Request for Adjustment to DSR Phrasing

UPDATE: New poll and discussion thread are here: Signal Request: Should we replace the weekly DSR governance poll with a DSR Spread governance poll?

I’d like to request a change to the way the DSR governance polling question is phrased. Currently we are being asked for the absolute values of the SCD SF, MCD SF and DSR Rate. I think it is more appropriate to ask for the MCD SF - DSR Spread to replace the DSR Rate poll. With this phrasing change it helps keep the two values as orthogonal as possible. DSR Spread adjusts for system growth/migration speed whereas MCD SF adjusts for DAI peg stability.

This will also help us avoid a situation where for whatever reason the DSR mistakenly ends up being voted higher than the SF. I believe this issue may come up once we start having to change the SF for peg stability as there is a problem of coordinating the values of the DSR and MCD SF to move in tandem. It’s not clear whether voters are signalling to increase/decrease the DSR spread or if they want to adjust for peg stability.

IMO keeping the voting parameters as orthogonal as possible is essential for demonstrating voter intentions in the clearest way.

EDIT Adding a vote as requested.

Question: Should we switch the phrasing of the DSR governance poll from absolute DSR % to a MCD - DSR spread %?

  • Yes
  • No

0 voters


Maybe set up a poll to see if this is a popular change? I don’t have a problem with it but it usually helps to get traction if they see that more people support it.

I don’t see a problem with the DSR temporarily being higher than the weighted average SF. At the time of writing, only 20% of DAI is in the DSR so we have a lot of slack that we can use to our advantage.


I’m not super familiar with the forums. How do i do that?

The problem comes with the fact that you’ve given vault owners a way to earn a guaranteed % return on their ETH stack. It may max out the debt ceiling pretty quickly.




For examples of setting up signalling posts, there are lots of examples under the signalling tag. There is also a post suggesting a process: Current Signaling Process

1 Like

Agree with hexonaut, I was actually going to make a thread about this. When I saw this I was happy I didn’t have to do it myself.

“It’s not clear whether voters are signalling to increase/decrease the DSR spread or if they want to adjust for peg stability.”

This is spot on. It needs to be made clear.

There are other reasons as well. Once Vault users begin hedging DSR they need someone who has the direct inverse exposure to them. If the spread is adjusted too much it makes it more difficult for them to secure that hedge.

Thanks for this. Updated!


EDIT: i’m wrong in my thinking, I misunderstood hexonaut (see comment)

What are the options for a poll like that?

What should the spread be between DSR and MCDSF?

  • 0% at 2/2
  • 0% at 3/3
  • 0% at 3.5/3.5
  • 0% at 3.75/3.75
  • 0% at 4/4
  • 0% at 4.25/4.25
  • 0% at 4.5/4.5
  • 0% at 5/5
  • 0% at 6/6
  • 0% at 8/8
    (then multiply that by 1% at, 2% at, 4% at, etc…)

You have to define the spread = {,} and the rate combination = {,} so the number of exact choices becomes a lot more unmanageable.

What you’re describing sounds like a separate poll to try to keep people consistent with their actual votes. People will express their opinion with exact values for the rates anyway.

It would be cool to vote on a matrix though if there was a user friendly way to do it.

1 Like

So it would be a range of values with a lower bound of 0%. It doesn’t matter what the MCDSF is. You want the DSR to follow it minus this parameter.

Since the DSR affects demand and SF affects supply in opposite directions it does not make sense to have these as separate parameters for controlling peg stability. Different parameters should affect different pieces of the system to keep things orthogonal in our collective brain.

An example:

DAI trading under $1.

Current Options:

Raise DSR or Raise SF

Fixed Spread DSR Options:

Raise SF


This is confusing to me. What would the new poll actually look like like David tried to define? Can you provide a concrete example versus the abstract one you provided?

Yeah so it could be something along the lines of.

What should the Stability Fee - DSR spread be?

0% DSR = SF (for example a SF of 4% means the DSR is 4%)
0.5% DSR = SF - 0.5% (for example a SF of 4% means the DSR is 3.5%)
4% (for example an SF of 4% means the DSR is 0%)

Some more examples for clarity:

DSR spread of 1% and SF is 8% means the DSR is set to 7%.

DSR spread of 4% and SF of 10% means the DSR is 6%.

Does this clear things up?

1 Like

That does, thank you. I think I still personally prefer the current implementation just because I’m used to it and it feels more binary to me (hopefully thats the right word choice to describe it).

I think my point still stands. In the end the actual rates need to be decided on, not just the spread. So your options will be the multiple of the two sets:

the spread = {,} and the rate combination = {,}

Could someone make a list of options that would work well for a poll like this?

Maybe I’m misunderstanding. I am not suggesting getting rid of the MCD sf vote. That would stay as is. I am talking about replacing just the current DSR rate vote with a question about spread. Both the current and proposed setup provide the same end numbers. It’s just phrased as a spread and not an absolute number.

Ah, I see what you’re saying.

So the spread vote + MCD SF vote would give a DSR value

got it. I understand now.

1 Like

I voted No, but only because I would like to try using the IAM to tune the system instead of weekly votes.

Can you elaborate on why the introduction of the IAM changes things?

The IAM will only need to be voted on when the bounds are set. Let’s say we vote to set the IAM DSR bounds from 1-8, and the ETH-A SF bounds from 1-16. The risk team could tune the SF and DSR at any time without having to worry about a vote pushing the DSR much higher than the SF.

If we do keep voting each week, then I’ll change my vote on this signal request to yes for this reason you put above.

This makes sense and I 100% agree, I just don’t think it will be necessary after the IAM are implemented.

So my main reason for wanting this change is not because of the accidental DSR > SF issue. It’s because I believe it is a more natural fit to how we should be thinking about these parameters. As I said above we currently have two parameters to effect dai peg stability when it should be reduced to one for simplicity and clear signaling.

With regards to the IAM, I think only the stability fee should be set each week with the DSR rate following it minus the spread parameter. The DSR spread should remain relatively constant because it effects growth/profit and not peg stability.

Under this change and after migration I think we can simplify things down to 1 weekly poll of the global stability fee only. If we keep the DSR in terms of absolute rate we will almost certainly have to vote on 2 things week to week (the DSR and the global sf).


OK, I’ve been pondering this for a few days. Why is it necessarily a problem to provide a guaranteed return? What happens when we add ETH vaults with more limited leverage and lower SF? If the DSR is rigidly limited by the lowest SF for any asset type then we’re artificially limiting our ability to govern effectively.

If I take my ETH, open a vault, generate DAI, and deposit it in the DSR then I might extract 0.5% or 1% risk free, but I’m also tieing up capital. It’s not necessarily a great investment. Also, if DAI is trading below 1.0 then I think it could be worth temporarily raising the DSR above the weighted average SF to defend the peg. What we don’t want to do ever again is to raise the SF to 20% to defend the peg. Vault owners just loathed that experience, adding injury to an already bitter bear market.