Signal Request/Discussion: Should we implement a tax during SCD Shutdown to recuperate unpaid Stability Fees?

  • whatever we do, somebody will be pissed
  • i would even be ok, with refunding people that payed fees during migration (from buffer). I am not sure how difficult is that technically. I also assume most of the community would be against it.

Sure. If I have to choose, then I want to piss off the people that don’t pay attention or are trying to get out of paying the fees that make the system work. Don’t piss off the good actors.

3 Likes

The people that will be affected by the tax are not all bad / lazy actors. Remember that some Sai has been irrecoverably lost, everyone can’t close their Sai CDPs, and that others may have good reason to not be giving this their active attention.

Using the Tax parameter will cost some people more than the Stability Fee they currently owe. This is the Maker Protocol taking their money by breaking its own rules when they have done nothing wrong (because the tax will definitely catch some innocents that aren’t paying attention / can’t source Sai). This is objectively worse than the Maker Protocol taking your money because you (and it) followed the rules as they were defined.

Yes, some bad actors will ‘get away with it’ if we don’t implement tax. That sucks, but it doesn’t mean that we should screw over the innocent to punish the guilty. I am near certain that doing so would be worse for MKR Holders in the long term.

Furthermore, how pissed are you really going to be because someone else broke the rules and through some messed up game theory got some free leverage? You haven’t lost anything other than money you could have had if you’d acted differently in some ‘what-if’ universe. In this situation you’ve lost nothing you hadn’t expected to lose. Someone else has gained something that you haven’t gained, but you are not worse off than you expected. You are only losing out in comparison to others, not on an absolute basis.

3 Likes

@Sirlupinwatson1 unless I’m misunderstanding you I’m not sure your last post was very relevant to the tax discussion in this topic.

Could you please move it to it’s own topic or find another that is more appropriate?

Sorry about this then.

I can’t move treads. Only Mod or Admin can do this.

I can delete it.

Here is a list of the most likely scenarios we have going forward. Please let me know if my proposals are flawed in any way and also your thoughts and opinions on them. You may notice some repetition within the explanations. I am just trying to fully articulate to all readers the implications of each scenario. After some discussion, I plan to create a poll after the governance meeting on Thursday, January 30th.

1. Lower DC, close when one of the parameters is hit (6 months, 10M supply, .05 peg deviation)

Explanation

Easiest plan logistically. Debt ceiling must be lowered to 0 in order to prevent system exploitation. This strategy is in line with the progression of signal requests in the past month in regards to determining proper shutdown parameters. A summary of this topic can be found here: Summary/Discussion of Parameters for SCD Shutdown

Maker community could still try and maintain peg through stability fee voting for good actors but it would be unclear how effective this would be, knowing CDPs can just wait to settle at GS and not pay any fees, peg may break.

Example

MKR holders vote to lower debt ceiling to 0. On-chain poll is crafted to include threshold parameters. On-chain poll passes, signalling governance facilitators intention to generate an executive vote for GS once any of the threshold parameters are crossed. MKR holders then vote to execute when available. CDPs wanting access to collateral during this time period will still owe stability fees. CDPs that wait until GS will be settled at the current oracle price of ETH/USD and will not owe any outstanding fees.

Pros/Cons

  • Community has clearly defined shutdown parameters
  • MKR holders gain to lose the most potential stability fees
  • CDPs can wait until GS to claim interest free leverage
  • No new Sai can be minted during this time period
  • Possible liquidity issues

2. Swap `Fee` for `Tax` , shutdown in **x** months

Explanation

It is known that at global settlement, CDP users who wait until then will not have to pay their owing stability fees. In a strategy to circumvent this from happening, Maker governance could vote to swap the Fee function for the Tax function. This would ensure that CDPs would be charged whatever % the Tax amount ends up being. More details on how the Tax function works can be found here: Global Settlement of Single Collateral Dai

Implications of this method are that CDPs would have no incentive to pay off their debt because doing so would incur past stability fees along with current tax fees instead of waiting until the end where they would just pay tax. There may be reasons they would want access to their collateral before then however so it is not a sure outcome.

