[Signal Request] Increasing the Surplus Buffer (2021-11)

I agree there should be some dynamic target. Especially if we also burn some proportion along the way.

8 Likes

Black thursday gets brought up a lot in this discussion - i just want to mention that we should not expect to see an auction keeper failure/dai liquidity failure combo again, meaning losses wouldn’t be anywhere nearly as high in a future crash that resembled the covid crash/black thursday situation

6 Likes

The discussion around the size of the Surplus Buffer is dissolving into a mess that involves multiple issues that could all be their own discussions.

  1. There is an unspoken feeling (maybe we could vote on it) that the current design of the buy & burn mechanism is not optimal. When something is copied by no other team in crypto you might want to reconsider. Instead of trying to increase the Surplus Buffer we could just vote to disable the Burner until a replacement is found if the community has lost faith in it as a method of returning value to MKR holders.

  2. Many forum members appear to confuse increasing the Surplus Buffer with growing the business. It is not. Maker has a vision and plenty of funds, but we lack both a strategy for how to get there and any form of long-term planning. In any normal organization, you would have investment proposals.

  3. I have previously proposed that Risk CU should come up with a formula for the Surplus Buffer and then we stick to that to avoid a steady stream of signal requests on the subject. I now realize this was a mistake since Risk will always work to increase buffers and so cannot make what is ultimately a political decision. What part of surplus funds goes where will have to be decided by a MIP.

7 Likes

The burn method is used by some of the largest companies in the world, including Apple. Crypto may not use it because the investors in this field are uneducated and drawn to 20000% APRs which are really just ‘stock splits’ they call stake rewards. I’d rather Maker burn supply because it the design is proven and sound.

8 Likes

Holds true only if MakerDAO always keeps increasing debt from volatile collateral. Regarding the formula, I do agree it would be easiest if each collateral type had additional governance parameter that equals risk premium and the Surplus Buffer would simply be a function of aggregate debt weighted risk premium. But instead of having Surplus Buffer signals, you’ll have signals and governance decisions for risk premium (because it can change) and parameter updates will need to be made one way or another. Maybe just less frequent.

4 Likes

I think this deserves it’s own discussion because I think the burn mechanism is an integral part of Makers value proposition, and I’m certain I’m not the only one judging by numerous Reddit threads. I do agree however that it deserves to be enshrined in an actual policy and not being left up to signal requests that pop up at random intervals. I happen to agree with SCV that the burn method is a bread and butter method from the financial world in returning value to investors. I think the biggest difference between MKR and other crypto assets is that unlike the majority, MKR is actually turning a profit.

I fully recognize that risk needs to be managed but I think it is also wise to take a step back and look at the bigger issue. Why do we need to manage risk? We need to manage it to maintain credibility. Users of DAI needs to trust that the value of 1 DAI is more or less equal to $1. In the same vein holders of MKR has to trust that MKR will not lose it’s revenues through mismanagement and lost collateral.

What I feel is underappreciated though is the loss of credibility that comes from not delivering on previous promises. Every time you do that you devalue your own credibility. If MKR was a company that in Q1 said that “We’re going to do a share buyback in Q4” they would tank their stockprice if they in Q4 instead chose to kept the profit in the company as liquidity reserves. This means that there are other risks to manage than systemic or financial risks.

There are good reasons to have and to grow the Surpluss buffer, but I think it has to be done as a balance with regards to burning some of the profits. It is not uncommon for an organization to have competing goals for its resources and then one need to come up with a meaningful compromise.

If anything I think this threat has shown the urgent need to establish a policy on distribution of profits, the level of risk and the size of the surplus buffer and how and when parameters regarding this is going to change. Predictability is a key word in this.

4 Likes

Imo, at this point, (partial) burning makes sense, if only, because no burning, divides MKR holders. We can dismiss it as irrational, shortsighted, dogmatic etc…, but many (smaller?) holders see it as the fundamental MKR value proposition.

Personally, I would go with something like 15%-25% burn fraction and 90-130mm surplus buffer since it would be likely non-contentiously passed → imo, starting to increase the buffer asap, is the most important thing.

There is nothing preventing anyone submitting new MIP to solve this (periodical) problem systematically. I would imagine large holders have strong incentive to do it. I also didn’t get the feeling, that burning proponents would oppose it vehemently.

6 Likes

@Aes I am onboard with you on 95% of the points you raised in this thread. Mainly the need to increase the SB and significantly. My only difference, I think, is how we are currently calculating rwa. In my understanding we actually should be using the official BIS percentages for asset risk weights, rather than coming up with our own. In that case, our ratio is well below than what we’ve been reporting. Meaning we have much less protection than we think we have.

Maybe a first step is to maintain clear and transparent documentation on risk weights for each single asset in Maker. Then a second step is to come up with a regulatory capital calculation first (standardised approach), before going on doing our own economic capital calculation.

