Black Thursday Response Thread

Right guys, here’s the thread where we discuss and figure out what we are doing, and when we are doing it.

We need to decide the following:

  • What are we doing?
  • When are we doing it?
  • Are we happy to ignore the regularly scheduled polls this week?
  • Do we want to consider triggering Emergency Shutdown in the short term?

I would like to see as many short-term options suggested and discussed as possible, I will collate these, and will be responding on this thread until we actually decide on something (via forum poll). Once we have the options together, we will start forum polling on what we want to do and what we don’t.

It would also be fantastic if someone can do a writeup of the situation that I can include in this initial post so that everyone knows what is going on.

At that point, we can take the consensus options and present them to the Foundation for inclusion in an executive.

Link to ongoing call - Note that there is a 100 participant limit and we are running into it every now and again.


Summary of the Situation

Courtesy of @silto - Thank you!

During the ETH price crash, some Vaults (containing ETH as collateral to mint DAI) became undercollateralized (under 150% collateral). Since the oracle ETH price fed to MakerDAO came under those vault’s collateralization limit, they became available for liquidation. Liquidations are triggered by users calling the contract, which starts an auction for the ETH from the vault. The auctions is used to raise DAI to reimburse the liquidated vault’s debt. In normal operation, user (named “keepers”) compete for this auction, and the winning bid is paid (in DAI) in exchange for the ETH in the vault. The DAI is then burned, the debt erased, and if some ETH is left, it is returned to the vault’s owner.

What happened today is that during the big drop in ETH price, a lot of vaults became available for liquidation, at the same time as the Ethereum network became clogged with lots of transactions, leading to a sharp rise in gas price. This confluence of circumstances led to keepers either stoping their activity for fear of slippage (price dropping fast + transactions taking a long time to validate) and for lack of liquidity to absorb all those liquidations. This lead to a situation where only one keeper continued triggering liquidations, and bidding 0$ for the ETH, with no competition. This situation lasted around 2/3hours, before other keepers came in and resumed normal operation.

As a consequence, some vaults where liquidated with 0 DAI coming back in the system, resulting in a net loss for the system. The MakerDAO had a +500k$ surplus before the price drop, and now has a -4M$ surplus that needs to be filled. The protocol covers this issue, the solution being to trigger an MKR mint and auction, the DAI raised being used to fill the surplus debt. During normal operation, MKR is burned when debt from vaults is reimbursed, this would be the opposite mechanism.

In order to avoid a 0$ bid from winning a liquidation auction (which was possible because of a short auction timeout - 10mins -, the biggest bid wins the auction if no other bid covers it) the Maker team is currently discussing the possible solutions.


Summary of the Call

Helpfully provided by @ryanberckmans
  1. Media should email [email protected]
  2. An emergency shutdown (not happening now) would cause DAI holders to take a haircut, whereas the social contract of MakerDAO is that MKR tokens take a haircut in the event of system failure. Therefore we should try and ensure that MKR holders take a hair cut by avoiding emergency shutdown if possible. I heard that emergency shutdown is not being considered as an immediate option.
  3. MKR buyers should prepare for a large MKR system auction in < 48 hours. If this is successful it will pay back the current $-4M surplus.
  4. For altruistic whales who want to help, the current best way to “donate” to this issue is to bid on the upcoming MKR system auctions.
  5. MKR buyers should consider buying MKR on open market to support the underlying MKR price. (down 44% in last 24h)
  6. ETH buyers should consider running a keeper bot to buy liquidated collateral and help prevent this issue from happening again. https://github.com/makerdao/auction-keeper
  7. The simplest root cause here (unconfirmed) is that a lack of keeper bot competition contributed to ETH collateral being sold for $0, possibly because of high gas prices and eth node sync issues, and many keepers run their own eth nodes.
  8. In the next 24 hours, we must work on ways to make it easy for anyone who can afford it to participate in the MKR system auctions. We won’t want to constrain the set of auction buyers to those who can afford it and have the technical ability.
  9. In @ryanberckmans opinion, keepers and MKR buyers should prepare for sustained high gas prices, and downward pressure on ETH and MKR. The Dow Jones hit sell-off circuit breakers three times in the past week and a half. It’s a historic week. We should plan for global markets to potentially crash further which may correlate with further crypto drops.

Current Issues

1. Dai is off-peg.
2. FLOP Auctions will occur concurrently, we may want to attempt to mitigate this.
3. There is ~4 million worth of bad debt in the system currently.


Suggested Options

1. Increase the FLIP auction lot size

Pro: Decreases the load on the ethereum network.
Pro: Easier to track instead of 1000 liquidations at once.
Pro: Lower transaction fees per unit collateral makes it more worthwhile to take part.
Con: Prevents keepers with lower levels of capital from participating in auctions.

  • Yes
  • No

0 voters


2. Reduce the FLIP auction lot size

Pro: Allows keepers with lower levels of capital to participate in auctions.
Con: Increases the load on the ethereum network.

