[Signal Request] Change the Stability Fee Structure

Discussion: Change the Stability Fee Structure

Based on the discussion over the past week it seems there were no strong objections to proceeding with this initiative. The main points of contention were around naming and the negative base rate. I’ve decided to run a 4 polls to cover these points of contention as well as a poll to cover something which wasn’t really discussed - the polling cadence of the vault portion of the fee.

The General Idea

The first poll is on whether you in general like the idea of splitting the stability fees into two separate components - one global fee portion (the Base Rate) and one per-vault fee portion (Name to be Voted).

  • Yes, I want to proceed with the change to split the stability fees into two components.
  • No, I like things the way they are.
  • No, I have a different way to proceed.
  • Abstain (I disagree with the poll options)
  • Abstain (I have a different objection)
  • Abstain (I want to see the results)
  • Abstain (I have no opinion)

0 voters

Naming

I saw no objection to the term Base Rate to refer to the global portion of the fee, so I’ll assume that is okay to proceed with. There have been three terms floated for the vault portion of the fee. I will summarize as best I can.

Risk Premium

Argument For: The vault-specific portion of the fee will mostly if not completely consist of the risk premium. Historically this is the term most people seem to use and understand.

Vault Fee

Argument For: The Risk Premium is not the only part of the vault-specific portion, and so we should generalize the term to include future pieces such as collateral-specific taxes/subsidies.

Vault Spread

Argument For: Similar argument to Vault Fee, but use the term spread as it may be less confusing to the vault owners. With Base Rate and Vault Spread as the main terms it is still clear that Stability Fee is what they will be paying and it won’t be confused with the vault-specific portion. There is also a symmetry surrounding the name Spread as the DSR Spread and Vault Spread both extend from the Base Rate in opposite directions.

Pick your preference:

  • Risk Premium
  • Vault Fee
  • Vault Spread
  • Abstain (I disagree with the poll options)
  • Abstain (I have a different objection)
  • Abstain (I want to see the results)
  • Abstain (I have no opinion)

0 voters

Negative Base Rate

Under this new system we have:

Stability Fee = max(Base Rate + Vault-specific Fee, 0%)

A negative Base Rate allows us to achieve a target Stability Fee of 0%. Without the negative Base Rate the lowest Stability Fee we can achieve is whatever the vault-specific fee of the collateral is. Keep in mind this is a governance construct and does not correspond to any system parameters being negative.

For example, let’s say risk comes back and tells us that the Risk Premiums are as follows:

ETH-A = 2%
BAT-A = 2%
USDC-A = 10%
WBTC-A = 12%

to encode the current setup of ETH-A Stability Fee = 0%, BAT-A SF = 0%, USDC-A SF = 0%, WBTC-A SF = 1% we can do this:

Base Rate = -11% (MKR is subsidizing the risk premium in exchange for peg stability)

ETH-A SF = max(2% + (-11%), 0%) = max(-9%, 0%) = 0%
BAT-A SF = max(2% + (-11%), 0%) = max(-9%, 0%) = 0%
USDC-A SF = max(10% + (-11%), 0%) = max(-1%, 0%) = 0%
WBTC-A SF = max(12% + (-11%), 0%) = max(1%, 0%) = 1%

I will attempt to summarize the pros/cons:

Pros

  • Allows more flexibility in what we can vote on in the weekly poll. Without a negative Base Rate reaching a Stability Fee of 0% requires initiating one-off votes for each of the vault-specific fees.
  • Makes it explicit when we are subsidizing risk with MKR. Base Rate > 0% means the system should be bringing in revenue. Base Rate < 0% means the system should be taking on debt. In an ideal world the Base Rate should always be >= 0%, but it’s good to have this option available.

Cons

  • Too complex. It’s simpler for voters and users to not have to deal with a negative rate.
  • MKR subsidies for risk should be a rare occurrence. We have the tools available with the vault-specific portion votes to set Stability Fees to 0% so why complicate this more than necessary when it will not be used that often.

