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

Based on ongoing discussions on implementing the shutdown of our legacy Dai Credit System (SCD), the latest is in regards to how we may extract the remaining stability fees from CDP holdouts. If you are unfamiliar with this topic of discussion I encourage you to read/watch the following resources:

With these resources you should start to see the picture that is emerging on the decisions that need to be made in regards to SCD Shutdown.

With that being said, I believe it is important to gather signal on whether the community even wants the ability to recuperate the outstanding stability fees using the tax function, or would prefer to forgive or find another solution entirely.

Questions you should think about before voting:

  • Are we ok with all CDPs being taxed the same regardless of their current stability fee %?
  • What is the game theory behind these decisions? (PETH/ETH)
  • Are there other ways we can get CDPs to migrate/close?
  • How should we best communicate our intentions to CDP holders?
  • Will the broader community be upset by these circumstances? Should we worry?
  • Are we willing to forgive a certain amount of stability fees? (~$2 Million of SCD fees outstanding)
  • If we forgave stability fees would CDPs that already migrated be upset? Should we worry?
  • If a tax is implemented, should we worry about a Sai liquidity crunch in the event CDPs want to all pay back at the same time?

Should we implement the tax function near the end of a planned SCD Shutdown in order to extract fees from remaining CDP holders?

Please discuss below!

If you have any concerns with wording / format / or content, please let me know. I tried to be as unbiased as possible.

Edit: Due to the possible implications of this poll, some community members have determined that it may be unsafe to signal the results of this poll at this time. With that in mind, we have decided to postpone the poll until we can decide on an optimal game theoretic next step.

4 Likes

I’ve been giving this some more thought after the call today. At first I viewed this purely as a way to re-coup fees to MKR holders, but I now think it should viewed as a way to prevent abuse of the system. If we leave things as-is and announce a shutdown date of say July 1st. From now until July 1st you are giving people a way to mint SAI at 0% interest. Furthermore the minted SAI could then be moved through the migration contract to earn the DSR. Free money. This is completely undesirable behaviour.

The only mechanism I can come up with to counter this is to prevent minting of SAI by setting the debt ceiling to 0 which may have some other adverse effects.

I would encourage everyone to think of the game theoretic full picture of this rather than just “should we punish the CDP holders or not?”.

7 Likes

That’s a great point.

I’ll attempt to lay out my arguments here as to why I’m against this tax.


First and foremost, this issue is only coming up because of the way SCD was designed, whether it was an accidental oversight or an intentional trade-off against some other criteria, it doesn’t really matter. The point is we ( technically the Foundation but outside of this community I would suggest most don’t see a distinction ) are responsible for this situation.

If we implement a tax, then some percentage of CDP Holders are going to end up paying that tax. Game theoretically, this percentage should be low. The only reason people would get hit by this is if they’re completely oblivious to crypto news/announcements etc over whatever timescale we announce the shutdown.

The issue is that we are punishing these users for a problem that is not their fault. One can argue that ‘they should have been paying attention’ and there is an element of truth to this, but I think it is important to consider why these people may not be paying attention. I would argue it is one of the following:

  • They’ve lost interest in crypto to the point they are willing to abandon the value in their CDP.
  • They are taking a break from crypto and trust the Maker Protocol not to screw them over in their absence.
  • They are for some reason unable to spend their time actively engaging with crypto news / announcements etc.

Now for point 1, we have no problem. However for point 2 and 3 we are fucking these people over in order to fix a problem that we created. This is unacceptable. Under SCD, fees were only to be paid when a CDP is closed. If you had a very low liquidation price, you might well have stopped paying attention safe in the knowledge that fees would not be due until you closed your CDP, by implementing the tax we are breaking that implicit promise.


Secondly, if everyone has to close their CDPs by a certain date or else get hit with a huge tax, it will cause a liquidity crunch for SAI. As soon as we announce the tax, there is a reason for everyone to buy and hold as much SAI as they can acquire for under $1 and sell it once the liquidity crunch hits and SAI is trading well above a dollar.

Well! You may say, if SAI is trading above a dollar, we’ll just emergency shutdown early. But doing that would defeat the entire purpose of the tax and would mean that we’d have to go against the timeline we had already loudly announced. Cue huge backlash to absolutely no gain. (We still won’t see the fees because we’ll be forced to shutdown long before the tax hits.)


Thirdly, the optics of this look terrible. I’m still not sure that I 100% understand the game theory around the tax, if it’s that complicated then no one outside out community is going to make the effort to bother. Instead they will assume that we are:

  • Charging people twice for SCD because we are greedy.
  • Making other people pay for the flawed design of SCD.
  • Forcing people to migrate because we can’t be bothered to manage both systems.

This has the potential to harm the future income from MCD.