  • Yes
  • No

0 voters


3. Reduce the DSR

Pro: Brings more Dai in circulation, hopefully moving the Dai price closer to the peg.

  • Yes
  • No

0 voters


4. Reduce the Global Stability Fee

Pro: Encourages more CDPs to be opened for arbitrage on the price of Dai, hopefully moving Dai closer to the peg.

  • Yes
  • No

0 voters


5. Create an out-of-order executive to pass the chosen changes as soon as possible

Pro: Faster response.
Con: Breaks the known schedule, possibly confusing MKR Holders.

  • Yes
  • No

0 voters


6. Ignore the current set of polls in order to do emergency response procedures.

Pro: Faster response.
Con: Breaks previous precedents around process.

  • Yes
  • No

0 voters


7. Increase the initial bid price for the collateral auctions.

Pro: Resolves zero bid-size issue.
Con: Requires a technical change to the smart contracts, might not be immediately available.

  • Yes
  • No

0 voters


8. Increase the minimum bid time.

Pro: Gives more time for keepers to bid on each auction.
Con: Increases market risk for keepers.

  • Yes
  • No

0 voters


8. Make the minimum bid time dependent on the spread between the bid and the oracle price.

Pro: Gives more time for keepers to bid on each auction if the bid is currently low.
Con: Increases the market risk for keepers as they no longer know for sure how long the bid will last.

  • Yes
  • No

0 voters


9. Increase the debt buffer before the MKR mint auctions happen.

Pro: Prevents MKR from being auctioned to recover from the bad debt.
Con: Prevents MKR from being auctioned to recover from the bad debt, this breaks the implicit promise that MKR will be used to recapitalize the system.

Edit: Turns out this isn’t as simple as changing a parameter, probably not feasible in the short-term but I will leave the poll up.

Second Edit: Turns out this is possible by increasing the debt buffer to greater than the current level of debt.

  • Yes
  • No

0 voters


10. Decrease the MCD ETH-A debt ceiling to 110million and the MCD global debt ceiling proportionally. (From Cyrus)

Pro: Reduces the severity of the OSM attack.

  • Yes
  • No

0 voters


15 Likes

Mentioned it here but it seems to me that the first move would be to drop the rates.

https://forum.makerdao.com/t/the-recent-jump-in-flip-auctions/

Would even consider putting out a new spread poll. I know that we recently voted to change the cadence but recent events suggest that we might want stop minting unbacked dai as much as we can until things stabilize. 5 year us t-notes are currently yielding around 0.267 % so i could easily seeing spreads as high as 3~4% being justified at this point.

1 Like

QE efforts may too be justified. IE minting MKR, but not sure we are quite there yet. Also not sure what the best way to spend the minted mkr would be at this point. Would be interested in what people ITT have to say.

Minting MKR to buy DAI doesn’t really help right now. That’s like QT.

Who says we would want to buy DAI? I just know that minting MKR is a lever we can pull. Unsure of the best way to spend the funds that come from it.

1 Like

Increasing lot size could be discussed. 50 eth might be too small. Not scalable enough.

Good point raised in zoom. Could have some keeper delay bots. Can delay the $0 bids by slightly outbidding. So that the $0 bids don’t execute.

I’ve seen arguments to the contrary, that 50 eth is too large, and bringing the lot size down allows more keepers to take part.

Not sure I quite understand this, can you elaborate on the mechanisms there?

Well currently, people are bidding $0 which doesn’t require capital. Bigger lots means less transactions being spammed. Less transaction fees to make a bid. Worth people’s time more. Should have less chance of a $0 bid. Easier to track instead of 1000 liquidations at once.

I think cyrus brought it up. If someone makes a bid it delays the auction. So you bid for $1, pay your $11 tx fee, delay the auction. Wait for keepers and liquidity to catch up.

This is kind of what i would be thinking when i say QE. Like could we mint mkr with the intent of using it to participate in these auctions.

Understood. Sort of like a capital raise of a corporation for emergency operations. Makes more sense than for using it towards the system debt currently.

Maybe it’s worth considering alternative auction systems. I can totally see that under “normal circumstances” bidding zero for something worth more would quickly get out bid.

But the problem we’re seeing it that there’s insufficient keepers, and those trying to participate likely fail due to the network congestion and gas price.

We should look into finding an auction system which is less exposed to these types of network conditions. Having more keepers might help, but if the network is like now, that might not be enough.

2 Likes

Probably not a short-term solution, unfortunately.

Not a short-term fix, let’s try and stay focused please.

2 Likes

Are we just worried that it would take too much work on the technical side of things? Is there an option where minting mkr for auction capital is performed manually?

Sure, maybe, but we find ourselves in this situation because of lack of long term thinking? We can keep on throwing short term fixes at this… but what’s the short term fix for ensuring longer term sustainability?

The executive goes live tomorrow (or earlier) this is our chance to make positive changes immediately, any other changes can happen once we’ve figured out what we want to put in that executive.

1 Like

Short term goals:

Bring down DAI price, so flush market with DAI, reduce DSR, reduce SF… anything technical?