[Signal Request] Stablecoin Action Plan

The Rates Working Group is looking for feedback from governance on how to proceed with the stablecoin vaults. In particular, the lowest collateralization stablecoin vaults will start dropping below 100% collateralization ratio (CR) on December 19th, 2020 at the current rate of 4%. To see why this is an issue please review the most recent G&R call (~57 mins in). I will attempt to summarize the pros and cons of each option below.

NOTE This post has been edited after the failure of Option 3 to secure enough MKR in the governance poll. I’ve added 2 options based on feedback - Do Nothing and Option 1 variants which are reflected in the new governance poll.

Do Nothing

In the do nothing option, MKR would be burned and stablecoin vaults will become undercollateralized.

Option 1 - Keep Raising the Surplus Buffer

In this option we let Stability Fees continue to accrue and increase the surplus buffer so we don’t end up burning MKR only to re-mint it later. The reason is that fees collected after 100% CR are virtual in the sense that, when the vault goes to auction, the revenue collected will be deducted to cover the bad debt. If we had previously burned this revenue it may result in a debt auction.

Variant A) Keep Stability Fees at 4%
Variant B) Lower Stability Fees to something still above 0%

Pros

  • We continue to collect fees from newer vaults.

Cons

  • The lower the CR, the less likely vault owners will close their own positions. Auction inefficiencies may cost us some money - even with auctions V2.
  • We aren’t burning MKR.

Option 2 - Lower the Stablecoin Stability Fees towards 0%

In this option we will lower the fees towards 0%. Initially we would move the fees down to 2% with the expectation of lowering fees further as we approach 100% CR.

Pros

  • Higher chance of vaults closing themselves out.

Cons

  • We miss out on fees from newer vaults.

Option 3 - Migrate Users to a new Vault Type

This is the option discussed in the call where at the threshold of 100% CR for leading vaults, we set the debt ceiling and stability fees on USDC-A/TUSD-A/etc to 0, and open a new vault type USDC-C/TUSD-B/etc with the same parameters and some positive debt ceiling. To be clear this will not result in any immediate action. We will schedule the new vaults and DC/SF to 0 whenever the leading vaults reach 100% CR which will occur approximately on Dec 19th.

Pros

  • Maximizes fee collection while minimizing the amount of vaults that need to go to auction.

Cons

  • Technical complexity - there is a fair bit of work involved with rolling out new vaults.
  • We are bloating the vault types for a possibly one-time event.
  • We may have to keep doing this if Dai doesn’t return to $1.

Poll

  • Option 1 - Keep Raising the Surplus Buffer
  • Option 2 - Lower the Stablecoin Stability Fees towards 0%
  • Option 3 - Create a new Vault Type
  • Abstain

0 voters

This signal request will run until Thursday, October 29th, and can be extended if there is not enough activity or consensus. Next step is an on-chain governance poll followed by the Rates Working Group implementing whatever course of action wins.

9 Likes

Thanks for bringing the options to the forum!

Could you elaborate a bit on the technical complexity of option 3?

Right now I see:

  • governance effort to bring the new vault types + SF/DC modification of the old ones into the system
  • keepers: no effort (since liquidation are off anyway)
  • … That’s it? No code beside the spell(s) right? Am I missing something?
1 Like

Another big downside of option 1: every DAI in the buffer is DAI taken out of circulation. Probably not the biggest effect on the peg, but not insignificant for sure.

1 Like

Yes you’ve got it. Not a huge amount of work, but it does require some effort from the smart contracts team.

1 Like

Thanks, confirmed my pick on option 3

New Vault Type :loudspeaker:

One main advantage of raising the surplus buffer: People seem to forget that its entirely possible that Dai could one day (?soon) fall below its peg again. If this happened, people with USDC vaults that were below 100% LR could be incentivized to pay them off including the accrued stability fees. Therefore, its much better to keep the SF accruing as it could lead to profit for the protocol if Dai falls below the peg.

3 Likes

In what way do you view the system as being broken, and what solution are you seeing that would fix it?

