Alchemix Request For Comment (ARC) #1

scoopy

The Alchemix dev team is nearly complete with the expansion to add an alETH vault, transmuter, and farm. This document is Scoopy’s vision for how to launch alETH, but comments/suggestions for improvement are encouraged.

Alchemix Vault Parameters: Collateral Ratio: 400%

Current yvWETH yields are 3-4%. A 200% collateralization ratio would mean loan repayment times of up to 10-15 years. At 400%, this would be lowered to 5-7 years. Still a long time, but much more reasonable than the alternative. If ETH yields increase in general, or are boosted via the transmuter, this value could be revisited.

alETH Debt Cap: 2000 alETH

It is important to distinguish that this is NOT a supply cap for alETH, but rather the maximum amount of alETH that can be minted via debt. The supply can expand when people repay their loans using ETH, much like how alUSD has a 360m supply with only a 100m debt cap. By setting the supply cap on the lower side, we can take a conservative approach, limiting the system’s alETH debt to not exceed 2000 alETH. This limit can be increased at a later time if it makes sense to do so.

Eth to alETH direct Purchases

The v1 contract already has a roundabout way of trading DAI for alUSD 1:1 via the liquidate function. We can add a convert ETH to alETH function, which executes the trade 1:1 ETH for alETH, sending the ETH to the transmuter. In doing so, it will boost yields in the same way that the DAI transmuter does.

Farms: Adept mode

Adept mode is the next evolution from mage -> sage -> adept

ALCX/ETH SLP
Sage 60%
Adept 58% (-2)

ALCX
Sage 20%
Adept 19% (-1)

alUSD3CRV LP
Sage 18%
Adept 18% (+0)

alUSD
Sage 2%
Adept 0% (-2)

alETH/ETH LP
Adept 5% (+5)

I propose we add 5% of weekly rewards to the alETH LP pool. Some quick calculations show that the alUSD-only pool at 2% attracts around $70m in liquidity. Normalized to 5%, that is $175m in expected liquidity. Assuming an ETH price of $4000, that would mean a pool worth 43,750 ETH, which would be roughly split between ETH and alETH. If only 2k alETH and 2k ETH can enter this pool without liquidations or direct trading for alETH to ETH (4k eth total), then yields for stakers would be considerably high. Such yields will attract more stakers into the curve pool until it normalizes along with much of the rest of DeFi yields. The end result is that there will be a tremendously beneficial ratio between Alchemix vault ETH deposits and transmuter ETH deposits, which will boost yield considerably for debt repayments.

We’d take some rewards from the Sushi and ALCX-only pools to fund the alETH/ETH pool, but only slightly so as to not disturb the stakers in the protocol and their expectations. The alUSD-only pool was only ever intended to be a limited pool, so cutting it altogether and redistributing it to the alETH/ETH pool seems reasonable now, especially considering the ~$600m of liquidity in the alUSD curve pool.

Final Note
This post is an open call for the community to comment on this launch plan for alETH. Input from the community is super important and we will take feedback seriously. Please indicate your sentiment of this plan and make suggestions for changes. Based on that feedback, I will create an AIP to send to snapshot to make the initial parameters of alETH official.

ChickenBunz

Well balanced proposal. Approve. The long repayment times should definitely come down post the merge as ETH yields increase. This is going to be a great instrument.

atlas-ship

Curious what is the justification for not favouring a long repayment time on vaults? Surely it’s a net positive for the protocol to lock up Eth for decades?

If anything, I see ethereum yield being far more viable long term than stables. When the bear hits, stable yields will likely crater, but ethereum yields have a floor associated with eth2 rewards.

daveytea

Agree with atlas-ship's comment regarding bear market yields, with eth2 staking resulting in a 'floor' yield.

400% collateral ratio seems too conservative as a start IMO, it should be closer to the current DAI coll ratio to reduce mental gymnastics for new comers and to create a sort of new listing standard. E.g. I deposit X in alchemix, I get an un-liquidatable loan of 50% of the value, no matter the original asset.

