we would like to lend across every chain that Notional is active on. This means that we can scale our external lending activities across all chains and currencies using a single venue.
In the future, we might want to use different yield sources for different assets to increase our yield or diversify our risk exposure. But for now, a single yield source saves us time and resources. Adding new external lending venues would have to involve vetting, development, and governance approval.
Interest rate threshold
We would propose to use an interest rate threshold to determine when we should lend and when the reward is not worth the additional risk. If the lending rate on an external venue is above this threshold we would lend, and if it is below this threshold we would not lend and redeem any funds that we have lent. In practice, this threshold will not necessarily be enforced on a rigid, automatic basis. A tentative value for this interest rate threshold would be 1% APY.
Revenue implications
If applied to Notional V3 on Arbitrum today this change would increase the protocol’s annual earnings by 84% based on current market data.
Here is a snapshot demonstrating how this change would increase Notional’s revenues:
This change would almost double the protocol’s current revenue run rate. We can’t be sure what effect this change would have in the future as the protocol scales and borrowing is driven more by leveraged vaults and less by harvesting liquidity incentives. But we do know that in the near term it would allow Notional to use its assets more efficiently and make Notional significantly more profitable.
Implementation
The code necessary to implement this change has already been written. Upon approval of this proposal, the code will be tested, audited, and deployed. Once the code has been deployed, Notional would phase in the change over time starting with one or two currencies and eventually expanding to all eligible currencies.
We would expect the time from approval of the proposal to full implementation would be roughly 6-8 weeks.