[Signal Requst] Adjust UNIV2DAIETH-A Debt Ceiling

[ultraschuppi]

Overview

Since the launch of UNI2DAIETH-LP token as a collateral the utilization has been running pretty good; we just hit the Debt Ceiling.

We started with a pretty conservative DC of 3 MM, I guess we can start increasing it now.

This poll is about raising the DC to allow more DAI from UNI2DAI-ETH.

Pros:

• more fees collected
• more DAI in circulation

Cons:

• of course we are raising our exposure to ETH-volatility here, but since the token is backed 50% by DAI, this is probably not a big deal

Ceiling Adjustment

A few options with +50%, +100% + 200%, +300%

• 3 MM (no change)
• 4.5 MM (+ 50%)
• 6 MM (+ 100%)
• 9 MM (+ 200%)
• 12 MM (+300%)
• Abstain

Next Steps

Poll will run until January 14th and depending on the result will move on-chain assuming the outcome the poll deems it necessary and both Risk Domain Team and Oracle Domain Team support the increase.

[jernejml]

@NikKunkel

The decision on debt ceiling is ultimately up to the Risk Domain Team, but the Oracle Domain Team is recommending an initial debt ceiling of 3M Dai given the enhanced Oracle risk.

@monet-supply

Summary Proposed Risk Parameters

Risk Premium: 1%
Liquidation Ratio: 125%
Debt Ceiling: 10 million DAI
Auction Lot Size: 50,000 DAI
Minimum Bid Increment: 3%
Bid Duration: 6 hours
Max Auction Duration: 6 hours
Liquidation Penalty: 13%
Dust: 500 DAI

Current low debt ceiling is because of oracle risk. We should wait for Nik to provide update.

[smaugho]

@jernejml

Current low debt ceiling is because of oracle risk.

Could we cost an external audit to increase our confidence in the oracle? LP-tokens are a very big source of income for us (almost guaranteed IMO), the Risk Team requested us to increase the DC only once we feel confident with the current oracle (and its risks), but how do we know when we’ll feel confident?, I don’t think we can define that, and if we measure that in time, that’ll be 6 months, or 1 year.

IMO, if we lack confidence here, we should invest in auditing the contract by 3rd parties (maybe 2 independent audits?). With the coming UNIV2USDCETH and ETHWBTC vaults we should be prioritizing this to don’t limit all the new vaults to 3M (which will not help us that much)… plus if an issue is found in one of them, I guess that it probably exists in the others (which means that we should then really have 1M per each one, to keep the 3M which can be covered by the Surplus?).

[iammeeoh]

@smaugho I totally agree!

Given the potential size of these collaterals, investing quickly on external audits (2 or more!) seems a pretty much obviously sound investment.

[Malgaph]

I hope we can get that DC boosted, there is a definite demand for leveraged yield farming in liquidity pools!

[psychonaut]

Assuming price feed is solid, how about a voting option to hook up the MIP27: Debt Ceiling Instant Access Module ?

[jernejml]

@smaugho

IMO, if we lack confidence here, we should invest in auditing the contract by 3rd parties (maybe 2 independent audits?

I don’t think we actually need formal audits (from well known companies). It’s not that big codebase, imo. Maybe paying 2 or 3 reputed developers familiar with uniswap v2 and maker code would be better. Now, how many such developers exist? Are they freelancing (or working for competition)?
Basically, how many engineers like ‘Lev’ are there?

(i am probably looking at this too naively?)

[alexis]

I don’t think the code is about uniswap, the feed is based on a formula plus existing feed, the formula is well known. What is unknown is probably all newish thing that can go wrong and the better to be safer. May be we can just wait to hear what Nik has to say.

[ultraschuppi]

@psychonaut

Assuming price feed is solid, how about a voting option to hook up the MIP27: Debt Ceiling Instant Access Module ?

Would totally love that. I guess we are better off adding another low DC vault type to it instead of applying it to ETH-A?

[NikKunkel]

Hi all. I’m really happy to see the DAI-ETH LP getting some traction, especially given that:

1. BTC and ETH have been on a tear leading to an expected decrease in demand for being an LP
2. No easy one-click leveraging solutions through Zapper / DeFi Saver / InstaDapp (yet )

This proves our thesis that there is substantial demand for LP tokens and I suspect that LP tokens (not just from Uniswap, but SushiSwap, Curve, Balancer, 1inch, etc.) will contribute a significant amount of stability fees going forward. Now that we’ve tested the waters, I think we should try to scale this as quickly as possible.

That being said, there is still significant Oracle risk given the smart contracts have not been formally audited. So here’s what I suggest. We get an audit on the contract ASAP. It would actually be great if we could fund this through the Surplus Buffer as it would showcase an end to end community led effort of how to onboard something into the Maker Protocol in a responsible manner and be a blueprint for future initiatives to follow. Audits will take time though as auditors are frequently scheduled months in advance.

In the meantime, the Oracle Domain Teams confidence in the Oracle contracts has increased given we haven’t seen a hack over the holiday break (knock on wood). Rather than increase our exposure by increasing the debt ceiling for DAI-ETH, I think it would better to onboard USDC-ETH and WBTC-ETH with 3,000,000 DAI debt ceilings (though I will defer to the Risk Domain Team if they have strong opinions on the matter). This positions us much better for the future given that debt ceilings can be raised in an instant but collateral onboarding is a process with significantly more friction. I would strongly recommend we do it this way instead of raising the DAI-ETH debt ceiling, though ultimately it is up to Maker Governance on how to proceed.

EDIT:
DappHub audited the Uniswap V2 code and has significant expertise with respect to the Multi-Collateral Dai contracts. If they’re willing (and Governance agrees of course) I think they would be an ideal auditor.

