USDC: Peg Arbitrage vs. Auction Liquidity

Hello again everyone. So it looks like the uptake on USDC as collateral has been very strong, so strong in fact that @cyrus is worried that we will not have enough liquidity available for auctions.

This issue was touched upon in today’s governance and risk meeting, but to lay it out simply:

The main goal of onboarding USDC as collateral was to provide keepers access to liquidity for the coming FLOP auctions (as well as any unexpected FLIP auctions.)

Apparently (and unfortunately, sort of) much of the debt ceiling is being used by people arbitraging the Dai peg rather than keepers stocking up for auctions.


Given this is the case, we have a few options that I can see:

1. Increase the stability fee on the USDC-A collateral package to discourage peg arbitrage.

Pro: Allow the remainder of the debt ceiling space to be used by keepers for auctions, and may cause peg arbitrageurs to close their positions.
Con: Will probably annoy those users that have been using USDC-A to arbitrage the peg.

2. Increase the debt ceiling on the USDC-A collateral package to allow more room for keepers to create Dai liquidity during the auctions.

Pro: Provides more liquidity for auctions.
Con: May still get eaten up by peg arbitrageurs.
Con: We add more blacklist and regulatory risk to the system.

3. Add an additional USDC collateral package (USDC-B) with a higher Stability fee that should only be used by keepers due to the prohibitive cost over periods longer than six hours. Optionally, also reduce the debt ceiling on USDC-A.

Pro: Provides more liquidity for auctions which cannot be used by peg arbitrageurs.
Con: We add more blacklist and regulatory risk to the system. Unless we also lower the ceiling on USDC-A by an amount equal to the new debt ceiling on USDC-B.


Personally I like option 3. I think so long as we keep the total debt ceiling between these packages equal to $20M Dai we are in a good position. I would rather not raise it beyond that unless we have to.

If we set USDC-A to zero debt ceiling but allow those with positions to maintain them I think we are also fine to move the entire $20mil ceiling to USDC-B. That said, I would rather see a $10M-$10M split, or at the very least a $5M-$15M split.

If I’ve missed any pros, cons or options @ me in a reply with the change and I’ll make it when I see the reply.

Isn’t the peg arbitrage crowd adding DAI liquidity to the market overall? What’s the difference between keepers opening vaults vs buying DAI on the open market?

My first inclination is #3 sounds good, as this would reserve some of the usdc DC for ultra short term usage during liquidity crises

2 Likes

Yes, in my view it is, and I’m not 100% sure why the peg being at $1 and having USDC Vault liquidity available aren’t equally useful.

That said, we have an Interim Risk Team who know way more about this than I do and I trust them to know what they are talking about. So, if Cyrus believes we need the liquidity available no-questions-asked in the form of a USDC vault, then I’m (personally) happy to back him up.

1 Like

So I think the goal of the vaults shouldn’t just be to provide keepers with liquidity during auctions. Keepers can also source liquidity from markets. The whole point of the USDC vaults should be peg stability. A shock absorber that can be tapped when there’s a big need for DAI lqiudiity.

Though, even with the goal of peg stability, you want to have room in the debt ceiling when there’s an actual shock in demand, sort of like the use case cyrus mentioned. So that means that the SF needs to be priced at a point where it’s mainly used for shocks in the system.

I support option 3, adding another collateral package USDC-B with a very high SF so that it’ll only be used for peg stability in extreme situations.

Also, it might still be a good idea to also slightly adjust USDC-A parameters. A 130% collateralization rate or 25% sf could be possible.

2 Likes

Very good point, there are definitely other more nuanced alternatives than the three I initially presented.

1 Like

Chart comparing the Implied leverage ratio with different collateralization ratios. You have to take the peg into account, as you can do more leverage if the peg deviates more. It might be make sense to use 130% for USDC-B also.

Well I cant see arbing the peg as bad necessarily. The peg at 1.02 is better than 1.08 or 1.1. I am seeing about 10 mil USDC locked up. I think your 10 10 split makes sense in the current state @LongForWisdom. We got a benefit from USDC-A, now if we can close the space in USDC-A while opening USDC-B with the high stability fee maybe keepers will have to use USDC-B, if there is any debt ceiling space they would use the option with the lower SF right? unless maybe if we have a lower CR on USDC-B, letting keepers take more leverage.

This is getting really complicated though. I am trying to contribute, but I just don’t understand this stuff well enough. I am concerned about how few people have a really good grasp on all this. Cyrus and co are amazing, I really hope they can get rest, burnout of the risk team is a serious risk to the protocol. Repeated intense stress/little sleep will lead to worse cognitive function no matter how smart they are.

1 Like

I think pool here is a good Idea,

my personal preference is C. Idea to have two packages of same collateral with different parameters is highly underutilized

one remark, It is obvious to me, that introducing new package of some type of collateral should decrease debt ceiling of other packages off this collateral. Not necessarily by same amount, but two should be correlated negatively.

My preference is also option 3, but with cumulative (USDC-A, USDC-B) debt ceiling of 20m.
We would need to determine ratio though.

1 Like

Are these mutually exclusive options?

The combination of Option 1 & 3 seem attractive in that:

  1. you can increase the stability fee on USDC-A just enough to throw the arbitrage off, have the arbitrageurs cease to recalibrate their models, and temporarily save the remainder of that liquidity.
  2. create the USDC-B to create new liquidity (with stability fees more than USDA-A)
  3. leave debt ceiling headroom untouched and thus avoiding any more complexity

The issue I see with that is why would a liquidity provider use USDC-B if USDC-A has a lower SF and debt ceiling space?

They wouldn’t, but that’s okay. We just want to ensure there is some space reserved for auctions, rather than ensure there is space for peg management.

So the current parameters are okay then? plenty of space in USDC-A unless risk thinks 12 mil isnt enough

Why is that a bad thing? After all this arbitrage will bring the peg down to $1. An arbitrager will mint DAI using USDC and sell it on the market, which will source liquidity. Given that the current peg seems to be stuck at $1.02, why not decrease the SF for USDC and incentivize even more down pressure on the peg price?

1 Like

That’s what I was thinking. We could add a USDC-B with a higher SF and lower the SF for USDC-A.

I agree, it does do that. But auction liquidity can also be created by having a higher SF effectively allowing for use by auctioneers only.

Decreasing the SF for USDC does make that arb more profitable, I agree.

Adding USDC-B looks like quite weird to me. If I want to open a CDP using USDC, I’ll always use the option with the lowest SF. In that sense, I don’t see value in adding USDC-B when there is still plenty of space for USDC-A to grow. Using the 1inch routing algorithm it seems that we need around 4.8DAI to be sold on the market to push the price to $1. That’s well with-in the current USDC-A debt ceiling.

Yeah, so encourage USDC-A by lowering the SF and add USDC-B for emergency liquidity.

2 Likes

The problem is your arb example only goes halfway. in your example the arb people have now made some extra USDC but they have to pay off the debt in DAI. Now they have to go buy DAI and pay of the debt, removing liquidity, and if they don’t they are losing 20% in interest. Unless I’m missing something I don’t see how you can Arb DAI right now.

Side note the peg has gone down MAYBE .5% over the 2 days USDC vaults have been available. the Peg was already lowered prior to the USDC vaults going live

1 Like