Example

MKR holders vote to implement tax February 1st. Shutdown is voted to initiate on August 1st. CDPs who pay back between this window would still owe the equivalent % of stability fees outstanding and any taxes accrued after February 1st. CDPs that wait until GS will only owe the taxes.

Pros/Cons

  • Shutdown date is known (certainty)
  • Debt Ceiling doesn’t have to go to 0 (more liquidity)
  • CDPs are incentivized to pay back before implementation (SF) or wait until GS (Tax)
  • CDPs are disincentivized to pay back before GS (SF+Tax)
  • Requires communication of changes to community.
  • Public Relations Issues

3. Shutdown in x months, flash `tax` at the end

Explanation

With this method, CDPs would still accrue stability fees like normal until global settlement. In the week before GS initiates, MKR holders would vote to swap fee for tax and proceed to hike the tax rate to a high % to be equivalent with what the stability fee would have been during this time frame. Tax % would be a blanket amount, so governance would have to determine what a “fair” % would be. If properly communicated to CDP holders and the broader crypto community, CDPs would understand that they would not be able to escape paying their fees and be incentivized to close before GS.

Example

MKR holders vote to initiate GS August 1st. 4 days before GS, MKR holders vote to swap fee for ‘tax`. MKR holders vote to increase the tax %1000, so over 4 days APR fee would be 11%. GS initiates and remaining CDPs are automatically settled with 11% being taken off and added to burn PETH, thus diluting their claimable ETH amount.

Pros/Cons

  • Shutdown date is known (certainty)
  • Debt Ceiling doesn’t have to go to 0 (more liquidity)
  • CDPs still accumulate SF until GS
  • Tax hits all CDPs equally, regardless of age
  • CDPs are incentivized to close before GS
  • Must be communicated effectively to community for PR reasons

4. Shutdown in x months, communicate an imaginary `fee` / `tax` swap, flash `tax` at the end

Explanation

This is a hybrid of option 2 and 3. Option 2 swaps fee for tax right away but the consequence of doing so is that from thereon out no MKR would be burned through fees. In this proposal we would be starting the counter on tax on a set date in the future but wouldn’t enact it until several days before global settlement (option 3). Through this proposal stability fees can still accrue and CDPs that want to migrate/close in this time frame can do so by burning MKR. Otherwise they can dilute PETH/ETH through tax at the end for the equivalent SF %.

Example

MKR holders vote to initiate GS August 1st. Maker Foundation/Community conveys the start date of tax beginning February 1st. Community members keep track of current SCD SF and 4 days before August 1st calculate over the past 6 months what the equivalent tax would be. MKR holders vote to change fee to tax July 28th. Say SF was 8% for 6 months. Tax equivalent over 4 days would be 365% APR. MKR holders would vote to raise tax to 365%. August 1st GS would initiate. All CDPs closed between February 1st and July 28th would have all previously accrued fees as well as fees from the final 6 months to pay while all CDPs closed at GS would be taxed a flat 4%.

Pros/Cons

  • Shutdown date is known (certainty)
  • Debt Ceiling doesn’t have to go to 0 (more liquidity)
  • CDPs still accumulate SF until GS
  • Tax starts from specified date.
  • CDPs are incentivized to close before GS
  • Requires the most communication
  • Has the potential to capture the most stability fees
  • Requires the most governance overhead

5. Lower DC, manage system for 6 months, stop supporting after 6 months, let collapse on its own, determine GS sometime in the future.

Explanation

Lowest overhead while still having the best chance of recovering unpaid stability fees. This method would signal that while it has been MKR holders duty to maintain the SCD system, we have given ample time for CDP users to migrate. We would promise to maintain the system for several more months before letting the system run its course until an unknown time when chaos has overtaken and we have no choice but to close down. This proposal is similar to option 1 but we would not close if the peg breaks or supply dwindles. Essentially it would no longer be our problem. This would give us the longest opportunity to let CDP holders migrate/close down on their own and allow MKR holders to capture more potential fees. Other considerations would be to keep the debt ceiling higher for market makers and liquidity providers to help maintain the system.