[hongbiao_li]

The development of LP tokens has broad prospects. We should seize the opportunity to speed up the entry of various LP tokens and increase the priority of LP tokens.

[brianmcmichael]

Seconding Nik. I have some concerns about the current oracles that I want to test out when I’m back in the saddle. The Foundation’s given us the week off and frankly I need to step back a bit.

This Oracle is currently in test using other people’s money. It needs a little more R&D before we go all-in. The original proposal stated that this was a test and that the $3m debt ceiling would stand until audits are complete. I have been having some concerns that liquidations can be triggered inadvertently and we need to account for that. Please do not vote to increase this debt ceiling.

[ultraschuppi]

Hi @brianmcmichael

@brianmcmichael

I have been having some concerns that liquidations can be triggered inadvertently and we need to account for that. Please do not vote to increase this debt ceiling.

This poll was meant to get community sentiment - of course we shall not proceed here if there are concerns by the corresponding teams. I adjusted the Next Steps section accordingly, HTH.

[ultraschuppi]

@NikKunkel

Rather than increase our exposure by increasing the debt ceiling for DAI-ETH, I think it would better to onboard USDC-ETH and WBTC-ETH with 3,000,000 DAI debt ceilings (though I will defer to the Risk Domain Team if they have strong opinions on the matter).

happy with adding other LP-tokens into the mix, but: aren’t non DAI-LP-tokens significantly worse than DAI-based once w.r.t. to risk? also: USDC-ETH and WBTC-ETH do not incentivize on being a DAI-LP

[alexis]

I like the “incentive on being a DAI-LP”.

Interestingly, the surface attack is probably difference too. And it may be worse to put them up at a low level (like 500k) for the hackathon bounty. (100k to win for bug).

You should have a dai-usdc decoration attack possiblity.

But that is the expert decision. @brianmcmichael.

[iammeeoh]

@ultraschuppi

happy with adding other LP-tokens into the mix, but: aren’t non DAI-LP-tokens significantly worse than DAI-based once w.r.t. to risk?

Yes, WBTC-ETH will have different risks characteristics than DAI-ETH or DAI-WBTC, but these differences will be taken care (possibly) by different SF settings.

@ultraschuppi

also: USDC-ETH and WBTC-ETH do not incentivize on being a DAI-LP

I don’t think that these collateral types were added to stimulate DAI-LP, but to profit (read: SF income) from the existence of DAI-LP holders who are likely to benefit from this Vault type.

In any case we can (but this is a separate discussion?) promote DAI-LP by, e.g., reducing the SF of the pairs DAI-ETH and DAI-WBTC.

[ultraschuppi]

@iammeeoh

I don’t think that these collateral types were added to stimulate DAI-LP, but to profit (read: SF income) from the existence of DAI-LP holders who are likely to benefit from this Vault type.

Who am I to question to profit I just say that DAI-LP-tokens - additionally to our income - also provide (more) DAI-liquidity as the non-DAI-LP-tokens - which is a nice sideeffect

@iammeeoh

In any case we can (but this is a separate discussion?) promote DAI-LP by, e.g., reducing the SF of the pairs DAI-ETH and DAI-WBTC.

we are set on 1% SF right now and the (quite low) DC was hit really quick. I don`t think we further need to promote this right now. having at at (around) half of ETH-A-SF seems pretty reasonable to me?

[bit]

The stability fee should be set to zero. DAI is above the peg. The only thing that keeps the DAI price from taking off are the risky (because of the regulations and centralization) stablecoins which also have 0 SF.

[NikKunkel]

@ultraschuppi

happy with adding other LP-tokens into the mix, but: aren’t non DAI-LP-tokens significantly worse than DAI-based once w.r.t. to risk? also: USDC-ETH and WBTC-ETH do not incentivize on being a DAI-LP

So I’m basing the USDC-ETH and WBTC-ETH as next on the fact that the community proposed them, they are very high on the Collateral Onboarding Prioritization Document, and the Domain Teams have already submitted the Oracle, Smart-Contract, and Risk Collateral Onboarding Assessments for them. The community already voted in USDC-ETH so all thats left is to put it in an Executive Vote. WBTC-ETH would still require a final Polling Vote. Notably these 2 pairs are also some of the highest liquidity Uniswap pairs which means they have the most demand.

I think promoting DAI liquidity is an interesting secondary benefit of the DAI based pairs. However, I think we still have to prove whether we can create demand for non-liquid LP pairs just by virtue of us adding them as collateral or whether we’re subservient to which pairs there is currently LP demand for. Until we confirm the former and can leverage it to bootstrap new DAI LP pairs, our resources are probably better spent taking the conservative approach and adopting LP pairs with existing demand.

Looking at the list of MIP6s, I see we already have applications for DAI-WBTC and DAI-YFI. Looking at which pairs actually have liquidity on Uniswap instead, I see DAI-USDC at 20M and DAI-USDT at 5M. There’s also some BAC and BAS based pairs with significant liquidity but those are highly risky to have exposure to. I think if someone submitted MIP6s for DAI-USDC and DAI-USDT those might be good pairs to target after WBTC-ETH and USDC-ETH.

[cmooney]

First, it’s great to see strong product market fit. However, I wanted to briefly mention here what my concerns are:

• We need audits (possibly three) of the new oracle code.
• We have not seen how liquidations happen yet. Pricing LP tokens and recycling capital is more complicated for keepers. It’s likely that we need to ensure they are tooled up first. I recommend reaching out to keepers in rocket chat.

Whatever approach we take, it should likely be a slow DC increase and other LP types as we get those audits and see how performant auctions are. I would not personally feel comfortable with 12 million unless we could see a number of auction participants and have at least one audit.