As a side note i thinkcollateral ratio may not be an accurate term, since it is used widely in lending protocols, whose positions can be liquidated at a certain ratio. Alchemix is different and special, so something such as deposit ratio may be more accurate and friendly.

scoopy

The worry with a lower collateralization ratio is that the transmuter won't get fed enough, and in the event of a peg destabilization event, the peg could recover quicker if the transmuter was seeing better flows. I'm still in favor of 400% to start, and then if the base yield and the transmuter boosted yield is good enough, we can turn it down to 200%.

It's MUCH easier to start conservative and then loosen up later than to start loose and then try to be more conservative later.

atlas-ship

scoopy I understand the economics for maintaining liquidity for alUSD and alEth pairs will be a different ball game. The likelihood of peg destabilisation in alEth should be higher. Imo I think it's better to charge a higher yield fee, say 20% for eth vaults as a means for stabilisation, instead of increasing the deposit ratio. Ultimately, Alchemix doesn't want to discourage depositors. AUM is the bread and butter- locking eth for long periods is beneficial for ALCX holders

boyeatzworld

I agree that the 400% is a bit restrictive. I also think that at that cap ratio it sorta punishing for those without fatter stacks of ETH. A few more considerations:

If this is something that could change quickly post launch then we're talking about eating another potentially high gas fee to get additional alETH. Not the end of the world but again sort of a tougher pill to swallow for smaller depositors.
- There's a potential tension between lowering the cap ratio and the alETH mint cap. If this is as popular as many thing it would be would we have to potentially increase the mint cap in order to lower the ratio?

Maybe we can split the baby here and consider 300%. Slightly less punishing to those with small stacks while not being overly ambitious given the lower ETH yields.

msulc

In my opinion, both keeping the peg and achieving a boosted APY are key features. That's what make Alchemix great and unique. A user can get their future yield in advance and also they receive more APY than yearn on their collateral thanks to the magic of the transmuter mechanism.

So I would also initially vote for a higher collateralise ratio than 200%, 400% seems fine, so we can feed the transmuter faster.

What I'm not sure is if a 5% weekly rewards will be enough to incentivise users to provide their ETH, so I would vote to go a bit higher, maybe 6% or 7%, and reduce it from the alUSD3CRV.

Either way, it would be great if both metrics (collateral ratio and pool rewards) can be easily updated afterwards.

feelingoodfeelingreat

msulc great input from msulc, i'm in agreement with both comments. For the suggestion of the reward ratios, here's the outline with 1% less to the ALCX pool and 1% more to the alETH/ETH LP.

ALCX/ETH SLP
Sage 60%
Adept 58% (-2)

ALCX
Sage 20%
Adept 18% (-2)

alUSD3CRV LP
Sage 18%
Adept 18% (+0)

alUSD
Sage 2%
Adept 0% (-2)

alETH/ETH LP
Adept 6% (+)

Start here and adjust in the future as needed to increase liquidity.

SathFenrir

Two thoughts:

400% is indeed a bit high.

We should take between 2 ~ 5% total from the ALCX single stake for the alETH/ETH LP. 1% feels too low. Providing LP is a far more useful service at the moment than single staking for rewards.

Reeserj

Thanks for putting the the proposal forward!

Vault Parameters: Collateral Ratio: 400%

I'm of two minds here

A longer term payoff is scary in crypto mostly because of how fast crypto can change. Locking up ETH for 15 years is hard to really wrap your mind around. (ETH isn't even that old!)
But I think once the merge happens we'll see a "risk free" rate of ₇% (Likely more since the tips will be going to stakers) and that would almost require other protocols to increase their yields above that threshold because otherwise it's not worth risking the ETH, that should reduce the repayment time quite a bit.