Fourthly, if we screw over CDP Holders in this way, there is a chance that they will be less inclined to take advantage of using MCD for leverage in the future.


Fifthly, the tax is not equitable. This is also horrible optics, and despite the game-theoretical fact that it should not apply to most users, you can bet that there will be a massive uproar from the people that understand the tax well enough to realise this.


Sixthly, there is unlikely to be SAI in the migration contract for this period, at least there certainly won’t be enough for everyone to migrate, which means we are not even forcing SAI CDP Holders to migrate, we are forcing them to close their positions and re-open them, potentially incurring taxable events.


In conclusion, this is an occasion where we should accept defeat. We cannot claim those fees without causing ourselves a whole bunch of damage. The CDP Holders that have held out until now have won. There will be plenty more fees to burn under MCD, we should cut our losses on the SCD fees and take a less dangerous path.

In my view the less dangerous path is the following:

  • Sai Debt Ceiling -> 0 ( Prevents more Sai being minted. )
  • Sai Stability Fee -> 0% (Irrelevant at this point since we aren’t going to see the fees.)
  • Set a date for Shutdown some months in the future. (We don’t want to be paying money to the DSR for Dai locked in the migration contract indefinitely.)
  • Make sure that MCD continues to be net-positive over the months until SCD Shutdown (Sai locked in the migration contract can accrue the DSR without paying the stability fee.)
8 Likes

I’m almost convinced, but isn’t there another group of people who would have a lot of regret if we went your way? Those who migrated early with the understanding that Maker would somehow recoup the SCD fees anyway will see that they could have stayed in SCD the entire time.

I have a few questions.

  • Would it be technically possible to set the tax such that it is either (a) too high for every user or (b) overall higher than the fees owed and then, using Maker surplus, make whole every overpaying address?
  • If yes, should Maker poll-vote on a promise to redirect some of the MCD surplus towards making those users whole?
  • If yes, could/should Maker executive-vote on a binding commitment to do that surplus redirection?

The last option seems like a large endeavour technically… but I thought it’s worth mentioning.

2 Likes

The differences are thus:

  • We aren’t directly taking anything from the people that migrated as expected.
  • We aren’t breaking implicit promises made to the people that migrated as expected.
  • Those users actively made a choice to migrate, meaning they share the responsibility for that migration.

It’s true that those users may be jealous or regret that they didn’t make the same high-risk high-reward decision as someone else who stuck with SCD. But ultimately this was their choice. We won’t have deprived them of something they have, but only something they could have had if they’d made different decisions. I suspect the sentiment from these users will average out at something along the lines of:

“Ah damn, I should have stuck it out, those bastards still in SCD must have had some serious balls to risk the liquidity crunch and shutdown.”

As opposed to people that get hit by the tax going:

“Those bastards at MakerDAO took my Ether away when I wasn’t looking.”

I think we could do this technically, but I don’t think we would want the governance overhead of having to make those users whole.

I had a response typed out about the 2nd part but then remembered tax goes to PETH so it’s possible to make overtaxed users whole, but not possible to end up in a situation where on net everyone including governance owns roughly what it should. My bad :slight_smile:

Are there any alternative plans? Otherwise too bad, but those who migrated early played cooperatively in a game where governance signalled that it would seriously look into making SCD fees real, and those who stayed played adversarially. It would be great to respond in kind if we can.

2 Likes

This argument has cropped up from a couple of people. I agree, it sucks to be caving in to people that are playing adversarially. But we need to remember our objectives. We aren’t trying to ‘win’ against those users. We are trying to maximise long term MKR burn.

I suggested an alternative plan above, but I admit that it hasn’t had much scrutiny.

1 Like

Fully agree that this isn’t about winning or losing locally, but it would be fantastic game-theoretically wrt other actors around Maker if we expended work to make good even on “aspirational promises”. It would also be net positive reputationally wrt the wider audience. That’s if there is a path to make good on those promises, of course.

What alternative plan are you referring to? Is it

?

1 Like

Yep, that’s the one.

Since the SAI Debt Ceiling will be going to an Executive Vote tomorrow at 45M—is there any way we can lower it to 27M instead? And then do a poll on Monday to grant the Risk Team the ability to reduce the SCD DC 1M at a time as CDP holders scale down? So the Risk Team can automatically scale SCD DC from 27M to 26M, etc. without needed the MKR holders approval? I’ve done some thinking… and I agree with LFW—this TAX thing is not worth it. We are rushing into GS and there’s no benefit to MKR holders. Let’s play the long-term game—after all, it’s only 27M exposure of SCD and maybe it will wind down on it’s own with due time…

1 Like

The tax, or not to tax, discussion reminds me of this talk from devcon by Brewster Kahle (fast forward to 9:15 for the relevant part): https://youtu.be/AyaWFicSihE

I think we should take a lesson from the Sumerians and hit the reset button. To echo what LFW said:

  • 2M in fees is nothing in the long run. We can easily recoup that if we get CDPs like 3088 to come back and create vaults in the future.
  • the negative optics/press. Consider reading this headline on coindesk as someone who knows very little about DAI “Maker implements 3000% tax on unpaid positions!!” Would you want to use this system? I wouldn’t. We should avoid this headline from happening and call GS.