1 Like

Just ignore him. He literally has never been for anything that has ever been proposed. Not sure why he even still shows up besides enjoying being a constant contrarian.

2 Likes

I) by the time we are below peg, the stablecoin-vaults get paid back regardless of their LR or the amount we have in the SB - just because the vault-owners do make a good profit out of that

II) addressing the increased risk by having the SB raised is already done in this weeks poll

III) this signal request is about, how do we want to deal with the issue of the current stablecoin-vaults that (soon) turn into undercollateralized vaults. option 1 is just a way of making this our future problem while option 3 is fixing the problem right now.

2 Likes

In option 3, the stability fee gets set to zero once USDC vaults reach 100%CR, so no new fees get earned on them anymore. The idea is that the protocol will never collect on them anyway. However, if Dai goes below the peg, we may in fact be able to collect those fees. I really don’t see any advantage over option 3 compared to raising the surplus buffer. Its just more complicated.

1 Like

I feel option 1 and 3 are postponement of a potential issue…

In a optimistic scenario regulators will approve of centralized stablecoins and so Dai will simply be wrapped USD. In that case we could have 0% stability fees on stablecoins simply to help build market share.

So option 2 it is.

4 Likes

We could always do a mix of strategies. There’s some good reasons to lower stability fees by a bit but still combine it with another strategy. You still get the benefit of a higher chance for vaults to close themselves out.

We can be less aggressive than 33 bps per month.

1 Like

@bit I’m going to start moderating you more aggressively when you make low-effort posts. I know you have a difference in opinion over the current direction, but posting stuff like this isn’t the way to convince people of your point of view. It doesn’t add anything to the discussion.

2 Likes

If anyone wants to propose and/or poll on a viable compromise option I think that could be valuable. These three options represent the major different directions we could go in but we can definitely take a more nuanced approach.

1 Like

I have had to think about this one somewhat.

I am going to abstain for the following reasons.

  1. Raising the surplus buffer only makes it so that we don’t buy MKR high and sell it low if the protocol loses money. I consider this unlikely - there are lots of reasons to raise the surplus buffer - what is going on with the USDC LR 101 vaults isn’t one of them.
    .

  2. Lowering stability fees. One could put this to 0 to buy as much time as possible but basically all this does is allow free minting at 1.01. From a system loss perspective it probably is the best option but it doesn’t clear the debt and also gives the system no fees.

  3. I have been for a new vault but not at the same 1.01LR. Also it isn’t going to migrate anything.

Let’s put this and my CC comments in perspective.

The bet here is that ‘at some point’ in the future DAI will trade lower than 1.01 with some liquidity and these positions can either be liquidated or closed out without the system losing money and possibly ‘in profit’. This is the bet that has been taken and 1% is the only buffer we have. Let be clear this 1% buffer IS only active if the DAI PEG price is 1 not 1.01 - at 1.01 there is NO buffer and any fees being generated by these vaults are ‘as I said in the CC’ an accounting ‘trick’ simply because if these vaults were auctioned right now the current price is 1.008USDC for 1DAI. This ignores the fact if these were auctioned/closed they would have to source DAI liquidity from ‘somewhere’ probably driving DAI price up again.

The idea that we have till 12/19/2020 magically somehow is premised on an idea that DAI is at PEG with 400M of liquidity to satify closing of these positions. I am thinkig the likelihood of this event happening to be quite low. But ‘there is a chance’ even if it is growing lower every day.

The right thing to do might be to simply lower the SF on this vault and don’t allow others to be opened. (i.e. if another vault type is created - or a vault structure I would suggest a structure as I have advoccated previously [Signal Poll] Use multiple stablecoin (e.g. USDC) vaults with different LR/DC combinations to manage value/price/risk exposure )

The whole point of these is to ‘attempt’ to manage the PEG with larger amount of capital at different price points (LRs) - not to lump everything into a heap at 1.01 which is exactly what was done.

