[Signal Request] Accelerate the PSM Launch

[hexonaut]

Background

On Monday I introduced MIP 29 - Peg Stability Module which is an implementation of the PSM defined here. While I originally wrote the PSM implementation and MIP with the intention of going through the regular monthly governance cycle, it has become apparent to me that accelerating this process may be in the best interest of the protocol.

Without mentioning anyone in particular, we are hearing again and again from credible sources in the rocket chat that we are losing market share to USDC with both integrators and users in Latin America. The reason being cited is almost always the issue of the peg not being at $1. I think with these past adjustments of USDC-A moving to a 101% CR has got us most of the way there, but we need to take that final step.

Currently we are a worst of both worlds situation where we have massive stablecoin exposure, but are still trading at $1.01 USD. From a risk perspective, there isn’t a big difference between having 500M of stablecoin exposure and 1000M. There has been hesitation in the past to go lower than 101% CR on stablecoins due to uncertainty about liquidations which has led us to have to consider some awkward options. At the time these options were the only thing available, but I now believe we have a better option with the PSM.

Proposal

I am proposing we deploy the PSM with a 0.1% fee from USDC → DAI and a 0.1% fee from DAI → USDC. We can initially do this for USDC, and if all goes well deploy to the other stablecoins in the following weeks.

It is important that we do this soon not only to preserve the peg, but because stablecoin vaults will soon be approaching a CR of 100% where we will lose an opportunity for a natural unwind by vault holders.

By taking this course of action we get the following benefits:

• Restore the peg. Dai will be forced to trade between 0.999-1.001.
• Collect almost exactly as much fees as we can get from the Stability Fees anyways.
• No longer have to worry about liquidations as the PSM will self-unwind when Dai drops below $0.999.
• Additional profit on the unwinding stage which would previously have gone to vault holders.
• We enable very large holders to go between Dai and Stablecoins without slippage.
• Most active vaults in all stablecoins should migrate over immediately if we don’t wait too long as there is a profit taking opportunity.
• Any leftover (abandoned/forgotten) vaults can be manually swept over to the PSM as long as they are above 100% CR.
• As we get more tools to maintain peg stability we can back off with the PSM fees to find a more profitable/sustainable balance. This is in contrast to the stablecoin vaults where once you lower the CR you can’t re-raise it.
• Switching to a fee model gives us all the revenue up front, and exactly.

To keep this objective, I’ll list the cons I can see:

• We will increase stablecoin exposure even more. Estimates are another ~400M Dai to push us to $1.
• Removes profitable opportunities for people speculating on the price of Dai.
• Price controls are not really a long term solution. We will need to eventually replace the stablecoin exposure with other assets.
• We are even more so at the mercy of USDC than we already are.

While this is definitely not a long term solution, I think it is good enough to both solve our current issues with the stablecoin vaults and fix the peg at the same time.

Technical Considerations

The code for PSM is actually very straight forward. It is mostly using existing systems, so the smart contract risk is not high. That being said, if we are to proceed on this accelerated route, we still want as many eyes as possible looking at it before hand. Some other members of the smart contracts team are already giving it a look over. Security always comes first.

Timeline

We are working against a ticking clock with the stability fees, and the sooner the peg is in-line the better. That being said, I still want to give ample time to discuss this, and if MKR holders don’t want to go this route so be it. I will run this poll until Friday, November 20th. If people want to proceed, I will craft a governance vote for Monday, November 23rd. Upon success of that we will turn off the stablecoin Stability Fees in the executive for Friday, November 27th and aim for PSM launch the following Friday, December 4th to give the smart contracts team time to set things up. A passing governance vote will override the Stablecoin Action Plan.

Note 1: I will continue with the formal process of MIP29 submission as well regardless of the outcome of this vote. If for some reason this passes and MIP29 gets rejected, we will work to close out the PSM.

Note 2: I will also continue to run the Stablecoin Action Plan governance polls in parallel with the assumption that is still the route MKR holders want to go.

Polls

There are a couple of options for the details of this proposal. I’ll split them out into separate polls:

Do you Support this Proposal?

• Yes, I want to proceed with the accelerated PSM
• No, I would like to stick with the normal MIPs process
• Abstain

What should the USDC → DAI fee be (tin)?

• 0%
• 0.1%
• 0.25%
• 0.5%
• 1%
• Greater than 1%
• Other
• Abstain

What should the DAI → USDC fee be (tout)?

• 0%
• 0.1%
• 0.25%
• 0.5%
• 1%
• Greater than 1%
• Other
• Abstain

[iammeeoh]

Just voted.

@hexonaut I find it confusing that you voted for several options in the tin/tout values.

