[AIP-26] Alchemix Fuse Pool

RogueItachi

One way in which Alchemix could add another use case with minimal strain on the devs, who are working hard on V2, is for us to utilize the dope product that is Rari’s Fuse pool by creating a custom Alchemix pool. For those that don’t know, Fuse is basically a customizable fork of compound. We could utilize a custom pool to create a new way to borrow and back alUSD. The treasury could mint alUSD, let’s say $5 to $10 million (which would be about 1/60th and 1/30th of the current supply of alUSD respectively), and create a custom fuse pool in which allows users to deposit collateral like ETH, sOHM, and others chosen by the team and community, which would be used to borrow this minted alUSD and alETH once oracles are developed. At the time of mint, this alUSD would be technically be unbacked, yet the alUSD couldn’t go into circulation unless a user deposits the collateral we approved at an overcollateralized level chosen by the community then pays to borrow the alUSD. We would basically be borrowing the DAI model. This overcollateralization would make the alUSD backed at the time of borrowing and would diversify the backing of alUSD which is beneficial to it as a general DeFi stablecoin in my opinion. The borrow interest on the alUSD would insure that people always paid back more alUSD than they borrowed. The Alchemix treasury could also take in the interest of these loans along with any liquidation fees to continue to build the treasury and add more diversity to the treasury. I’m very much of the opinion that a treasury should be as diversified as possible and owning stablecoins are a great way to do that. As we have seen with sOHM in other pools, the demand is very much there for a borrowable stablecoin at reasonable rates, and we could get the collateralization ratio and borrow rates to our liking for what’s best for Alchemix and borrowers alike. There is a possibility that we could further our build upon partnership with Olympus with this pool as well. We can target an 80% utilization rate where interest raises in a gradual slope until the 80% utilization rate, and then goes exponential. Since it's not my strong suit, I'll leave more loan specifics up to the hive mind.

This custom pool would provide alUSD with more utility and spread it across DeFi more greatly. It would also diversify the backing, so it’s not so heavily built upon by DAI. If this is effective, we could add other al collateral types to the pool in the future, and it could be a great way to get additional revenue for the treasury.

Unfortunately, I won’t be around for a bit after posting this, so I will have to ghost my own proposal. I trust you all, my fellow community members, with it.

hekmat

RogueItachi
This is a bit of a stretch and would probably involve a lot more discussion + negotiation from the core team but it popped in my head after seeing the snapshot announcement.

But how would we feel about doing a combined Fuse Pool with another protocol or two? That would mean liquidity for the joint assets comes from multiple communities expanding the reach. It would also symbolically add to partnerships without any real binding commitments at the protocol level. Defi 2.0 projects are frens and stronger together, Olympus would have been an obvious choice if they didn't already have their own pool.

Think D4 saddle pool as a kind of loose analogy. Maybe turn off lending of ALCX and any other governance tokens to make it mildly more difficult to short.

Definitely would be a thought for later even if people do like the idea. No clue how much negotiation it would take behind the scenes from the team and I don't think any of us want to delay v2 for any reason at this point.

LogrisTheBard

What is the advantage to a user to borrowing alUSD this way compared to minting DAI and transmuting the DAI to mint alUSD? Just fewer steps and less gas? A different spread of collateral options than what Maker supports? The point you raise that I most agree with is that the treasury should be diversified to protect the alUSD peg. That is what the devs are focused on most anyway with v2. I dislike the principal of minting unbacked alUSD in any form. I'd rather we just raise the debt ceiling, mint the alUSD using treasury DAI, and use that DAI as the seed money for this pool if we're going to. Yes it's less capital efficient but unbacked alAssets are a schelling fence I'd prefer not to cross.

bostero2635

LogrisTheBard I think the point would be allowing more collateral options to borrow alUSD with little effort from devs. This would lower the barrier-to-entry for alUSD and therefore spread it further into the community. Not saying i’m for or against the idea, but that is the gist of it I think.

ov3rkoalafied

LogrisTheBard I don't think the DAO has enough DAI to mint the amount of alUSD asked for here. However, I'm wondering if we can allow users to deposit their alUSD as liquidity to be lent? Then we wouldn't need to mint as much alUSD (just an initial amount to seed the pool), and users would have another place to use their alUSD - win-win.

scoopy

LogrisTheBard I think one important thing to know is that since the alUSD won't be in circulation until it is borrowed from an overcollateralized position, this isn't quite the same thing as "printing alUSD". For reference, FEI and MakerDAO are doing the same thing with rari and aave respectively.