Pros/Cons

  • Shutdown date is unknown (uncertainty)
  • Debt Ceiling must go to 0 or be below Sai supply (less liquidity)
  • CDPs still accumulate SF until GS
  • Lowest governance overhead
  • CDPs are incentivized to close before GS
  • Peg may destabilize
  • General liquidity issues possible
  • PR implications unknown

Glossary of terms

GS - Global Settlement (Emergency Shutdown) - https://community-development.makerdao.com/makerdao-scd-faqs/scd-faqs/emergency-shutdown
CDP - Collateralized Debt Position - https://community-development.makerdao.com/makerdao-scd-faqs/scd-faqs/cdp
SCD - Single Collateral Dai

Looking forward to discussion below.

Edit: added pros and cons to options and glossary.

5 Likes

I would say that it is worth noting that Supply and Demand is not the only mechanism through which the peg is maintained when ES is involved.

When ES happens 1 Sai is redeemed for $1 of collateral. If everyone knows ES is coming then it doesn’t make sense to buy Sai for greater than $1, and it makes quite a bit of sense to buy Sai below $1.

This effect is stronger the closer and more certain the future emergency shutdown is deemed to be. I think that if we announce a date then we will not have too much trouble with the peg regardless of the debt ceiling setting. Unless we do something else that could trigger a liquidity crisis.

We don’t want to make Sai CDP holders afraid of Emergency Shutdown. If they think that ES is worse than buying Sai and closing their CDP then rush for Sai will cause the peg to break upward.

In that regard, what shut down method do you think makes most sense?

Probably option 1. I think the ES game theory will maintain the peg, especially given that there is no real need for users to migrate considering that the fees will be forgiven on shutdown.

I also think we should announce a date sooner rather than later. I think it will put a lot of minds at ease. At least once a week someone asks when it’s getting shutdown.

Options 2, 3 and 4 involve us having to explain the tax parameter to the space, and convince people it’s the right way forwards. Even if it weren’t for the PR nightmare, I think we need to define what we are trying to achieve.

Option 2 leads to most of the fees going to PETH holders, there is no incentive to pay both, so most users will hold until shutdown and pay the tax to PETH holders.

Option 3 has the best chance of recovering the fees, Options 2 and 4 are middle-road options that I don’t think are worth the cost of communicating tax.

Option 4 has a weird outcome where depending on the currently owed stability fees for a CDP, each CDP is either incentivised to close before the tax starts, or wait for ES. Not sure this is good, as I said, middle of the road option, recovers some fees, less chance of a liquidity crunch, still need to explain the tax parameter to the space. Imo not worth the effort given that we aren’t going to recover all of the fees.

Options 2, 3 and 4 increase the chance of a liquidity crunch of Sai, forcing us to shutdown early anyway. Can you imagine going through the comms issue arround the tax, only to have the Sai price balloon into insanity as everyone tries to close, and we’re forced to shutdown early after taking the tax PR hit. Worst of all worlds if that happens. No significant fee recovery, damaged credibility, much governance panic.

Option 5 feels like abdicating all responsibility which I don’t think it right considering the implicit promises made in the past. I don’t want us to be in this grey area of ‘oh, we haven’t shut it down yet, but we don’t support it either, it’s just sort of there.’ Even if we tell everyone we’re not supporting it, people will still look to us when it breaks. Human nature at its finest.

2 Likes

Great summary. I’m still debating between the options myself.

2 Likes

I also do favor option 1, i don’t get why tax is benefiting PETH-Holders in the first place as it feels like a replacement for the fee (and this is benefiting MKR holders)

also i do not see a reason to wait for this, why weeks not months? :wink:

The tax is not so much about who it benefits, but more who it penalizes. The reason we have to set the DC to 0 in option 1 is because otherwise people will be able to mint SAI, migrate it to DAI and stick it in the DSR for free money. With a tax this is no longer an option. CDP owners from this point on (assuming option 2) will have to pay the stability fee just like every other person.

