[Signal Request] Adjust the WBTC Debt Ceiling and Risk Premium

The WBTC DAI minted has been effectively capped by the debt ceiling for a week. WBTC is the most popular collateral after ETH with 8.08% of DAI minted. There has been strong demand for minting DAI from WBTC, maybe even stronger demand than from ETH. This is important to help stabilize the peg, and add DAI liquidity in this 0% rate environment. WBTC is in a unique position being the only active Bitcoin collateral on MCD.

There has been some discussion in the community regarding raising the WBTC debt ceiling. Members discussing the risk profile of WBTC, and also an increased risk premium that should come with a Debt Ceiling increase to balance the increased risk exposure.

WBTC Collateral Discussion - It should be noted that WBTC has some custodial and counterparty risk.

Increasing the WBTC Debt Ceiling

Increasing the WBTC Debt Ceiling would allow more DAI to be minted from WBTC. MKR Holders will be exposed to more risk. Particularly WBTC custodial risk and WBTC liquidation risk.

  • 10M - No Change
  • 15M
  • 20M
  • 25M
  • Abstain

0 voters

How to increase the WBTC Debt Ceiling

There has been discussion on how a WBTC Debt Ceiling increase should be implemented. We may want to keep WBTC Debt Ceiling room available for retail users. There are also some additional risks if the debt ceiling is being used by one vault.

The winning option will be included in the poll with the WBTC Debt Ceiling increase. And if the poll passes, the debt ceiling increases will be included in the weekly executives.

  • All at Once
  • 500k per week
  • 1M per week
  • 2M per week
  • 5M per week
  • Abstain

0 voters

Increase the WBTC Risk Premium

There has been support for increasing the WBTC Risk Premium with a WBTC Debt Ceiling increase. Though, there may be some mixed opinions on how much to increase the WBTC Risk Premium.

Some factors:

  • WBTC custodial Risk
  • Comparison with ETH-A Risk Premium
  • DAI liquidity
  • Rate shocks

To keep it simple, you should assume a 0% ETH Risk Premium for this poll. If that changes before next week, that’d be added to this poll result.

If the Debt Ceiling Increase gets spliced up, the Risk Premium increase will also get spliced up.

Note: Stability Fee = Base Rate + Risk Premium

  • 1% - No Change
  • 2%
  • 3%
  • 4%
  • 5%
  • Abstain
  • I think the risk team should decide

0 voters

Polls will run until Sunday June 21 at 7pm UTC to be included with the weekly governance polls. Possibility for extension if there is not a clear answer by then.

5 Likes

I like the idea of coupling a gradual debt ceiling increase with gradual risk premium increase.

I think someone on a governance call suggested something like raising 0.25% risk premium for each 1 million Dai of increased debt ceiling.

I wish this signal request made it easier to express that, or at least the first step of it (for example, 11 million debt ceiling and 1.25% risk premium).

4 Likes

I didn’t hear of this before. Seems like @cmooney mentioned it. It seems sensible to amend the signal request. If ‘All at Once’ doesn’t win, the Risk Premium increases would get spliced up.

If other people missed it too, check the June 4th transcripts for some discussions on WBTC. No discussion in the June 11th meeting.

1 Like

In my personal opinion I don’t see how raising the W-BTC ceiling makes sense. Simply because it has hit the cap does not justify increasing it (then what’s the point of a cap anyways?). Right now we’re still suffering from (slight) peg strength issues (see yesterday’s discussion in #governance-and-risk). No collateral asset is charging a risk premium right now. If we need to artificially incentivize Dai supply to fix the peg, then it should (IMO) be only from ETH, additionall trustless collateral, or USDC-A, as they are the safest options. W-BTC contains risk that we don’t fully understand. The only relevant question here is “how much debt ceiling exposure do we want towards centralized assets.”