Limiting the risk initially I think is a good idea, but I think some kind of threshold for when the Collateral Ratio (Or Deposit ratio as @daveytea said) can get to parity with DIA is important. I also think that letting users see how long it's going to be until their loan is paid back would put that into their hands. Is the user ok with not having their ETH paid back until 2036? if so, that's probably ok. If not, they won't use it.

alETH Debt Cap: 2000 alETH

This is great, I think this limit will be reached very quickly, but that's ok. Better safe than rekt. Again though if there was some set (hopefully publicly available) metrics that needed to be reached to increase this cap. (Time, Volumes, liquidity, etc)

Eth to alETH direct Purchases

Would it be possible to use the spare ETH in the Transmuter to put into the yvSTETH or yvSETH vault to juice the returns some? I know you mentioned that may introduce some slippage, but I think I'd agree with @squidgle that it'd be worth the extra dev work to allow users to take advantage of the increased yield

Farms

I think I'd be in favor of reducing the single sided ALCX by 5% and adding it to the alETH/ETH Pool as suggested by SathFenrir

KW710

My only concern with the long lockup period is that it may disincentivize people who want to park their ETH during the bear cycle. If they feel that there's a chance the next bull rolls around while their position is still maturing, they may hesitate to deposit it at all. My assumption here is that their goal is to avoid selling their ETH at all costs, so liquidating part of their collateral would be off limits for them.

Pirateed

Just curious, why use the yvETH vault instead of the crvSTETH vault? It would boost yields significantly and mitigate the leverage & paydown timing issues.

scoopy

Pirateed the bookkeeping require for going to eth -> stethcrv LP to yvstETHCRV vault could affect users with slippage and would require a longer dev cycle to ensure it is safe. We were planning to go with stethcrv but the zaps to get in and out could make people incur losses too. Using the yvWETH vault means we can implement our current contracts with minimal changes. So if we choose to go with stethcrv LP instead of eth, it means people would have to deposit stethcrv LP instead of eth, or we'd have to delay this a few weeks at least

squidgle

scoopy i think the small effort up front to gain almost double the APY is worth it when you're talking years to get the return back.

even if that effort is put on the user to acquire the stethcrv themselves. zaps could come later?

Gravity

scoopy will v2 bring the ability to implement/deploy too multiple strategies as yearn v2 vaults do?

Re. Above discussion.
As a long term hodler of ETH (in various forms) I would not be deterred by longer maturing periods. Ie. 200% vs. 400%. I would obviously prefer 200% over 400% as my goal for many years has simply been to accumulate more ETH, which this ratio would compliment. On this same note, a 200% ratio will imo, enable greater participation in the ETH:alETH pool.

To an earlier point, an ability to borrow 50% as a meme is powerful. Simplifies UX mental models and if an eta maturity date is prominent enough, should not impact adoption negatively, especially if it is known that rates will improve over time.

jamesk

scoopy it would be good if we could chooae the best "flavour" of ETH available (meaning the highest APY). It would be very attractive to newcomers

aLeo

scoopy

Data maxima venia

alETH v1 vault with 400% collateralization and low APY of 3-4% is a less-than-optimal product and I think we all know it.

If I deposited my ETH on any Alchemix vault today, I expect it to be on par with the best yields available on DeFi.
stETHcrv is rewarding 11,44% APY -- if we can't match that, no matter the collateralization settings debate, I don't think it's worth launching.

By Alchemix launching this iteration I don't see any angle that this would benefit the community, au contraire, I can see it only tarnishing it's pristine branding.

Consider that Alchemix is already a very hard to grasp concept for the average user.
We (I mean you guys) are working very hard to educate users and present the benefits of utilizing ALCX vaults.

Even amidst all the confusion happening in normies and no-coiners minds, in my perspective, there are certain aspects about the actual v1 DAI vault that allure as crucial concepts of Alchemix (as a service provider), in simple terms:

user deposit X (100%) -> user can borrow X/2 (50%) upfront
meanwhile, 100% of deposited amount is kept "working" in DeFi
loan pays itself as fast as DeFingly possible

I think those are more than settings. To me they are fundamental characteristics and a piece of the branding concept of Alchemix. This is already being memefied amongst the web. I don't think we should give up on them in favor of shipping a v1 ETH vault fast.

In my view the launch of alETH with low APY and higher collateralization would only contribute to further confusion and unhappy users. Or even worse, the market could dislike it at large and label us as a flop, or a one-time product company.

If we decide to launch vaults with superior collateralizations and lower than best APY's, the end user will get further confused and letdown.
This may be worth doing in certain cases, like with alBTC maybe. I don't think this is the case for alETH at all.