1 Like

My basic question is whether anyone who has deeper knowledge of the entire SCD system has any other options beyond what has been laid out above? I have this feeling there are other options but we simply have not had them presented.

I also wonder if it makes sense (or if people agree) that it seems the community favors the SAI holders vs. CDP creators in the GS or whether there are mechanisms that could be used to incentivise not just CDP holders but SAI holders to convert that we are missing.

IF after the above is answered with no additional information or ideas and I am left with the choices posted by @Aaron_Bartsch I also would favor his (1) I think. As I have a hard time doing analysis on the above scenarios due to lack of sufficient depth of understanding of the SCD mechanics I’m going to defer to others with more knowledge on this topic.

Last comments on regarding the migration. So far I think it has gone quite smoothly and at or slightly faster than the pace I expected I think IF MCD SF’s rise and the DSR rises with it above 8% we should start to see some of the Nexo SAI deposits move to DAI and into the DSR. In principle IF the DAI savings rate with DSR beats all the secondary markets (SAI/DAI saver carrot) we should see the maximum movement of SAI to DAI and this should provide liquidity for CDP closures. This event should it unfold should give the Maker community a solid view on how much SAI is going to move via a reward vs. what will be left that will have to navigate the least punishment options.

I don’t see this as a - we need to collect the revenue issue - but more about being as fair and open as possible with the entire greater community (SAI and by proxy DAI savers, as well as SAI/DAI borrowers) and not just Maker holders.

My .02DAI opinion to go with my few MKR that does vote on occasion.

3 Likes

While reading in chat found the following comment by @mrabino1

https://chat.makerdao.com/channel/governance-and-risk?msg=KL57urF9KsdAswvuJ

I have been maintaining this opinion as well. There are so many people that are like close the SCD down now! The community has given enough time to CDP holders to close. Really? Honestly if I’m a large CDP holder right now I’m still waiting to see what this community is going to decide because it is starting to look like I’m going to get out of my previously accumulated SF fees at a minimum. We honestly don’t know dink about the remaining CDP holders and I loath the idea that people who got out of the CDPs early by paying fees will basically feel screwed for doing what Maker considered ‘right’. I want them to feel rewarded.

People wonder why the SAI PEG is so robust. I think it is due to the bridge still being active with at least ‘some liquidity’ but this is speculation. I also think LFW concern over a SAI liquidity crunch is reasonable. Set the DC to zero and this may start to cause issues that could become impossible to manage and could float over to DAI.

Lastly your comments above regarding ‘what is the goal?’ - I echo them all and add the questions above.

One thing I really don’t see being discussed is the following proposal that I will write from a PR perspective as an example.

Maker today has decided today to do the following with the SCD.

  1. Drop the DC continuously so that no more than 4% extra SAI liquidity can be minted at any given time. SAI outstanding as of today is $25M and we are reducing the DC to $26M and will continue to lower it (dynamically if necessary or possible). The SCD DC will not be going up.

  2. Keep the SCD open indefinitely with the following caveat. The SCD GS can occur at any time any of the following conditions are met. SAI outanding below $10M, the SAI PEG off by more than 5%, or longer than 6 months has passed. There will be no announcement of a specific closure date and this can come at any time any of the above conditions are met.

  3. While the SAI:DAI bridge remains open, fully bi-directional, and with a 1:1 exchange rate this can change at any time. First to being a uni-directional bridge, and finally with a exchange rate that is different than 1:1

  4. Expect that on shutdown that all CDPs that have not closed will have a fee via a tax applied that is intended equal or greater than the sum total of the SF fees because we want people who close their CDPs early to expect that they will end up paying less by closing early than waiting until the GS.

  5. the SCD-SF fee will condintinue to rise at the rate of .5%/week or 2%/month until GS.

  6. It is in the spirit of cooperation and open communication that we make the above announcement today regarding the status and future of the SCD. We strongly urge SAI holders and in particular CDP holders to migrate. This is the last call to SCD/SAI users/holders! There will be no other warnings or PR regarding this issue until Maker decides to execute the Global Shutdown of the SCD system We want to be clear at some point in the future the SCD will be shut down but the community has decided the ‘exact’ details of the shutdown will not be discussed or even decided until the event happens.

