[Signal Request] Adjust the Surplus Buffer

Overview

The Surplus Buffer (SB) is currently set to 4 MM, which is only ~0.33% of the DAI in circulation. In the last signal about the SB, there was a rough consensus that we should first flap some MKR. As of now we burned almost 2k of MKR already bringing is under 1MM MKR supply again :rocket:

The SB should protect the system against bad debt, my feeling is that having just ~0.33% of coverage is not high enough.

There are some interesting conversations around automatically increasing the SB, but I think we should not wait for them to resolve in an implementation.

Please also note that increasing the SB is also revertable - if we for some reason decide that the SB needs to get lowered, we can do that the same way as we do the increases.

Pros:

  • likelihood of flopping new MKR in a “too much bad-debt”-event gets decreased
  • maybe also allows us to experiment more with lower liquidation-ratio vault-types like WBTC-B backed by B.Protocol

Cons:

  • DAI in SB is DAI taken out of circulation
  • no more flapping as long as the SB is not filled if we combine this with the lerp we can stretch the filling of the SB over a longer period so a fraction of the surplus goes into the buffer while the other fraction goes into flapping

Currently we generate ~19 MM fees per year. Assuming we stick to this

Adjust the Surplus Buffer to

    • 4 MM (no change)
    • 6 MM (+ 2MM / + 50% / ~5 weeks without flap)
    • 8 MM (+ 4MM / + 100% / ~ 11 weeks without flap)
    • 10 MM (+ 6 MM / +150% / ~ 16 weeks without flap)
    • Abstain

0 voters

shall we lerp it and if yes how much should should get burned?

Lerping: increase the SB over time so we can burn MKR and increase SB simultaneously.

Upside of lerping: a fraction of the fees can still get used for MKR burning.

Downside: it takes longer to fill the SB

  • Burn 75% (4x the time to fill the SB)
  • Burn 50% (2x the time to fill the SB)
  • Burn 25% (1.5x the time to fill the SB)
  • Burn 0% (as above)
  • Abstain

0 voters

Next Steps

Poll will run until January 14th and depending on the result will move on-chain assuming the outcome the poll deems it necessary.

9 Likes

as the peg is looking a lot better now, i am not too concerned about this Con anymore

Very good signal request.

I think we all want to scale up and increase fees generation. While we don’t know exactly what the appropriate size of the Surplus Buffer should be, most will agree that 10M is not too much.

Let’s have room to experiment and take fees yielding risks.

1 Like

Hi,
They were a discussion to increase the param progressively, 2 weeks ago. Using the deployed contract from psm. The idea was to burn some too while increasing the surplus buffer.

It was MIPxx from @mrabino1

5 Likes

Yes we can use the lerp module to set the rate of increase over time so as to not shut off the burner for weeks at a time. For example, we could set the target increase length to be twice the amount of time it takes to burn that amount of Dai to reduce the burner to 50% instead of 0%.

11 Likes

that’s really cool, was not aware that we can use lerp for this usecase as soon as PSM hits prod.

so in case 10 MM wins we can configure the lerping to increase it to 10 MM over 32 weeks, so half of the surplus gets burned and the other half will go into the increased SB.

great collateral effect @hexonaut :smiley:

3 Likes

There is only one hic about raising it all in one go with the lerp,
We can’t stop it.

12M sound the way to go, but I think it is better to do it in two time.

2 Likes

I think we should also note that this is a negligible amount of Dai relative to the overall supply. I don’t think about this very much when considering the buffer.

1 Like

This is by far the best Signal Request you have created :stuck_out_tongue_winking_eye:

I polled for 10M and honestly, if you had provided the option for 12M and 22 weeks I would have opted in for that as well. And IMO I am okay with reducing the burner gradually over weeks, or immediately to 0%. One thing I learned in this ecosystem is that you can have :poop:tokens that go to a $3B Mkt Cap with Zero Earnings within a few minutes–courtesy of Degens diving-in, and you can have solid protocols (Maker) trade sideways for weeks, earning real profits, generating a Surplus. I’m in for the latter. Hence, the more the SB the better IMO.

4 Likes

If we can’t get a solution to automatically raise the Surplus Buffer, I’d prefer a signal request/governance poll with a commitment to doing smaller Surplus Buffer increases in the weekly executive.

Hm… Raising by 2MM would mean we would “need” to adjust once a month. Isn’t that small enough?

I voted for “No change”.

My reason is that a buffer raise of 4,6,8 or 10 million Dai is all based on gut feeling anyway. I think we should pause and then discuss a reasonably scientific framework for calculating the size of the buffer. And as suggested by @hexonaut - just use parts of the funds that go to the burner.

Then we can vote on the parameters of that framework.

6 Likes

You can actually stop the lerp module with another vote to de-auth it from the targeted module.

2 Likes

Let’s all read Basel III EU implementation during Christmas and have a scientific discussion in January :slight_smile:

I don’t know the exact number but it will be higher than 10M for sure. And we all expect to increase our exposure next year, not decrease it right?

We currently generate $20M of revenues per year for a $1B DAI exposure. I want to pause growth unnecessarily (like waiting for capital generation because we were too greedy) until $1B annual revenues.

I understand you want to increase MKR price but burning shares is the solution when you have nothing better to do.

MKR price = (revenues - expenses) * PER / #MKR
(with PER = price earning ratio (market dependant))

Expenses will come in 2021 so increasing revenues is mathematically a better solution than decreasing the number of MKR shares thanks to operational leverage. And to increase revenues we need to increase the surplus buffer.

5 Likes

Thinking, there is a little difference with the EU.

By increasing the surplus buffer we basically increase our USDC exposure as the dai will all go there.

To be fair, I think there is little benefit to increase the surplus buffer now. When we will have purge the USDC vault, there is huge benefits but now it is like an accounting some zero.

Just changed my vote. :upside_down_face:

3 Likes

If you think the loss will come from USDC, that’s good.

In September, @Primoz showed that with a base case scenario the 350M ETH-A might generate an average loss of 14M DAI (23.5M for a VaR 99.9%). Probably, this isn’t perfect, but the parameters make sense. Now the problem is obviously just bigger (even before adding ETH-B, YFI and a tail risk).

No one wants to take away the punch bowl just as the party gets going. But you can feel how that will end if the bull market stop.

Removing USDC from the PSM vault will be easy (we are all working on that with RWA and farming collaterals). Generating surplus not so, that takes time (I take for granted that we don’t want to issue MKR).

3 Likes

We can extract a certain percentage from our monthly income to increase SB so that we can continue to increase SB without suspending FLAP. I think the growth of SB must be continuous, but you can’t rush to realize it in one step. It is better and more stable to keep increasing continuously by a certain percentage every month.

3 Likes

I believe this surplus buffer discussion needs to be strategically planned in conjunction with @Hugh_Karp’s proposal for offloading risk from purchasing cover with Nexus Mutual. If the economic and financial analysis reveals that Maker is better offloading a portion of the risk than parking tens of millions of dollars (eventually hundreds of millions hopefully!), then we should plan total target, what portion Maker takes on, and finally, what portion is offloaded to Nexus.

In any case, raising the SB now can be undone at any point of time. Imho there is no need to delay any action right now even if there are other solutions in the midterm futures

2 Likes

Would be nice to have a combined proposal with the lerp. Currently the poll options don’t sound like a mix of burning and saving to the SB.