[Urgent Executive] ETH-A, USDC-A and PSM-USDC-A DC increase

ETH-B is at cap too, when can we raise the debt ceiling there?

3 Likes

Good job handling this @SebVentures and @hexonaut. I agree with proposed changes but I do want to note that we need to increase PSM-USDC-A DC as soon as possible to a higher level since increasing USDC-A DC is not ideal. The issue is that USDC-A SF is set to 0% and speculators have zero cost of capital to perform 100x leverage trade on the DAI price even before the $1.01 price is reached. This literally happened in the past few days. And this means USDC-A DC can get utilized very fast and can not be used right when needed mostly.

As soon as we feel confident about PSM we should increase PSM DC more drastically and set DC for old stablecoin vaults to 0. This also prevents increased exposure to TUSD.

As for ETH-A exposure I think nobody wants to miss on new DAI minting demand, but we should be aware that ETH exposure is increasing heavily and current rates are probably not applicable anymore at these risk levels. We have been undercompensating risk for ETH-A already in the past but now it is becoming more obvious. Good news is that competitive rates in DeFi increased heavily.

If we are going to be aggressive with growing debt exposure increasing SFs won’t be enough to protect though. Community also needs to think about 1) limiting further DC growth of certain collateral types such as ETH-B 2) Increasing box parameter 3) Increasing Surplus Buffer.

Imagine we have a big crash soon after this bull run and we get ETH-B liquidations queued due to the 15m box parameter and the price keeps tanking. Not sure though how we can argue higher box value apart from hoping the keeper ecosystem improved in past months. There was also discussion about increasing Surplus Buffer and it seems the community is divided between increase to 10m and leaving it at 4m. I think a solution with a Surplus Buffer increase and the ‘lerp’ module might be a good compromise.

12 Likes

There was already an onchain poll about setting the PSM DC to 500 MM and the SC of the legacy stable coin vaults to 0. Let’s do it!

Yes, the current box size is quite the limitation for the debt size and ETH-B. When the box parameter was first set we had $0 in the ETH-B vault and much lower exposure for other vaults.

I do agree that it’s a difficult discussion. Some arguments FOR can be necessity and the PSM.

Anyways, I’ll be excited when we’ll be able to transition to the new liquidation system.

Would like to see an increase in SF for ETH and USDC vaults and a higher DC for PSM.

I think if anything it’s become worse due to the lack of substantial liquidations (people do not run infra that remains idle). Is there a way that the DC IAM can help limit the risk of this happening?

Completely agree with this. Is the PSM with USDC going to do PEG management or a USDC-A vault with LR at 101?

Can’t agree with this enough. FYI: Limiting DC growth in say ETH-A pretty much means a locked LR. So again having multiple vaults potentially is a way to control LR exposure but not a KISS model at all.

I have been harping on raising surplus for what seems like forever, as well as creating a secondary surplus/treasury out of other assets (mostly stablecoins) but hey we have almost 500M on the books in stablecoins that are probably under their 101LR and can be liquidated at any time.

The real concern here is allowing everyone to rachet up debt exposure on ETH predominantly and then a ETH price crash on part with Black Thursday back to back 25% drops in 24hrs. Couple this with mempool manipulation to block out bidders and a new liquidation system not up yet and we have another recipe for people not just losing millions but a good fraction of a billion.

Fortunately we do have the IAM shutoff - but I am not up to the details of whether those are exempt from the GSM delay and/or how liquidations get shut off again should we end up in the same boat as Black Thursday - gas at 1000gwei and someone controlling mempools to limit smart contract access?

I honestly have no clue where Maker sits here with respect to another significant downturn event.

4 Likes

“The real concern here is allowing everyone to rachet up debt exposure on ETH predominantly and then a ETH price crash on part with Black Thursday back to back 25% drops in 24hrs. Couple this with mempool manipulation to block out bidders and a new liquidation system not up yet and we have another recipe for people not just losing millions but a good fraction of a billion.”

Is this really a concern? Rapid price drops were never a problem before for Black Thursday for Maker. And yeah there were keeper problems but havent there been some improvements since then? And bc it happened once it probably is less likely to happen again as market forces come into play.

My preference would be to see Maker get a lot more aggressive here. Raise the limit for Ether an additional 500 million, ratchet the PSM up to full throttle. Pulling in assets means pulling in more eyeballs as well. Lets grow the network while the times are good.