4 Likes

I have ask, if we are much less protected than we think, but we aren’t accruing fees fast enough to fill a buffer anyways, why did we ever reduce the fees we had back in May? The stability fees should be large enough to maintain a good buffer and burn some MKR. It seems to me we should be doubling and even tripling in some cases the fees we charge now. I think we should focus on slowly increasing fees until we see material pushback from the market. I think it is likely that borrowers aren’t too concerned about the difference between 2% and 6% stability fees, because this market is highly volatile and they think they can profit. However, we care a lot.

1 Like

We needed to encourage DAI minting (as demand is higher than supply, and it still is: the PSMs are filled). Without the PSMs we would have had a huge peg problem.

How do you encourage DAI minting? Low SFs…

2 Likes

I think that’s is the best way forward but the 25% burn / 75% surplus is a good in-between until we figure this out.

6 Likes

On Black Thursday we did not have Liquidation 2.0. If we had, the losses were much lower or even zero. As a long time MKR holder, I’m in favor of burning, as this was the original promise since the beginning.

Chance of having another Black Thursday is close to zero, because now, with the Dutch Auction one keeper can bid on multiple liquidation auctions with the same capital. Back then the keeper had to wait for the end of the auction to unlock their money they used for bidding. Therefore there was no option to participate in simultaneous auctions if they had money only for one bid.

3 Likes

This sentiment poll is flawed. Users were able to vote for more than one option. Many users voted multiple times for different SB increase amounts. If you strip out people that voted multiple times, it appears the community favors keeping the SB at 60mm.

4 Likes

That’s actually intentional. The idea is for people to vote for all the options they would support on-chain. Any options with higher than 50% support (excluding abstain votes) will move to an on-chain poll.

Please vote for everything that you support.

The idea of this is to promote compromise and consensus. Some people would prefer increasing it by a lot, but would accept increasing it by a little. Some people don’t want to increase it at all, but would accept increasing it by a little. Having multiple votes allows us to capture that signal.

Conceptually, it’s easier to consider it as a single yes/no decisions on each option. Any options that a majority votes ‘yes’ on proceed as options on-chain. ‘No Change’ is also always an option on-chain, except in the very limited circumstances where doing nothing is not a viable option.

9 Likes

I believe Maker needs to develop Policies that address how to manage the Capital Surplus (CS) and burning (among other things) programatically as opposed to debating/voting on incremental changes to both.

Some Key Observations :

(1) Maker has a leveraged balance sheet where
A. Liabilities are basically demand deposits pegged to the US $
B. Assets include loans over-collateralized by crypto assets CBL), other stable coins, and a small amount of RWA

(2) Maker needs to understand, measure and manage the risks in both its liabilities and assets.
Maker’s key balance sheet risks including the following.

(3) Key Liability Risks
A. Liquidity risk: Maker does not have sufficient liquid assets to enable DAI holders to convert back to US $ as desired, particularly when many DAI holders want out quickly

(4) Key Asset Risks
A. Credit risk: assets do not receive all their contractual cashflows due to default by counterparties
B. Crypto Currency/Liquidation Risk: The crypto collateral backing the CBL falls so quickly in price that the Maker Keepers cannot liquidate under collateralized vaults at sufficient prices to cover the CBLs.
C. Interest Rate Risk: market interest rates move against Maker’s interest rate sensitive assets (for liabilities), devaluing them.
D. Currency Risk: currency rates move against Maker’s currency sensitive assets (or liabilities), devaluing them

(5). Capital Surplus: this is a prudent reserve that helps Maker manage the Key Risks, including the ones listed above. A reasonable Capital Surplus Policy requires that Maker be able to understand and measure its risks!

5 Likes

So based on “no change” currently at 32%, it would not be an option to vote for on-chain?

In polling science, if you have multiple options that are relatively close to one another (and each of those votes increases the denominator), then the outcome would skew towards the clustered options. If you had 3 candidates running for an election – 2 democrats and 1 republican – and voters could pick multiple options, then the republican would appear to poll very low.

That’s what is happening here. Pretty much everyone who voted for SB to 90mm also voted for SB to 130mm. They are substantially similar. If you had a simple poll of raise SB or keep flat, the polls would tell a much different story.

3 Likes

I would like to encourage voting for the 1MM increase per week. The numbers in the original post are already outdated as DAI supply and revenue increased a lot over just one week.

Even with this option a lot of MKR will get burned.

2 Likes

No change will always be an option in the onchain poll.

The additional votes do not increase the denominator. Each option is measured as:

votes_for_option / (total_voters - voters_for_abstain)

Note total_voters rather than total_votes.

Any of these that resolve to >50% will be an option on-chain. ‘No change’ will almost always be an option on-chain, and will be in this case.

This is not what’s happening. Give me (and discourse) some credit, the percentages shown are votes / total_voters. This is mostly correct, though it doesn’t account for abstain, so we account for that manually, this means the actual percentages for each option are slightly higher than displayed.

Given this is a moderately complex poll, and is high-inpact, I’ll clearly present the results and the methodology when the poll is finished and will confirm what the on-chain polls will look like.

2 Likes

Ok, that’s helpful. Thanks LFW.

2 Likes