Update the yETH vault strategy to mine UNI in the ETH:DAI pool
Should this proposal be implemented, the yETH vault will update its strategy to use ETH as collateral in Maker to generate DAI, and then provide DAI and ETH into the ETH:DAI Uniswap pool to mine UNI. Half the UNI will be sold for increased yield, and the other half will be used to build a vault for Uniswap governance.
Currently the highest risk vault in yearn.finance is the yETH vault. The vault takes deposits of ETH and puts them in a Maker CDP vault, mints DAI, and then deploys that DAI to become working capital and provide a return on investment. However, this is an inherently risky venture, as it takes out a loan of stable value (DAI, theoretically) against a volatile asset as collateral (ETH). Should the value of the loan exceed the value of the collateral, this will cause a liquidation of the vault and generate a loss against the depositors. Furthermore, DAI is frequently in higher demand than other stablecoins, leading to a potential DAI crunch and an inability to unwind the loan. Currently, the yETH vault is closed for deposits, with an unknown date for reopening.
Should this proposal be enacted, the strategy would be updated to mine liquidity on Uniswap. The detailed schematic is below. The ETH would be used to open a Maker vault to generate DAI. An amount of ETH of equal value to the amount of DAI generated would be held separate. These assets would then be generated to Uniswap’s DAI:ETH pool to mine UNI. Half of this UNI generated would be held in reserve to be used in UNI governance for the betterment of the yearn ecosystem. The other half of this generated UNI would be sold for ETH to compound the return on the vault.
This methodology has notable improvements to significantly reduce the risks associated with the Maker vault. First, by holding an amount of ETH outside of the Maker vault, we have a supply of collateral that we can use to add to the vault in case of a drop in the value of ETH. Furthermore, by utilizing the DAI:ETH Uniswap pool, we have multiple benefits. In the case of an ETH price drop, the pool would sell DAI and accumulate ETH to supply to the vault - same story in case of a DAI price spike. With this strategy there will be far less risk of a DAI crunch - we will always have some available to unwind the loan in this Uniswap pool.
In the unlikely case of maximizing the Maker vault, we can follow the same logic for other lending platforms. By supplying ETH to Compound, for example, a stablecoin (whichever is cheapest) can be borrowed, converted to DAI, and supplied to the Uniswap pool as well. Generated COMP would then be sold for ETH to compound the value. This second strategy would only be used in the case of profits being greater than the loan rates, and if the Maker strategy is maximized.
For: Use the above strategy to mine liquidity on Uniswap in the yETH vault.
Against: No change.
I promise I’m done proposing strategies for a while
Please let me know thoughts and opinions. I know there’s another yETH vault strategy being worked on behind the scenes, but I wanted to get this proposal out there while the UNI farming is new Yearn has the greatest chance of gaining traction in Uniswap governance.
good strategy, I will vote for.We need quickly change.
@mattdw I believe we spoke a little about this strategy on Discord. Great to see it here fleshed out more. I voted For but have a question: Who holds the Uniswap governance tokens (the half that is not sold and reinvested into yETH)? Don’t those tokens belong to the ETH depositors?
Hey @Matternhorn, we did! My idea was that we’d hoard the UNI tokens in the same way that the yCRV vault uses CRV tokens to boost their deposits. By having the vault collect the tokens and keep them, we could engage in governance and influence the protocol to be more valuable for our uses.
Should the vault decide to just sell them all and liquidate the holdings (in the case of deciding Uniswap isn’t worth the hassle), then we could do that too. Or if folks prefer, we can just do the grim reaper thing and sell all the tokens we collect.
While I do like this idea in theory, and I think it’s an unlikely situation, my biggest issue with this vault is the possibility of ending up with less ETH than you put in– if ETH moons while you’re in the UNI LP, your LP is worth more, but a good portion of your ETH will dump into DAI.
One of the core tenets of vaults is that (barring the withdrawal fee) you will never end up with less of the underlying asset than you put in. Again, I don’t think it’s a likely situation, and I imagine this strategy would generate good yield, but it’s just something to consider.
I agree, this is the biggest risk of this vault strategy. I also agree that this scenario is unlikely, but thank you for bringing it up.
I can think of one possible option to alleviate that. Instead we could put ALL of the ETH into the Maker Vault, and then swap half the generated DAI for ETH (essentially having the entire strategy on leverage). This would provide less working capital (and therefore less ROI), and potentially re-introduce a risk of a DAI crunch. If ETH dropped in value, you’d sell DAI to buy more ETH - which means you’d be less likely to be able to unwind the vault, though I imagine that’d be a small risk.
Alternatively - make it a completely unique vault whose entire purpose is to mine UNI for the rest of the ecosystem? I’ll have to think on that some more. Suggestions would be very welcome!
Yeah, I know that “Vaults” have the tenets of not risking initial capital loss, but maybe this could be a form of the YOLO Degen Vaults that the team has discussed before.
It’s far from being actual YOLO, but with a chance of IL hurting your ETH if you withdraw at a bad time, we would probably want it in a separate interface or something like that, just to keep the brand consistent.
I think this should be a different vault, as there is a risk of gettting less of the underlying (ETH). That idea seems to be the main aspect for the original vaults and strategies. If this strategy could introduce that risk, it should be a separate type of Vault. In fact, Pickle Finance is already doing this: https://pickle.finance/jars
Perhaps we should identify them as something other than “vaults.”
Great to see this proposal. I believe that Uniswap is def worth the hassle. It is going to be massive longer term and dovetails nicely with Andre’s offer at yuni.finance to act as a delegate to tally enough votes to pass proposals within UNI Governance to help the Yearn Community have a voice there.
It is also my understanding that stableCredit is being built on Uniswap, so a good working relationship with Uniswap going forward is paramount.
Absolutely. I already delegated my UNI to @andre.cronje, and we could have the farmed UNI from this strategy also feed into that delegation.
I’m not sure if anything has moved on this over the last week, but wanted to suggest that yearn absolutely go into more risky strategies so long as its clear to the user that LESS of their principal investment may be returned when they withdraw. Have there been any discussions around this?
I don’t support any IL vault strategies at the moment, and as far as I know none are being worked on or seriously considered. But I do recognize your point, that perhaps in the future there could be vaults with IL strategies, as long as the depositors are made aware of the risks and the Vault itself is not unnecessarily risky. Obviously some discretion used in determining that.