The advantage over doing this in the alchemist contract is that in Alchemix, only DAI (and later more stables) can borrow alUSD. Whereas in the rari fuse pool, we could have eth, wbtc, and other collateral types, which would turn it into a mini makerDAO of sorts, and expand the ways in which people can acquire alUSD.

FWIW, I'm in favor of rolling this out with a small amount -- no more than 5m alUSD. Then we can collect data on it to see what, if any, side effects it has on the Alchemix system.

Also, it is important to note, that since Alchemix v1 prohibits smart contracts from interacting with it, then it would have to be managed by an EOA instead of a multisig, and that is a whole can of worms I do not want to open.

state

I would explore both the options of creating an alchemix DAO-owned pool with more conservative collateral ratios than the current most popular fuse pools , & also providing to some of the existing pools, personally would prefer the less risky ones like Olympus party versus some of the riskiest ones for Alchemix to mint & lend into.

This would also be a great opportunity to launch a pool with a plugin that stakes ALCX in the background and lets people borrow against their ALCX without losing on the ALCX APR.

LogrisTheBard

scoopy

When we write up the post with parameters be sure to include which EOA owns the deposit to RARI. That's a single point of failure of someone walking away with 5M alUSD.

I'd still rather the treasury just take a MIM loan with it's ALCX, swap it to alUSD and seed the pool that way using the multi-sig. I don't even like the fact that a multi-sig can authorize the creation of alUSD through a back door. The fact that this is possible is worrisome. The fact that people are in favor of actually doing it is madness.

0xFelix

I'm a bit conflicted on this proposal. On the one hand, I think it's a great idea to have a DAO-owned Fuse pool that would give alUSD another use case and that would effectively become another income stream. Depending on the collateral types this could be quite a good move (especially if we include sOHM which has crazy interest fees in the other pools).

On the other hand, even though this alUSD is not technically unbacked (since users need to put up collateral), it is unbacked from the transmuter's perspective. Meaning: there won't be any DAI in the system/vaults to act as a backstop and to guard peg stability. So launching this pool could potentially put a strain on the alUSD peg, since it would take longer to transmute alUSD to DAI.

So it might makes sense to do a trial run with a small amount (maximum 5m, but I would say even less than that), but we need to be conscious that this is not a free lunch and that it could have potential negative impacts on the protocol.

scoopy

0xFelix
I agree with your point, but since these are overcollateralized loans with interest, borrowers will ultimately have to repay more alUSD than they borrowed. Thus, any short term pressure put on the market or transmuter, will be negated by borrowers repaying their loan. In fact, since they have to pay back more than they borrowed, it will ultimately have a positive effect on the peg

long1486

Not sure why do we need this. There is already a great way to borrow alUSD, which is by depositing DAI, and it's interest free, and it's self paying! Isn't that the reason why Alchemix is so attractive? I feel like the current proposals around treasury is getting more and more complicated.

thetechnocratic

I echo the concerns from @0xFelix and @LogrisTheBard.

Having the treasury mint alUSD out of thin air is a bad move.

The idea that the treasury could bypass the Alchemist, mint alUSD, put it into a lending pool, and still consider it "not in circulation" is flat-out wrong. As soon as the alUSD hits the market (ie. enters the fuse pool and affects interest rates) it is in the market. It need not be borrowed to be considered "in the market". Since the alUSD was not collateralized in the Alchemist, it is unbacked from the perspective of the Transmuter, the main pegging module.

When the price of alUSD drops, the Transmuter is looked to as a way to restore that peg. Users purchase and borrow alUSD to exchange it to DAI 1:1. By diluting the lending interest rate, we are reducing this upward pressure on the price of alUSD, thus destabilizing the peg. Having interest paid to the treasury does nothing to help the peg, since the treasury does not sit within the Alchemist-Transmuter loop.