I would rather delay the launch and consider stETHcrv for higher APY or forget about it entirely and focus on V2 where we could chase better yields wherever they are. Last not forget that ETH has EIP-1559 and the merge are happening soon, and with this, ETH as an asset is expected to deliver much higher yields and those should be harnessed by our vaults.

Also, if the APY (on launch) of any vault is higher than 10%, I would not consider changing the collateralization ratio (would keep 200%). If the APY by any chance lowers after launch then the user can always self-liquidate to retrieve its principal.
Although I think it is prudent to keep alETH mintable ceiling low and increase it slowly (as it is already being proposed.)

TL;DR - postpone alETH vault in favor of higher APY chasing (with stETHcrv or wait for V2)

AcedaBook

I'm in favor of conservative 400%, but would suggest a higher % of ALCX weekly rewards at the start. alUSD rewards can certainly be removed, but a small reduction in the alUSD3CRV pool would be merited to for the early ETH depositors. 2-3% higher on ETH would really bootstrap the new pools

ElB

400% is not ideal, but I agree it's better to start conservative and roll down once we see usage and understand the impact of doing so. Biggest callout is on the weights, because of the 2k cap, alETH/ETH would have a more favorable APY than other pools. I recommend the below, which still incentives capital into the new pool but makes alETH/ETH LPing closer in APY to SLP:
ALCX/ETH : 60% (0)
AlCX: 18% (-2)
alUSD: 0% (-2)
alETH/ETH: 4% (+4)
3crv: 18% (0)

Pirateed

Thanks for the explanation Scoopy. Given it's an initial rollout, I'm supportive of the proposal. The cap is lower than I would have hoped for but can understand why. I see this as the product of the future to drive ALCX growth and want to ensure there's enough incentive here to pull people away from higher LTV debt projects. I'm not sure a 25% LTV + 5 year paydown schedule is enough but can recognize it's the first iteration.

rjrudolphjr

Thanks for all your work scoopy. This is such an incredible tool.

Given the ease through which someone can use leverage to obtain a desired level of ETH price exposure, it seems to me that much of the demand for ETH collateralized borrowing might be generated by tax management strategies. Since borrowing through Alchemix offers the user the ability to liquidate as they see fit, I think it might make sense to prioritize the collateralization ratio over the expected debt repayment date.

I've know quite a few people (myself included) who are excited about the alETH development and are considering the following strategy:

Deposit ETH as collateral
Borrow 50%
Convert 50% of the borrowed collateral to USD / Convert 50% of borrowed collateral to ETH2x-FLI
Cash out USD

This works at a 200% collateralization ratio but becomes increasingly less effective the higher the ratio goes.

Additionally, for those who are willing to accept a lower net ETH exposure, a higher collateralization ratio decreases the utility of borrowing through Alchemix. Since people would have to pay taxes on the income earned that is paying down their debt, the tax burden of larger and faster income (which at least in the US would be taxed at ordinary income rates, not capital gains) may get to a level where it outweighs the consequences of simply selling for USD.

My only other comment is that I can understand wanting to have less of the rewards allocated to the alUSD farm. That said, I would advocate for lowering, but not eliminating it. As people are still getting comfortable with what Alchemix has to offer, allowing newcomers or people with a lower risk tolerance to participate in the Farm with alUSD might help generate more confidence over time as well as keep more capital within the ecosystem.

Thanks again for all your effort.

Pronoia

I admire the conservative approach on more volatile asset could it become a standard for other large caps in the future? What is the concern for limiting the system’s alETH debt to not exceed 2000 alETH?
It seems some would prefer to ape into a 10 year debt, that they can forever extend, But that's so fiat.
How do we incentivize 2-5 year loans and feed the transmuter to boost yields.
Maybe giving the option for lower collateral ratio with a deposit fee of 2%.
Limited debt + boost yield = Alchemix is to me.

Gravity

Pronoia you could always suggest auto liquidations after 5 years. But I’m not sure that go down very well.

i personally see our debt positions in Alchemix as cash flow generators, not debt obligations.