I think there is prob more danger in getting too conservative here as it could throw things out of whack.

3 Likes

I’d also add that the current “box” parameter is 15M Dai, so that is the maximum that can be auctioned off in any given OSM update. This naturally has trade-offs in both directions.

1 Like

Yeah thanks for reminding me/us on that. Frankly what I saw on Black Thursday there were a good number of vaults that would have completely avoided liquidation if there was an auction pause. Mostly because within I think 2hrs of the price break below 100 the price was above 110 and heading to 120.

So is the rate 1M/minute or 15M/15minutes. I think the biggest issue here would be something I have been harping on since I first came to Maker, DAI liquidity. USDC-B only has 30M available and that would be exhausted quickly. PSM is pretty much limited even if it is increased to 30M.

We are back to the same situation where if there is a market crash DAI liquidity is going to run out almost instantly. There still has been no movement on creating a liquidity backstop for emergency events that tracks with DAI outstanding. So IF/when something like Black Thursday happens again Maker will be in a similar boat. With Defi space having grown probably 5-10x since then the cascading liquidations across the space the liquidation demand will outstrip network capacity no matter what the price.

I don’t think people have considered what happens to on-chain liquidation systems with gas at 2,000 to 10,000 gwei.

1 Like

There’s some more on-chain liquidity out there and during Black Friday, Curve was just in its infancy.

I think the bigger constraint right now seems to be quantifying Keeper capital.

My opinion is that you want to be able to have the system open enough that anyone can be a Keeper during a period of mass liquidations and be able to place a bid. Just needs some publicity. That’s more scalable than requiring keepers to have enough capital, and limiting the box parameter based on that. Of course Liquidations 2.0 will be even better.

Yeah after a 2-3x run lets just rachet all of this up.

What could go wrong?

I thinking raising DC’s and not touching SF just allows people to put the system more at risk. Here is what I would be for.

For each DC increase of 10% up the SF on all assets by .5% and increase the surplus .5% of outstanding to stop MKR flapping for the moment. Later if things settle down and the protocol has cash in the bank start flapping MKR again.

This would be for every 100M of total DC, .5% increase in SF (about 10% increase in average Maker earnings) and bumping the surplus before flap by now 6M a bump.

Good point on the SF. When is the Rates group planning on submitting a proposal to adjust rates?

That is a design flaw. We should consider keeping ETH as a reserve currency to mitigate that problem. ETH is still below ATH.

You probably mean LR. The LR deals with the risk and the SF with the peg.

We should increase the LR of the stablecoin library and accelerate the expansion of the use of PSM. In addition, close USDC-A, create USDC-C and set SF = 4%, which will benefit our adoption of PSM.

No I meant what I wrote.

As long as the PSM has DC the SF is rendered irrelevant for PEG management and instead should be used to manage risk in terms of a system fee to control borrowing. I believe risk to the system here has increased quite a bit. If we are going to add 100M more DAI of DC and open up the PSM to 500M+ our borrowers can certainly afford another .5% SF. At some point when the SFs are high enough they will reduce their borrowing. Until that point they will MAX out the DCs over and over and over until we DO raise the SF’s or we finally hit their LR borrowing limits (which we’d like to avoid). Given the 200-300% ETH/BTC price rise I would advocate a pre-emptive rise of the SF across the board by .5-1% with the increased DC. In fact were I in charge of this I would study the markets generally and try to normalize borrowing rates across the board on all assets in the MCD. This is particularly true of ETH borrowing at 2.5% btw. (this could be 3-4% easily).

Regarding all stablecoin vaults with LR 101. I would also be a fan of zeroing the DC’s on all of them. Now. In time increasing the SF just enough to put the stablecoin LR 101 vaults in liquidation territory and then piecemeal liquidating them into the new PSMs. This is particularly true if Maker is going to continue down this route of using stablecoins as the DAI PEG management mechanism into the foreseeable future.

The key point to pay attention to here is that using the PSM to manage the PEG (with effectively unlimited DC) frees up the SF for risk management and overall rate normalization across the DeFi space. I am surprised no-one has talked about this as a sea shift in the way Maker incentives work relative to DAI and as @mrabino1 would suggest will probably have unintended consequences.

Uncomfortable with this new concept that we are absorbing stablecoins as anything other than a last resort for peg management.

1 Like

OSM updates once per hour, so it’s 15M maximum every hour (on a 7 hour delay given the OSM itself is delayed an hour an auctions are 6 hours)