[Signal Request] Adjusting Dust Parameter (2021-2)

Hi everyone!

The dust parameter is the minimum possible debt for a vault. You can learn more about the dust / “debt floor” in the documentation here. The dust value of all vaults is currently set to 2,000 DAI.

Governance most recently raised the dust parameter in January, when it was increased from 500 DAI to 2,000 DAI due to increasing gas costs and rising ETH price. Since then, the potential costs of participating in auctions has grown, as evidenced by the recent market drawdown where sustained gas prices reached well over 1200 gwei.

This table shows the gas costs in DAI for liquidation based on ETH and gas price, and also shows the minimum debt (dust) tolerated for auctions to clear successfully without a loss to Maker.

For instance, when liquidating a 150% collateralized position in a distressed scenario where gas price increases to 1,500 Gwei, auction costs (kick, tend, dent, deal, trade) for keepers amount to about $1,580 of gas fees. This means the debt size (dust) can not be below 3,160 DAI in case the keeper makes such a transaction with zero profit or loss and no loss is made for Maker. If you assume the market price of collateral continues to fall during the OSM delay and auction period, and the final auction settlement price is 15% below the vault’s liquidation price, the minimum safe debt size (dust) would be 5,750 DAI to ensure Maker does not make a loss.

Note that ETH-B vault is potentially a special case, as there is only 30% excess collateral available to protect Maker from losses. Using the same hypothetical inputs as above (1,500 gwei gas and 15% deficit of final auction settlement price vs the vault’s liquidation price), the minimum safe debt (dust) for ETH-B vaults would be 15,060 DAI to ensure Maker does not experience losses.

Also note that unwinding a vault costs approximately 400k gas, which makes it less economical for vault owners to deleverage themselves during a market decline where gas prices spike.

I believe governance should consider raising the dust parameter due to the risk considerations mentioned above. Given ETH-B vault type’s higher risk, I will include a separate poll for the ETH-B dust parameter versus other vault types.

Drawbacks of increasing the dust parameter include less access and worse user experience for small size vault users. While community members have raised the possibility of deploying vault infrastructure in another environment with lower fees (Ethereum L2, xDAI, etc), this would likely take substantial time and resources to set up and won’t provide a solution to small users in the short term.

Signal request poll:

Please select all options you would support in an on chain vote (multiple choice)

What should the dust value be set as (all vault types excluding ETH-B)?

  • Abstain
  • 2,000 DAI (current value)
  • 3,500 DAI
  • 5,000 DAI
  • 7,500 DAI
  • 10,000 DAI

0 voters

What should the dust value be set as (ETH-B)?

  • Abstain
  • 2,000 DAI (current value)
  • 3,500 DAI
  • 5,000 DAI
  • 7,500 DAI
  • 10,000 DAI
  • 15,000 DAI
  • 20,000 DAI

0 voters

Next steps:

This poll will close on Wednesday, March 10. I’ll submit this for an on chain poll if a majority of forum voters support changing the dust parameter.

6 Likes

Abstain option needed. :smile:

1 Like

Thank you! Gonna remake the polls real quick.

1 Like

Maybe EIP-1559 will help with this situation a bit.

1 Like

Here’s some gas history by hour. ‘Fast’ is the average value of the lowest gas price in a block.

Luckily median gas prices weren’t sustained above 1000 gwei for that long. Though who knows about the future. Well one benefit is that Maker Liquidations are a bit less time sensitive than AAVE or Compound and place less gas pressure.

3 Likes

I’m not so optimistic about this. EIP1559 should make gas prices more predictable and less easy to manipulate, but not necessarily lower than they are now. L2 and rollups will help with scaling, but even this is not likely to reduce transaction costs on L1. By pooling many users’ transactions together and spreading the cost of posting proofs among all of them, L2 adoption might actually make L1 transaction costs even more expensive because the L2 users will be less price sensitive.

Potentially liquidations 2.0 could help with the dust issue more generally by improving auction efficiency, and subsidizing gas for the “kick” stage of auctions.

To continue serving small scale users sustainably, I think Maker will eventually need to deploy vault infrastructure to alternative L2/L1 platforms.

This time I’m with you on this one brother. I have come around and realized that High Gas Tx Fees and High Dust Parameters protect the novice and Non-accredited investors.

Kudos to @Joshua_Pritikin who I disagree with in previous discussions, but looking back he was right :slight_smile:

6 Likes

I’m not all that happy about being right @ElProgreso. Just to be contrary, I’m in favor of rolling out Maker to L2 chains like xDAI, etc. :grinning:

5 Likes

Maybe not lower on an absolute basis, but we are in an unfortunate situation where we’re forced to process most of our liquidations at 1000gwei rather than 100gwei. If we can liquidate at 2x the mean gas price rather than 10x, I think that’s a big win for being able to keep dust down. A dust of 5000 Dai might be necessary, but it’s starting to become cumbersome for even sophisticated individual investors.

1 Like

DAI 5000 for dust breaks my heart.
We are becoming a whale-only tool if this continues, totally the opposite of what I want for the system. @Joshua_Pritikin you (and others) have mentioned L2 solutions, but can these really help? Would it be possible to work out a “L2 overview”? What are the statuses of the various L2 solutions and how much can each of them lower fees such as Dust?
I mean there is no point in doing an L2 effort unless the benefits are both real and permanent.
Maybe it is better to just wait for Vitalik and his crew to do their thing?

4 Likes

I hear you. 5k DAI minimum debt is a pretty large amount even for relatively wealthy users.

In the short term, a potential ETH-C vault could be modified to have a slightly lower dust value than the other vaults. Looking at the chart for an ETH vault at 175% collateral ratio, the same 1,500 gwei gas and 15% auction price drop vs vault liquidation price yield a minimum safe dust value of around 3,500 DAI.

ETH-C will charge lower borrowing rates than ETH-A, so in a way we would still be subsidizing risk from smaller users. But this could be a more feasible way to continue offering small dollar loans with less risk.

I think the best course of action would be to wait for ETH-C to be onboarded to the protocol, and then we can create a signal request to reduce the dust value of that vault type if the current gas and ETH price environments still support the reduction.

3 Likes

Today dYdX launched an alpha version of their L2 solution using StarkWare. Maybe this is something that Maker should be looking into more seriously.

4 Likes

I think that too much dust will only make the participation of small and medium investors lower and lower, which is not conducive to the development of Maker.

Hey everyone, reminder that there’s one more day to participate in this signal request poll :slight_smile:

1 Like

The on chain polls are active for voting through Thursday! Please vote: Maker Governance - Polling

2 Likes

Looks like one of these (ETH-B dust) was accidently omitted from the governance portal on Monday. We’ll get it put up next week instead as it’s a bit late to start it now with less than 24 hours to go. Apologies for the inconvenience all.

We’ll attempt to figure out what happened and what we need to do to avoid the issue in the future over the next few days.

5 Likes

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.