Currently tin=0.25% seems to be winning vs the 0.1% value proposed in your post (unless I misunderstood something).

Perhaps better to vote for a single value?

I voted for tin=tout=0.1%.

[hexonaut]

With multiple choice votes you are supposed to vote for everything you would be okay with. So while I think 0.1% is the best option, I’d be okay with compromising on 0.25%.

[iammeeoh]

@hexonaut

So while I think 0.1% is the best option, I’d be okay with compromising on 0.25%.

Ok thanks. Now I think it’s super clear to everybody.

[LongForWisdom]

@hexonaut

With multiple choice votes you are supposed to vote for everything you would be okay with.

Yep, I know I’ve expressed this a number of times. But just to confirm, the signal process will work more effectively if people vote for everything they would be willing to support in an on-chain poll.

The theory behind this is that we will end up with a greater level of consensus for the winning change which both sends a strong signal to MKR Holders when the on-chain vote takes place, and helps prevent the community from fracturing when making decisions. I like to think most of us are here for the long-haul. The fewer people that disagree strongly with the winning outcome, the better.

[ElProgreso]

Thank you for posting this Sam and the great work. Perhaps the irony-- we aim to turn off the Stability Fees with an executive on what Americans like to call “Black Friday” (the day after Turkey Day) Nov. 27th. There’s no way we can run the poll until Sunday 15:00 UTC? And get a governance vote going on Monday the 16th?

Also, Thanksgiving Week is a very slow week IMO–must folks take the entire week, or half-the-week off. Somehow I feel people work more the week prior to Thanksgiving week, and then go on vacation mode starting 11/23… but perhaps I am wrong…

[alexis]

Great work Sam.

[hongbiao_li]

PSM, please hurry up, we really need you.

[Jiecut]

±0.1% for the PSM seems much to tight. Stablecoin exposure could easily balloon. Would also work much better if you had a bonding curve.

Are we really setting up DAI sell walls at 1.001?

Switching to a fee model gives us all the revenue up front, and exactly.

0.1% up front instead of requiring 1% extra collateral.

If you really want peg stability, it is best to set a wider range. The goal of the PSM should be as a backstop.

Curve already provides decent liquidity in the ±1% range.

[alexis]

Thinking, Don’t we want to do it progressively?

Let is say decreasing every week. There is a lot of fees to miss by doing it violently.

[hexonaut]

This is a great point. Yes we will want to do it bit by bit to maximize fees.

[Doo]

But this also creates an issue of some gaming the system by waiting till later to help fix the peg. The value of fixing the peg and avoiding stablecoin collateral vaults getting undercollaterized should outweigh potential small increase in fee gathering

[LongForWisdom]

That’s a coordination problem. Everyone in the market would need to agree to wait til later to make the trade rather than taking a small immediate profit at each price-step. It’s very unlikely to happen.

Decreasing the tin progressively would result in more total fees. The major downside is that it increases the time that DAI is further from the peg.

[alexis]

Small increase in fee … assuming 400M more dai:
400M *0.5% --> 2M

that is half year of income for Maker.

Also let’s say
first week -> 0.75%
second week ->0.5%
3 week -> 0.25%
4week -> 0.1%

that won’t take too long neither. You can cut the first week if you want.

[LongForWisdom]

Without stablecoins the stability fee revenue is ~12M a year. So it’s not quite half a year of income. I agree it’s a non-trivial amount though.

[iammeeoh]

@alexis

Small increase in fee … assuming 400M more dai:
400M *0.5% --> 2M

Regarding this, should we perhaps bundle a significant increase of the Surplus, together with this accellerated PSM launch?

In my opinion, since this operation is new and might reserve some surprises, it is better to keep the fees in the surplus and not burn any MKR for the time being. We can always burn MKR later, when the peg is safe and the operation was successful.

[alexis]

Seems a very good idea.

[OneBiteLawnOrder]

Is there any reason why we need to have the Tin and the Tout set at the same value?

My feeling is that because we are currently about 1% above the peg we start at 0.75% or 0.5% with the Tin and ratchet it down over a few weekly governance cycles until we get to 0.1%. This would allow us to see how the peg responds and what the debt ceiling looks like as we get closer to the peg. As @alexis points out this also gets us more fees.

I would set the Tout at 0.1% or less right away and have the PSM clear as soon as possible as the peg is broken to the downside.

[hexonaut]

@OneBiteLawnOrder

My feeling is that because we are currently about 1% above the peg we start at 0.75% or 0.5% with the Tin and ratchet it down over a few weekly governance cycles until we get to 0.1%.

I’m planning to add a contract which will enable dropping the fee down continuously over a specified time frame (such as a week). This will give us the maximum possible fees.

[alexis]

That would be amazing.