I'm all for creating a lending market and bootstrapping it with alUSD generated via a legitimate loan from the Alchemist, but the treasury should not be in the business of printing alUSD.

scoopy

thetechnocratic

When the price of alUSD drops, the Transmuter is looked to as a way to restore that peg. Users purchase and borrow alUSD to exchange it to DAI 1:1. By diluting the lending interest rate, we are reducing this upward pressure on the price of alUSD, thus destabilizing the peg. Having interest paid to the treasury does nothing to help the peg, since the treasury does not sit within the Alchemist-Transmuter loop.

Hard disagree with this. Since all borrowed alUSD has to be paid back with interest it will actually relieve pressure off the market and the transmuter in the long run.

WRT to the point of it being in the lending market and thus "in circulation", who exactly are we diluting in lending markets when there is effectively no lending market for alUSD at the moment? If more people pile into it, then we can withdraw our alUSD and burn it.

Again, I think scale is the key here. If this is modest in size, max 5m, that is 1/60 of the supply of alUSD, and with the depth of our curve market, it won't affect the peg, full stop.

hekmat

If this AIP goes ahead. What would be the parameters of the fuse pool?

IIRC its up to us to decide the LTV of any given collateral token right? We're 200% and 400% collateralized on pegged pairs within alchemix. I doubt we could entice many people to our fuse pool when they can get far more use out of their ETH unpegged pairs on other protocols/fuse pools. That's not necessarily a showstopper though, I wouldn't have any issues if there was a 120% collateral ratio for ETH-->alUSD on fuse... Rari are pretty degen friendly, most Rari users are well aware of the liquidation risks involved.

I'm all for treasury diversity + partnerships as I've said a bunch in the past. From the general conversation here it does seem like starting with a small amount and seeing how things go would have negligible impact on the peg.

If we move forward considering this an experiment (as others have suggested) I'm easily for the proposal.

reginald-grarbs

A lot of good debate here and concerns. This just makes a lot of sense to me to diversify the backing of alUSD. I am personally uncomfortable with DAI these days. Living in the US and the current regulatory climate surrounding "stable value coins" and DAIs use of USDC is concerning to me. I think it's imperative to diversify treasuries that rely on DAI (unfortunately) and a fuse pool is one way you could do so.

We have no idea if there will be a market for alUSD, so best thing to do is (as suggested) start really small, decide the collateral assets, the collateral rates, utilization, etc. and give this a test. If there is strong demand, this could be a great way (without too much burden) to diversify the treasury and backing of alUSD and as Scoopy points out, the net effect (given borrowers will have to pay back more alUSD than their loan) is positive.

I would vote in favor so long as this is a small step, test the market, gather some info and then push further if successful.

thetechnocratic

The treasury printing alUSD is a bad idea.

It does not "relieve pressure off the market", since it dilutes the market, and thus adds pressure to the peg. The argument for creating a lending market when there is none falls flat, because if the market wanted this, it would add more alUSD to the current fuze pools.

The only perceived benefit here is filling the treasury with funds from interest made on lending. Again, THIS DOES NOTHING TO STABILIZE THE PEG. It only dilutes the lending market, which in turn destabilizes the peg. The treasury should not be printing money to make itself more money. This circumvents the protocol as it was designed, and breaks system invariants.

If you want to take a page out of the fed's book and turn on the money printer, then go ahead and vote yes, but the perceived benefits are not worth the risks.

ov3rkoalafied

thetechnocratic If there is no demand, can we close the pool and burn the alUSD? And then if there is demand, can we remove our alUSD and burn it? Ie, it's more of an alUSD loan from the protocol to itself? Or a temporary "seed" for this pool. Then the protocol would not earn revenue, would net have created 0 alUSD, and the pool would exist as a new alUSD sink.

hekmat

Good to see a strongly dissenting opinion and fair point.

It would be terrible for the dao if we turned into a yes-man circlejerk.