Great point. We don’t want people to mint SAI, migrate it, and then deposit into DSR to get a rate arbitrage if they know we aren’t doing a tax. Looking forward to hearing more about what we can do about this. Maybe do GS sooner to avoid much undesired arbitrage.

2 Likes

Guys we have a serious problem if “No” wins as of right now. If we provide a clear signal to the crypto community that we are not going to impose a penalty then we have just provided a free money incentive to max out the SCD debt ceiling.

We need to re-run this decision with the SCD DC set to 0, so that the outcome can’t be gamed. If “No” wins then we keep the DC at 0. If “Yes” wins then we can raise the DC back.

I think we need to scrap this poll and proceed with a plan to redo it with the DC set to 0 first.

4 Likes

I’m not sure that we need to scrap the poll, given that there is no end date and the outcome is not certain, but I agree that we should set the SCD debt ceiling to zero as soon as possible and re-raise it if ‘Yes’ wins.

Remember this is just a signal between a few community members. This would still have to pass an on-chain poll to be approved, but yes, we definitely don’t want to signal to the community that we may forgive outstanding fees.

Due to the possible implications of this poll, some community members have determined that it may be unsafe to signal the results of this poll at this time. With that in mind, we have decided to postpone the poll until we can decide on an optimal game theoretic next step.

3 Likes

Do I understand correctly that tax simply transfers ETH from CDP with low collateralisation (little PETH relative to debt) to CDP with high collateralization? So It incentivies to increase your collateralization as much as You can, so or CDP owners will push ETH they have into CDP or will try to decrease outstanding DAI (pay debt)?

You can have infinite collateralization ratio (no debt), right? Does this mean that every “informed” ETH/WETH holder would be incentivized to “deposit” to PETH when the taxing period starts?
Meaning that largest CDP owners can’t really guess in advance what percentage of tax goes to them?

No debt PETH owners would also needed to be careful, not being too greedy and convert back to WETH just before global settlement is triggered.

1 Like

As it stands right now, we have been having some productive conversations in rocket chat.

Yesterday, @hexonaut came up with the solution of keeping track of the stability fee starting today and making this the tax at global settlement. With this solution, no CDPs would be punished for opening before today while at the same time ensuring that CDPs are still penalized/charged for not closing before GS.

Before coming up with this idea, the main arguments against using tax were that it would be punishing CDPs equally when some have much lower SF % than others. CDP users that haven’t been paying attention to governance and trust the system to have their best interests would not be punished for it. In effect this would be forgiving previous stability fees for users that still choose to wait until global settlement.

From here we would have to determine from what date makes the most sense to start calculating from. I said starting today above as an example but this could be anywhere from last year to once communications go out or on-chain polling passes. We should seek consensus on this parameter. With that being said we also need to determine if this is a popular option to begin with. Perhaps a majority of the community/MKR tokens wants to still tax everyone equally to the max, or maybe they want to just forgive all debts and get settlement over with. We’ll have to seek consensus on this as well.

As of right now, here are the shutdown options we have as I see (thanks @swakya) :

  1. Lower DC, shutdown in 6 months, forgive all fee debt
    -> nothing to add
  2. Swap Fee for Tax, shutdown in 6 months, forgive debt from before migration (or starting as soon as agreed upon)
    -> Maker forgoes past accrued SF but this reestablishes correct incentives for CDP holders
  3. Shutdown in 6 months, flash tax at the end, no debt forgiveness
    -> not the same as #2 since flash tax would not be relative to CDP age at all
  4. Lower DC, keep managing the system, close when one of the parameters is hit (6 months, 10M supply, .05 peg deviation)
    -> No debt forgiveness and no trying to recuperate fees through tax
  5. Lower DC, manage system for 6 months, stop supporting after 6 months, let collapse on its own, determine GS sometime in the future.
    -> Gives CDPs time to get affairs in order then they’re on their own.

**Edited 2. & 3. per @hexonaut 's post.

Hopefully we can have some meaningful discussion on these options and then we can open a poll with the options that make sense and can come to some form of consensus. Let me know your thoughts. Thanks.

4 Likes

One point to add is that options 2 and 3 do not require lowering the DC to 0 which is good for peg stability. In fact any option with a tax does not allow abuse and any option without a tax does allow abuse (which is why the DC needs to be 0).