Pick your preference on allowing a negative Base Rate:

  • Yes, allow a negative Base Rate.
  • Disallow a negative Base Rate. 0% Base Rate will be the minimum.
  • Abstain (I disagree with the poll options)
  • Abstain (I have a different objection)
  • Abstain (I want to see the results)
  • Abstain (I have no opinion)

0 voters

Vault-specific Polling Cadence

One of the reasons to make this fee split is to reduce the number of weekly governance fee votes from 1 per vault type to some smaller number for scaling purposes. The two options I see are having the vault-specific portions of the fee voted on at either an infrequent, but regular voting period or on an as-needed basis. An example of the as-needed basis is if the risk team shows that the risk premium has changed for ETH-A and it needs to be adjusted.

  • We should vote on vault-specific fees at an infrequent, but regular interval.
  • We should vote on vault-specific fees on an as-needed basis.
  • Abstain (I disagree with the poll options)
  • Abstain (I have a different objection)
  • Abstain (I want to see the results)
  • Abstain (I have no opinion)

0 voters

Polls will run until Thursday May 14th at 11:00am EST to be discussed on the weekly governance call. Possibility for extension if there is not a clear answer by then.

5 Likes

Any chance I can get people to reconsider the use of ‘Risk Premium’ as terminology for the Vault Spread?

We already use Risk Premium to refer to the portion of the fee which balances the risk taken by the Protocol, this is distinct from any monetary policy adjustments to that number which combine to form the Vault Spread.

It’s undesirable to name the combination ‘Risk Premium’ because it implies that we are correctly balancing risk for these Vault types whereas in actuality we would not be.

I’d literally prefer we called it something nonsensical like ‘Jellybean Premium’ than ‘Risk Premium.’

5 Likes

I always considered risk premium just a general term, that is just part of the financial vocabulary.
Also, protocol is not really formally specified and versioned. I mean, i am the guy who is bothered by token being called Dai, instead of Dai 1.0., but i just don’t have the energy to fight for clarity, backwards compatibility, transparency etc… I don’t think majority of holders really care.

I think I may be wrong on my definition anyway, had a quick look here: https://www.investopedia.com/terms/r/riskpremium.asp

I’d still rather we didn’t overload the term though. If we call the combined rate ‘Risk Premium’ what do we call the two components?

Risk Premium = Monetary Policy Modifier + ???

And then we have:

Stability Fee (should be Vault Fee) = Base Rate + Risk Premium?

Edit: Nevermind, risk free rate is something different :frowning:

Agree with LFW. We should be thinking forward on this terminology. Stability Fee has already been grandfathered in as the term for the total fee, but really only a portion of it is used for “stability”. Risk Premium may cover exactly what the vault portion of the fee is right now, but we are almost certainly going to be adding a monetary policy component later on.

Also, I would like to hear some arguments for why people do not want the negative base rate? I feel like we are shooting ourselves in the foot by not including this. In the current market it’s clear MKR holders want to subsidize the risk of ETH to defend the peg. We will lose this monetary tool if we only allow for a positive base rate. I understand we can initiate one-off votes for each collateral type, but why do this when we can encode this goal into an existing weekly tool?

2 Likes

I don’t think negative base rate makes any sense and have no problem if we are off the peg for a longer periods of time. I am not 100p against it, but such things should be taboo.

But yeah, we are in bad situation since we are pegged to something fundamentally broken (current financial system does not allow for asset depreciation, and borrowers are perpetually bailed out).

1 Like

Could you elaborate on what aspect doesn’t make sense?

I edited to explain the negative Base Rate more in the hopes that maybe the no votes are coming from not properly understanding what it means.