When I think about the gamability of the above by both sides of this (SAI holders and CDP creators) I honestly think the above IS the best option.

Why? Because if we hard cap the DC to no more than 4% above outstanding (and drop this only if the outstanding drops sufficiently) CDP holders will have no way to get money or at their collateral except by paying down their loans and migrating. There really are no other options for them And if they see that SCD-SF fees are rising.

SAI holders basically are stuck waiting for potentially quite some time AND with significant PEG risk before they get anything in the GS. The fact the bridge is put into this and that there is no ‘formal’ date for shutdown. What exactly is there to game here except a possible liquidity crisis causing SAI price to rise above 1. I am less worried about the SAI PEG going up because once 1.05 is hit this triggers possible GS action to just close it down. The SAI PEG going down will mean CDPs can close at reduced cost and won’t go below .95 anyway.

Honestly if one said the SCD would stay open with increasing SF fees indefinitely and a DC that is going to do nothing but drop and a looming GS that has a community which has signalled it will apply a tax and the idea that this could go on for years. I think the above would lead to enough uncertainty and mounting costs to drive people to exit SAI. I also think if we can get the SF and hence the DSR to something north of 9% we should see another strong drive to exit the system. The only people who are going to be left are ones that have lost keys, or are dead, or don’t care. Anyone else - well I read the Maker documents and if they claim they didn’t know the hazards it is to me like someone saying they are ignorant of the law.

Sorry if this is long but on second thinking on this. I am honestly back to keeping the above alternative options to any of the other options presented on the SCD.

4 Likes

How would we ratify/draw consensus on these parameters when they are uncertain? All of the other proposals can easily be voted on in a governance poll on-chain while this one seems more complicated as we aren’t certain if and when we would actually initiate GS or even a tax. So isn’t this proposal just kicking the can down the road to make decisions later in order to create uncertainty for CDP holders? The best we could do is a loose, non-binding agreement to uphold these edicts. This is due to the nature of continuous voting and the way system parameters are controlled. Each time we want to change a parameter it requires an executive, and this seems to require a larger amount of continuous executives. If you can think of an elegant solution to my concerns I’m all ears.

1 Like

Off the top of my head response here on the easy parts and will have to mull over the more difficult parts. Apologies for length.

  1. Decide how the DC will be managed - make the DC downward management basically automatic so once we decide this we probably won’t have to mess with it again in governance. I was suggesting a 4% buffer putting the DC at $26M, cyrus was advocating $30M or a 5/25 approx 17% buffer - perhaps 10% is a good number (DC to 27.5M currently). I think we all agree the SCD-DC needs to be pretty tight to the outstanding here to limit SAI minting but if we threaten at tax the exact headroom should be moot. I think we want a tight reign on SAI minting at this point to be no more than 10% of the total outstanding. I have been thinking we should just debate this in forums and when we here come to a consensus then put the value up for yes/no governance poll.

  2. Vote that GS on the SCD happens according to the conditions laid out in (2) that I think we at least have some forum consensus on - or need to check again. The point here is that this is a vote effectively to keep the SCD open for as long as possible to give everyone the chance to close out CDPs before the GS. There will be no ‘date’ set other than one of the conditions is that after 6 months the SCD can close at any time at the discretion of Maker governance. In fact the more I consider the 6 month ‘condition’ I think we should just remove that. SCD stays open until $10M SAI is hit, or PEG is lost and then it closes as the pleasure or displeasure of Maker governance but not before. This is to give more than enough time for people to show up and close. One could argue that we could wait one more month before putting the SCD-SF rates into ramp mode to give everyone fair opportunity to close with the least cost.