Additionally, we don’t know if the W-BTC debt ceiling has hit the cap because of organic demand or because of massive rate arbitrage (likely the latter IMO). So once rates start to normalize (whatever that may be. If we want to discuss W-BTC or pure BTC risk premiums separately, let’s do that), and then we can see what organic W-BTC demand is, and then we can maybe try to balance risk vs adoption / user demand. Otherwise, we’re just incurring needless risk and printing Dai for no real benefit to the community.

7 Likes

We see that demand for ETH mortgages has stalled. Even if the interest rate is zero, the number of ETH is still between 1.9 million and 2 million.

Not sure if this has been mentioned, but this risk premium increase should be viewed in conjunction with other proposals to increase the base fee. If the base fee increases, this will effect the above votes. Or does it…?

For example. If the majority here believe wBTC should be at 2%, then with a 1% base fee, it will be.
Or are we actually saying, we believe wBTC should be at 2% +1% base fee?

This distinction must be made clear to give an accurate reflection of sentiment here, which is not yet the case.

2 Likes

[reposted from rocketchat]

Additionally, we don’t know if the W-BTC debt ceiling has hit the cap because of organic demand or because of massive rate arbitrage (likely the latter IMO).

@cyrus I don’t think there is as clear a difference between “organic demand” and “massive rate arbitrage” (or, I would invite you to define the difference, if you can). I would argue that DAI issuance against ETH is a similar “rate arbitrage”, and probably even more of an arbitrage since the ETH stability fee is 1% below that of WBTC while market rates for ETH have been consistently well above that of BTC since mid-late May (as evidenced by the futures curves, for example: all of the ETH curves have been in steeper contango than BTC by a few % ). So all in all I’m not sure why we would class the 0% ETH CDPs as “organic” while the WBTC ones “arbitrage”. This isn’t only addressing your point above but I also saw people making this distinction a few times in the past.

I of course do agree about setting the debt ceiling based on our view of the risk of WBTC (and in particular the custodian/redemption risk), to avoid concentration of risk in the collateral portfolio. I think that is what should almost always be the only reason for setting a debt ceiling.

n.b. I am not suggesting that we ignore the fact that 90%+ of the DAI issued against WBTC was issued from one CDP.

I voted for 10mm or 15mm debt ceiling, with adjustment made at 500k or 1mm per week, since I think we should be gradual with the increase and try to observe more data from future liquidations before building up exposure. I think a risk premium of 1 or 2% is fine since it is really only meant to represent a custodian risk spread (as BTC in general is more liquid and less volatile than ETH which is currently serving as our de facto “baseline” with it’s 0% rate).

7 Likes

@MakerMouse Yes, the risk premium is added to the base rate. And the vote is on the risk premium.

(And more specifically, the spread with the ETH-A risk premium)

I think raising the debt ceiling is the correct choice so long as the stability fee is also raised. There is risk with having WBTC and this risk should be offset by additional fees (clearly the demand is here currently to support a fee increase if we are already maxed at the debt ceiling). If we do not raise the ceiling I still think the risk premium should be raised.

1 Like

So I guess in a situation like black Thursday where a large amount of wbtc would be liquidated: since the secondary market will be much smaller than the supply of wbtc locked in maker, keepers will rely on liquidity through coinlist. So any keepers would have to have a coinlist account, unwrap the wbtc for btc, sell that for dai in order to recycle?

It seems fine to me to raise the ebtc debt ceiling as long as there’s some feasible process for keepers to rapidly liquidate wbtc during a crisis. Unlike other erc20 tokens, this liquidation process for wbtc will not likely be done on ethereum and so will fully rely on centralized exchanges, and mostly the single exchange coinlist at that.

There are centralization risks here that aren’t even custodial risk. Even just coinlist having temporary downtime due to high traffic could prevent the wbtc liquidation process from functioning properly.

2 Likes

We are dealing with non-important issues here. IMO, the priority should be to find the lasting solution to prevent DAI from rising much above $1 (now it is $1.013 at Kraken DAI/USD) which probably means abandoning partially or completely the current peg control system.

The DAI supply was around 135M about a week ago and now it is 116M. We can play with the wBTC parameters because we can but I don’t believe it has any effect (or the effect will last only a week or two).

