Harvesting Interest Rate Spreads

Prior to the beacon chain I expected that the borrow rates on money markets like Compound and Aave would normalize to something close to the staking reward rate APR. They clearly have not, and are still sitting under 1% most of the time. A surprising amount of capital has been just dumped there and left to sit regardless of what else is going on in Defi. There is $4.5B in ETH on Aave mainnet alone earning ~1% despite common Cefi lending options at 5%, staking APR at 4% (soon to be higher with the merge), various LP options including single-sided LPing with Bancor at 4%. Is Aave really that much less risky than wstETH or rETH? Why accept the lower returns? Is this going to self-correct over time when staking withdrawals are here?

I actually suspect not. I think part of the answer to this mystery at present and increasingly in the future is that other protocols are going to use Aave and Compound as a parking place for idle capital. Bancor for example is launching their “infinity pools” that park idle capital in “strategies” such as Aave. Balancer is following a similar approach. It seems to be the growing alternative to the compressed liquidity ranges of Uniswap v3 for capital efficiency. What I’m getting at here is there is a large pile of ETH sitting in Aave at irrationally low borrow rates because people putting it there have reasons for putting it there at any APR. It seems silly for us to be able to borrow ETH at 1%, sell it for rETH at 4% and profit from the spread but that’s where we’re at.

So how do we profit from this market inefficiency? Up until recently the best you could do to capitalize on this inefficiency was deposit some unproductive asset like UNI or MKR to collect the spread yield on staking income vs borrow rates. The problem with that approach was the liquidation risk of whatever else you’re using as collateral. If the UNI/ETH price crashes enough you lose your UNI while getting at most 3% here. So you up the collateralization ratio to 200-300% which means the actual yield falls to 1.5% or less. Basically, while it was minorly profitable, unless you were pushing around $100k or more of equity with this approach it just wasn’t worth the gas/risk/headache.

Recently though, Aave added wstETH as collateral. This… changes the game. This removes the liquidation risk, allows us to crank up the leverage on that rate spread trade, and go really long on the difference between the staking yield and the borrow rate.

stETH on Aave has a max LTV of 70% and liquidation threshold of 75%. For those not familiar with Aave this means is if you deposit $1 of stETH you can borrow up $.70 of ETH. Using the same flash loan machinery as The Double Logris you can get over a 3x leverage on this in one transaction. This of course means you can roughly triple your stETH yield in one transaction, going 3x long on the wstETH/ETH ratio and capitalizing on the irrationally low ETH borrow rates I discussed above.

For each stETH you have, you borrow 2 ETH on a flash loan, trade it for ~2 stETH, deposit 3 stETH (and earn the yield on it), borrow 2 ETH using your 3 stETH as collateral, and repay the flash loan. For the sake of demonstration, let’s say stETH makes 4% APR and the ETH borrow rate is 1% APR. Your costs are 1% on 2 ETH borrow. Your profit is 4% on 3 stETH. Your net yield is ~10% on your original 1 stETH. Now, when the merge hits and the tips all go to stakers and the staking yield goes to 8% you’d be making 22% APR.

What’s incredible about this setup is it doesn’t at all depend on issuance. All of the yield is denominated in ETH. No costs to claim a spread of tokens and sell them like Convex. No 20% Yearn performance fees to automate it. It’s just clean yield, denominated in ETH, that can be packaged as capital gains in case this actually lasts a year.

As an investment strategy, being long the stETH/ETH ratio and getting paid 10% on that bet is pretty solid as things go. Here are the basic risks though.

1. As more and more capital is attracted to this new arbitrage position the rates will normalize. We’ve already seen the variable borrow rates steadily climb from 0.3% to 1.2% since stETH was added as collateral to Aave. If this spread diminishes as it should then you’ll be better off in my Logris Vaults where you’ll conveniently get double yield on your rETH or stETH doing leveraged farming with Alchemix.
2. Imagine if Lido had a mass slashing event on stETH somehow. Well, the ETH underlying the stETH would be less, so the stETH/ETH ratio (which you are basically 3x long on) would plummet to correct to the new redemption ratio. This event would create a bigger splash in Defi than most think and we’d certainly have more to be concerned about at that point than a liquidation event on $1B leveraged play on Aave.
3. These positions are public and whales could potentially go liquidation hunting on this market. IMO the arbitrage profit fighting against them is too high for this to be feasible, but maybe it wouldn’t be if this market position grows another 10x.

If you accept all that and want to get on board you have a few options at this point summarized in the table below.

Protocol	Leverage	Performance Fee	Withdrawal Fee
DIY DefiSaver	Up to you	0	~.16%+slippage
TokenSet’s ETHMAXY (Deprecated)	3.0	1.95%	0
Index Coop icETH	2.5	0.75%	0.25%
Instadapp Lite	3.0	10%	0.01%

ETHMAXY and icETH are also tokens you can just buy on Uniswap v3. They can be unwound if the liquidity dries up but the process is a little messy since they don’t actually use my flash loan logic but instead go through the recursive process manually.

Now, for you degens out there who just aren’t satisfied by making 10% on your ETH the crazy fucking pirates at Galleon have managed to get their ETHMAXY token added as collateral on Rari Fuse. So, if this 10% yield was empowered by a 4% stake to 1% borrow spread, imagine what you can do with a 10% interest to 1% borrow spread. The obvious math is (borrow liquidity allowing) you could make 10%x3-1%x2=28% APR on ETH. Now, there isn’t $4.5B of ETH available for borrow on Rari Fuse and of course this means you’re compounding platform risk at every step but 28% is just impressive. Post-merge, at 8% staking APR and if the current borrow rates were to hold, this could be looking at 64% APR. Now that’s cookin’ with Defi.