  3. Not at all sure how governance handles the Bridge mechanics. Leave that for now.

  4. If we are going to vote for something - There are only two pieces here. Yes/no on levying a tax. And yes/no on the amount of the tax. I think prudent here is the extra fee/tas being equal to the largest CDP debt to loan ratio so the cost to CDP holders to wait until GS is > then the accumulated SF fees. Again best if we just get consensus in the forums about having a tax. How much to levy to be discussed. I want to be clear here - the message being sent to CDP holders stays consistent. The earlier you move the more you will likely save. Wait for the GS and all the old CDP holders will basically pay their accumulated SF + the fee, anyone who entered later will basically be punished more because the fee applies equally to everyone. I want to leave the amount and mechanics of the tax at the end and a possible redemption still open to discussion. Key point is that the message here to CDP holders is that they will not avoid their fees and if they wait for the GS they may face additional costs/fees so it is better to pay early than late. Someone posted that we could have a kind of redemption system (manually executed) which would take the ‘tax’ or ‘extra fee’ and put this into a fund to try to balance out some of the GS fee/tax inbalance due to what I’m proposing.

  5. This one should be pretty easy - we debate the automatic SCD-SF increase rate %/month or %/week and start date of this ramp once and then just bundle the actual changes into governance. Monthly/weeklyraises due to governance cadence might be more appropriate as I’m not sure spells cast can actually implement automatic changes (for instance to ramp the SCD-SF X%/week, or to lower SCD-DC under condition X by amount Y) without contract/governance changes.

In my proposal here is what I would like to see clarity develop around.

  1. The SCD will be shutdown - at some point!
  2. Conditions needing to be fulfilled before the GS could be activated. (If these conditions are not met the SCD should be open)
  3. SCD-SF on a steady and slow ramp up and always higher than the MCD ETH-A SF rate after some future date (3/1/2020 for example).
  4. That a tax > the accumulated SF fees will be applied across the board.
    4a) Whether or not there is a kind of rebate using the extra fees is done left unknown for now but under consideration.
  5. The DC is only going to go down and the available DC headroom will never be more than X% (say X=10)

What we want to keep unknown to some extent.

  1. The exact value of the tax other than it will be > than the accumulated SF fees and be equally applied to all CDPs regardless of their age and ideally will be in amount so that even the older CDP pays > accumulated SF fees on their loans during the GS. In effect what one wants to do is collect enough fees from the remaining CDPs outstanding to cover the SF on all the outstanding DAI all the way back to the oldest CDP that is reasonably large (say use CDP 3088 as the reference to determine what fees to charge everyone and then rebate the excess if possible to people who paid excessive fees at the end).
  2. There is no formal ‘date’ for shutdown only two conditions, outstanding SAI, PEG off by more than 5% for say a few days?!
  3. Whether the bridge will stay up, bi-directional, and 1 to 1 ratio until GS.
3 Likes

I am strongly opposed to Option 1. My preference is currently for Option 3. Here are some dot points summarising my thoughts:

  • Good actors in the MakerDAO system who have converted their CDPs from SAI to DAI should be nurtured. They will be pissed off knowing that they would have been financially better staying in SAI all along. They will rightly be dubious of trusting MakerDAO into the future.

  • Option 1 sets up a precedence not to trust MakerDAO and that there is no punishment for behaving badly. It may even convince users that MakerDAO would never go through with an Emergency Shutdown to punish a bad actor. MakerDAO needs to set a precedence of being fair but tough when the requirement is there.

  • Option 3 provides a clear narrative that the earlier you transition to DAI, the greater the benefit will be. The narrative will be that smart actors will transition early and that stragglers will be cleaned up at the end in an emergency shutdown.

  • I am strongly in favour of having a set date of GS instead of a GS when ‘one of x conditions are met’. I believe that a large portion of CDP owners have preference to close their CDPs naturally and an early GS would cause confusion and may seem to some users like they have been scammed.

  • There would be a large portion of SAI CDP owners who intend to transition to DAI but have not done so due to laziness. If we communicate the conditions of the tax system and a date for GS, they will happily pay what they owe and transition across.

  • I believe lowering the debt ceiling to 0 is potentially destabilising and may trigger an early GS. This is very undesirable.