I voted for no increase and highest risk but it really doesn’t matter because we are just buying time (or wasting time) if we set it to i.e. 20M/2%.

1 Like

@befitsandpiper https://www.curve.fi/ren is a new renBTC/WBTC bridge. $800k in reserves and growing. Bonding curve is more efficient than uniswap. It’s beta still but you can even swap trustlessly, directly with the BTC blockchain.

(Increasing trading volumes - https://beta.curve.fi/volumepercoin)

5 Likes

Great signal request @Jiecut. Just two comments.

Ending on Sunday evening UTC might not be enough lead time to get a poll the following day. After the forum stage, a poll template needs to be submitted to github, and needs to be reviewed by the facilitators (and ideally others, but that’s not setup yet) and deployed on-chain. It may be possible, and I haven’t spoken to @rich.brown about it yet, but in the future I think it would be better to leave a few days between.

Second comment is addressed at everyone generally. Going forward we want to try to emphasise reaching consensus as part of the signalling process. That means ideally we would like the leading options to reach majority (have greater than 50% support.) So I would encourage people to consider compromise options and keep in mind the mantra of voting for everything you would accept compared to the status-quo.

Not to name names, but as an example I’d note that there are people voting for a 25M ceiling, who aren’t voting for a 20M or 15M ceiling, despite these being an improvement over the current 10M. Remember you can vote for multiple options, and that you can change your votes!

5 Likes

Someone let me know if I’m wrong… but I thought the caps were in place to limit the damage to the system in case an exploit was found. Less cap head-space = less exploitable opportunity?

Agree that rate arbitrage isn’t helping restore the peg, just making the current peg price (high) more robust

Going to follow up on my first point. @Gov-Facilitators need at least 24 hours (not on a weekend) to review a PR submitted to the polls github. Given the timeline on this poll and the PR requirement (and assuming everything else goes well) it will go on-chain week beginning the 29th June.

Ideally we’d like way more than 24 hours notice. Hopefully we’ll be producing an on-chain poll deployment process in the near future that provides clear cut-offs and requirements for those wanting to get polls on-chain.

2 Likes

Thank you, I was aware of the requirement to make the pull request to the polls github. I was prepared to submit it right away, I knew it was tight but didn’t know about the 1 full business day requirement.

Also, the deferral of decisions to the risk team was also on previous polls. Though I haven’t contacted the Risk team if there’ll able to give guidance on this and how much lead time they’ll need before a governance poll gets published. So if that options won, it would probably have to be delayed also.

Depending on how much lead time is needed, it might be reasonable to delay ending the poll to Wednesday, June 24th, 2020.

A bit unfortunate that the poll could be ready for June 22nd deployment. Some discussion for 2 weeks but a signal request wasn’t made.

To be fair, we’ve never said there was a requirement before because we haven’t needed to. I think it makes sense though. Depending on the number of polls going in and the length and complexity it can be a little manic if it’s last minute.

My understanding is that @cyrus is against increasing the ceiling at this time. I’m not sure if he has views on the risk premium.

Extending the poll is fine if you believe there is reason to do so (chance for getting more consensus, etc.) Though obviously dragging it out if you dislike the result is a no-no.

Yep, the earlier signal requests get posted, the earlier changes can happen. Potentially if you’d set the poll end date sooner we could have had time to do it next week, but then I’d be complaining that people didn’t have enough time to vote :wink:

1 Like

I wanted to add that https://wbtc.cafe/ is now live. This reduces redemption risk for WBTC by using Curve’s renBTC/WBTC, straight to RenVM with redemption in real BTC. As of the time of this post, RenVM has 233.03 BTC under custody so one can’t yet swap the entire pool of WBTC under custody, but IMHO this redemption risk reduction warrants a reconsideration of the DC and stability fee.

9 Likes

I have to 100% agree with this line of reasoning.

1 Like

Perhaps the best way to fix the peg trustlessly is to introduce negative rates for ETH CDP’s, thus incentivizing more ETH deposits and dai generation

1 Like