Under this new system we have:

Stability Fee = max(Base Rate + Vault-specific Fee, 0%)

A negative Base Rate allows us to achieve a target Stability Fee of 0%. Without the negative Base Rate the lowest Stability Fee we can achieve is whatever the vault-specific fee of the collateral is. Keep in mind this is a governance construct and does not correspond to any system parameters being negative.

For example, let’s say risk comes back and tells us that the Risk Premiums are as follows:

ETH-A = 2%
BAT-A = 2%
USDC-A = 10%
WBTC-A = 12%

to encode the current setup of ETH-A Stability Fee = 0%, BAT-A SF = 0%, USDC-A SF = 0%, WBTC-A SF = 1% we can do this:

Base Rate = -11% (MKR is subsidizing the risk premium in exchange for peg stability)

ETH-A SF = max(2% + (-11%), 0%) = max(-9%, 0%) = 0%
BAT-A SF = max(2% + (-11%), 0%) = max(-9%, 0%) = 0%
USDC-A SF = max(10% + (-11%), 0%) = max(-1%, 0%) = 0%
WBTC-A SF = max(12% + (-11%), 0%) = max(1%, 0%) = 1%

4 Likes

We are in the world of crypto assets. As long as the collateral comes from crypto assets, the effect of low interest rate and negative interest rate is basically the same. Negative interest rates will only work if future collateral is added to real world assets.

Suppose we have a USDC-B vault type with a high stability fee (30% or whatever) for emergency liquidity. In your new scheme, can we keep the stability fee at 30% regardless of the base rate?

i think in the way that this is currently proposed it would move in lock step with the base rate, so for instance you could imagine right now we have a base rate of 0%. Well if we wanted to add this 30% SF USDC-B vault type i think it would end up with a “risk premium” of 30%. Suppose in 12 mo ETH starts on a bull run and we need to raise the base rate to say 3% well once that happened USDC-B would have a SF of 33%.

IF we wanted to keep it at 30% i think that would be doable but you would need an executive vote to reduce the “risk premium” on USDC-B to 27%. That said, IMO i don’t think this would be necessary given that in this hypothetical we would be looking at a more bullish crypto-economic cycle, but i think that is how it would work.

2 Likes

The issue with a negative base rate is it opens up the possibility of a negative total rate. Which maybe could work, but I feel like we haven’t properly discussed the ramifications of a negative total rate, especially for long periods of time.

I don’t care so much about the terminology, but risk premium seemed like the most sensible thing to call it. Maybe collateral premium? Token premium? Vault spread doesn’t make any sense cos the rate is per form of collateral, not per vault.

Alternatively we could call the total rate just “total rate” instead of risk premium, but that could be confusing with older use of defunct terminology.

I also like Andy mcCall’s idea for a multiplicative base rate instead of an additive one, which would then allow us to move all collateral types to a total rate of 0% without using negative base rates.

3 Likes

The way I defined it is so that negative total rates are not possible.

The rate is actually per vault. For example ETH-A and ETH-B could have different rates.

2 Likes

A multiplicative base rate is interesting, but a special case would be needed for a USDC-B vault with high stability fee for emergency liquidity.

1 Like

Why do we need two udsc vaults? Wouldn’t that be redundant?

It has been a popular idea thrown around lately that we need a second USDC vault type with a high stability fee so it could be used to generate DAI mostly in the context of a flop auction. Would go into it in more depth but it is probably off topic for this forum thread. Maybe start a separate post for your question here?

1 Like

FWIW i like this one

1 Like

I was originally liking a multiplicative base rate (ax+b). But realised that you then need to have two global parameters (a and b) plus a vault specific parameter (x). I think this adds too much complexity. I think having an additive system (b + a) strikes a better balance of utility and simplicity.

I’m personally in favour of the negative base rate. Without it, we are saying that every collateral with a stability fee currently at 0% should start with their collateral premium at 0%, which I don’t think many people would agree is correct.

1 Like

Yeah, it’s a bit too complex to poll for. And b+a should be fine as long as the vault spread gets periodically adjusted.