  • SAI CDP holders have been warned from the beginning that they are entering into a Beta system. They have been aware that they are paying fees to use the system. No one can rightfully be pissed off that after utilising the system to the very last moment at GS that they are charged what they owe (plus a penalty). There will be a small percentage of people who are unaware of the whole process. However, they will not lose anything more than what they should have paid in the first place.

  • Note that I strongly agree with @hexonaut that the tax is not about recouping fees for MKR holders, but penalising SAI CDP holders for abusing the system to the last moment.

I also have two questions, the answer to which might change my opinion.

Question 1: Say for example that Option 3 is implemented with GS scheduled for July 1st 2020, User A has opened a SAI CDP on July 1st 2019. They put in 1 ETH and they have drawn 1 SAI. User B purchases 1 SAI on the open marketplace.
At the point of GS, User A’s CDP is Open and ETH is trading for USD100. What are User A and User B entitled to claim after GS?

Question 2: I believe that Option 3 would have the effect of wanting to increase the peg of SAI above 1 as CDP owners search for SAI to close their CDPs. Does a 1:1 SAI:DAI bridge, game theoretically prevent the SAI peg from rising much higher than the DAI peg prior to GS? ie. If SAI is at 1.05 and DAI is at 1.00. Can a user print DAI convert to SAI, then sell for ETH and sell for a quick ~5% profit?

edit1: fixed some erroneous uses of SAI vs. DAI.

4 Likes

You classified users who migrated to the ‘good actors’ class and people who did not, to the ‘bad actors’ class. Convenient, but seems discriminatory.

Some additional thoughts on the gaming of the SCD here for some perspective. Thanks to @mrabino1 @lix @LongForWisdom @Vishesh for insightful discussion.

A quick summary of the below is as follows:

  • Keeping the SCD open effectively with a tight DC limit effectively punishes both CDP and SAI users equally. Closing the SCD rewards both participants as discusses below.
  1. Restricting the DC to be very close to the outstanding means that CDP holders will not be able to extract value on their collateral and that the only way to access collateral between the 100-150% liquidation limit is effectively to pay down debt with SAI. In fact since most CDP holders don’t want to face liquidation they keep more collateral than the minimum and basically can’t access the additional collateral above the 150% liquidation mark without risking liquidation. In effect without the ability to mint more SAI on their debt CDP holders collateral is held hostage while the SCD is open and SAI SF fees + debt is not paid.

  2. SAI holders while the SCD is open basically are facing greater and greater PEG risk. Sure it could go above $1 (good for them and liquidity probably will show up to sell if the price is > 1.01). What SAI holders are betting on here is the idea that after the GS SAI effectively becomes a non-expiring in time option contract on ETH at the SCD GS oracle price. If the ETH USD price on the markets is < than this price everyone will keep their SAI which will be trading < $1 and not redeem for collateral at an effective loss. IF the ETH USD price is > oracle SAI will trade > $1 and will exist as an option to redeem collateral in the future. In the loosest sense then SAI becomes a new proxy on ETH and will trade against the GS oracle price above or below $1 depending on ETH USD price.

While the SCD is open there is no ‘option’ to redeem for ETH and the PEG hazard exposure if the GS happens the SAI PEG will float (and do this in rather ill-liquid markets btw) but then people basically can hold SAI to claim interest while waiting to exercise the price option. So there will be a price gap between the actual redemption value and the premium (or lack there of) based on future price perceptions of the market.

  1. During the GS basically all CDP holders get a free sell of their ETH collateral at the oracle price - without fee. They then get access to all of their collateral above the $1 SAI borrowed for $1 ETH oracle price (within 3 days of the GS). In effect the GS frees up at least 33% and probably much more ETH collateral for withdrawal without having to pay anything. The other bonus is that SF fees are avoided if no tax is applied in this scenario which for some could be substantial. The GS for CDP holders is basically a win for them. During this time if the price of ETH trades below the GS oracle price they basically get the option to free their ETH at a discount. If ETH trades above their collateral sells below market as SAI redeems (if it redeeems vs. waiting).