Right now without a buyer who would take this 400M USDC off our hands for even 399M DAI the protocol is stuck. The only reason the PEG is at 1.01 is because there is still DC open in any of the LR 101 vaults. reduce DCs on these vaults to 0 and replace them with vaults at 101.5 and the PEG price is going to rise. Why? Because put simply it is easier for people to just use a vault to get DAI and abandon it vs. buying on open market. These vault owners get a kind of free DAI short position bonus (i.e… if for any reason DAI drops to 1 or below - they get to close - possibly at a profit). Unless something crazy happens I am skeptical the markets are going to provide liquidity for these vaults to close particularly when they can make more farming.

I don’t see this changing in 1, 3, 6 or even 12 months here. Now one can ask a better question (I hear @rune in my head on this one). Why is it bad for the protocol to have all this USDC backing DAI at 1.01? If effect if we take away the SF on this to make the facility DAI neutral to the surplus accumulation this vault with infinite DC acts as a decentralized PSM which will buy below 1.01 and sell above it.

Turn this around another way. Why not give this vault a very LARGE DC and lower the LR even further. Vaults under water won’t be able to borrow much more. Eventually move this LR to 100 or even say 100.2-100.4 and simply let the PEG move down even further. In the loosest sense this facility and approach has been VERY good at managing the PEG at 1.01 +/-.003

The prime issue with the system here is lack of generating any return from the now 400M USDC in vaults. If there was a suggestion it would be to take this USDC and pair it with DAI and drop it in the most liquid and lucrative farming contracts to earn some return for the system. If people are going to give us their USDC in vaults - why not put it to work. (I understand this is in complete contradiction to how Maker works with vaults - but if people give you 400M USDC to do something with - perhaps the onus is on us to DO something with it).

It is a good thing to have 400M USDC in case the PEG ever does drop below 1, and if RWA comes on line and wants to swap their DAI for USD.

The only issues for Maker here are:

  1. Accumulated SF will eventually eat the 1% - so lets put this off as long as possible.
  2. Earning return on this USDC vs. letting it sit there doing nothing.
  3. The already taken on risk of blacklisting.
1 Like

I like this idea however I feel there could be some downsides to that like the following

  1. Currently we hold the custodial risk, when investing we would swap it for smart contract risk which I believe to be higher (perhaps using Nexus for that would be prudent but then again they are not enforced to pay back although not doing is against their incentive structure)

  2. The main difference between maker & secondary lending platforms is that we keep the collateral meaning in principle one may interpret it´s safer, how do we know that vault owners won´t want to get their collateral back? They did not agree to have their funds used when depositing. Infering is kinda risky. Perhaps if there as a way for us to have a confirmation from them & share the profits I could see it favorably (we could even use yearn for that)

In summary if there is a safe way without a PR backlash, in agreement with vault owners and solely to be used for non volatile crypto assets I could in principle agree

Is that really a big risk?!
Just thinking at 1.01 that means we sort of have an unknown 0.01% reserves. Around 10m.

An other option is to keep the usdc to be liquidated as reserves. Probably as good as dai in case of issue on the protocol as the dai will probably plum.
This reserve can be put as maker man suggest into some farming or whatever uniswap you want.
Keeping dai as reserve is not really a great idea if we want to put the peg down.

I hope you will apply the moderation in the same way when someone makes a positive but meaningless/spam comment like “I can’t wait to see that implemented” or “New Vault Type :loudspeaker:”… Otherwise it is just plain censorship of a different opinion/view.

The “low effort post” is my explanation why I voted for a certain option (abstained). I could’ve rephrased the statement so that it does not offend people who think that the system works fine/as designed - I think it would not be censored then. I mentioned the problems with the current DAI design elsewhere so it would be off topic to repeat that here.

It would be offtopic to repeat that here. I think I wrote about the problems a few times - maybe I should create a new topic but I’m not sure whether it will be censored or whether anyone cares about the different/critical view.

I think that the burden of proof is on you to show why you think the system works fine/as designed when there is no peg, signal requests for patches like this one every few weeks, low governance turnout, SFs copied from competition manually…

For a solution you might check the forum/signal proposals (not just from me) that have been rejected.