In summary it looks to me that the worst case for all parties involved here is to leave the SCD open as long as possible because then the SAI holders have to hazard PEG risk and don’t get their ETH exchange option. CDP holders face excess locked capital that they can’t borrow against and have to face liquidation hazards. IF the DC is kept tight and creeping downward, and the SCD-SF is steadily on the rise I think CDP holders over time are going to slowly pay their fees to free collateral because depending on the total fees every $1 SAI they pay frees up $2-$3 worth of ETH. SAI holders facing PEG risk over time may also find reasons to migrate their SAI to DAI.

As the summary suggests for both parties here keeping the SCD open is a worse case scenario for all parties than doing the GS.

There has been a lot of discussion regarding the PEG.

https://chat.makerdao.com/channel/governance-and-risk?msg=aPCSXDe9apESuksg6

While there are some scenarios above that reveal what could drive the SAI PEG off one I am thinking some real thoughts about the scenarios driving the PEG off and what mechanics could come into play to bring it back to $1 (or not) really need to be discussed here. The PEG being off by 5% being one of the critical GS criterion I’m suggesting and will follow up later as I am able to work on this. I think the $10M outstanding is a reasonable debt point mostly because we have not seen any activity on CDP 3088 which has what 8.28M outstanding. https://mkr.tools/cdp/3088

I like your post The_Ratj and agree with your first point regarding Good actors. I also think it may be a bit harsh to call the remaining that has not moved as ‘bad actors’ as we don’t know the reasons people are holding opne CDPs. The issue of semantics is important and avoiding derogatory terms for classes of people here is probably a good idea. I mostly agree with all points except for wanting a fixed shutdown date.

Anyway to try to answer your question (1). User A on GS basically owes $1SAI + Tax presumably they will only be able to withdrawal $100-($1SAI + Tax) worth of ETH. User B basically gets .01 or $1 worth of ETH. PETH holders end up with the Tax difference.

question (2): Generally I think a lot of these options will have an effect on the PEG. The whole issue with the bridge is that it has to have SAI liquidity for DAI to be able to arb. My general reaction is that as long as there is liquidity the bridge will and does act as a PEG stabilizer. Without liquidity it does nothing. I’m still mulling over PEG mechanics here. PEG < $1 and borrowers get a discount to close provided they can find liquidity. PEG > $1 and SAI holders come out to arb the price gap > $1. The only issue here is whether the SAI holders value holding their SAI so that during the GS they effectively get a non-expiring price option on ETH collateral vs. the % price arb. My thought here is that a substantail number of SAI holders like the idea of getting interest on their SAI and it becoming a ETH collateral option at the GS oracle strike price.

In general I think before the GS this is going to cause the SAI price to rise > $1 and as long as there is SAI liquidity in the bridge or CDP holders can still mint some SAI there should be arb opportunities but in general thinking about this I think that the end game here is tilted in SAI holders favor, and against CDP holders to be able to close and this means we are going to see a rising SAI PEG. This is going to depend on what bets people are going to make on the GS ETH oracle price during shutdown and whether people holding SAI think this will go up or down after the GS.

It is the primary reason I think NOT putting a firm date on the GS but conditions that can trigger a GS is a better option for the community in the SCD GS shutdown game being played here.

I am going to add as a SAI holder earning 8% in nexo I plan on holding some SAI (for better or worse) just to get that ETH exchange option and the sooner the GS happens the better. If I thought Maker would just keep the SCD open, wind down the DC and steadily increase SCD-SF rates might I even consider taking out my SAI into a PEG > 1 event. My reasoning here is that the value of SAI as a indefinite ETH option contract is FAR more valuable if I think the GS is happening soon than if I think it might not happen at all. This would be particularly true if I can earn more than 8% elsewhere by converting that SAI to DAI and I am losing an oppportunity cost by waiting for a GS that might not happen for a year or more for me to be able to fully unlock the option value of SAI to ETH.

Let me be clear here. In my humble option: The worst case for all SCD players is for the SCD to remain open with no firm GS date with a winding down DC and steadily rising SCD-SF fee rate and a looming GS tax should the GS ever be triggered.

1 Like