Mirror governance incentives are broken

Currently, providing liquidity to any mAsset pair provides over 300% APR in the form of MIR liquidity rewards, as well as getting price exposure to some of the world’s most popular equities (AAPL, TSLA etc). In contrast, the average yield for MIR staking on the Governance module over the course of the past two weeks has been ~15%, while voting on proposals adds an illiquidity cost, as tokens used to vote on proposals can no longer be transferred for the duration of the vote.

Obviously the economically sane choice is to forgo gov staking for most users, which makes voting on proposals highly illiquid & subject to manipulation. The Mirror community needs to find a way to improve the incentives for participating in governance in order to secure the protocol.

Potential solutions:

  1. Use MIR-UST-LP tokens for governance staking rather than MIR staking: Here, the user can both enjoy the high APR from MIR LP rewards, while also getting some additional rewards from gov staking. This solution would likely have to be used in conjunction with 2 or 4, as a user already earning 300+% APR will be indifferent to the 15% being earned today
  2. Apportion additional incentives from the community pool to reward gov staking and/or voting: This makes gov staking more attractive, which offsets opportunity costs of illiquidity
  3. Allow votes to be canceled, and when such, the tokens to be withdrawn: Obvious, but makes gov voting much more vulnerable to manipulation (process is much more complicated, especially when market cap is this low)
  4. Shorten the voting period: 1 week? should do it?

Looking forward to feedback.

3 Likes

Personally i like option 2 by itself. But it would be cool if we setup a contract that can be used to fund gov rewards, and if gov could allocate a certain percentage of MIR apy to the gov tax pool. And with this approach if gov decided to it could allocate community funds and or adjust the interest allocation to it’s own tax pool to further incentivize gov participation if it were necessary.

Annyung DoKwon,

  1. Using the LP tokens makes sense. Would they be locked if used in a poll for example?
  2. Incentivizing gov staking and/or voting is a must to get the community involved. Over at $CVP governance voting no incentives were offered and high gas fees just to vote were a big turnoff.
  3. I think if you decide to vote you shouldn’t be able to cancel it.
  4. A shorter voting period of 1 week sounds more reasonable.

Komapsumnida!

I like option 4. At the very least, you can start with this. 7 days for voting-will increase the speed of adding new shares to whitelists!

1 Like

+1, option 4 is a given and a good starting point, but i do think long term gov should be able to manage gov rewards and adjust them to encourage gov participation when necessary (opt 2).

Our goal is to increase the incentives and lower the opportunity costs associated to governance participation

For that matter, I would oppose 2) as we should dedicate our funds to non-recurring costs focused on growing the ecosystem (grants, r&d). With the same approach I would not use these funds to incentivize liquidity on new mAsset pairs, but it’s a debate for later.

I like what Do mentions as option 1 (use MIR-UST LP for governance staking and discontinue MIR token staking) and agree that we need to make MIR-UST staking into governance more attractive.
Therefore I think we can also divert some of the MIR/UST pool rewards to governance staking:

From Github:
Mirror Protocol rewards stakers of LP tokens for each of the mAsset-UST pools as well as the MIR-UST pool. All mAsset-UST staking pools receive a weight of 1, while the MIR-UST pool receives a weight of 3. Therefore, there is a stronger reward incentive to stake MIR-UST LP tokens as they confer 3 times the reward when staked relative to mAssets.

If we could make the MIR-UST pool weight 1.5 (instead of 3) and divert the rewards corresponding to the extra 1.5 point into governance staking of MIR-UST LP tokens, then governance stakers could earn inflation rewards + CDP redemption fees for a much better yield than LPing MIR-UST alone.

The difference in yield would not be between 300% and 315% , but between 150% and 315% (simplified here), also the added benefit is that governance staking would yield rewards every block and not only during minting hours.

In summary
Yes to option 1 (coupled with a reallocation of MIR-UST LP rewards to Gov staking of these LP tokens)
No to option 2 (let’s look for better ROI options)
No to option 3 (too risky, and not needed if illiquidity cost is offset by sufficient extra yield)
Yes to option 4 - it’s underway with MIP 30 I believe anyway

4 Likes

Arguably as MIR starts to list on more CeFi exchanges there is less and less need to continue subsidizing LP on dexes, so we could technically drive this to zero

Interesting dimension is that while we can change terra side incentives with a gov proposal, Uniswap MIR-UST-LP incentives cannot be changed on the contract (no keys), so would require a contract fork of some sort or shuttle pointing balances to different contracts to migrate

1 Like

What do you think is the appropriate amount of rewards to allocate if we were to pursue 2? @sebnondzee’s argument that community pool should be allocated to dev grants & community development is fair, and we should be cautious to not overspend the community pool

  1. 1 weight on MIR-UST and 2 governance will incentivize hodl for MIR and spill more liquidity to terra chain from other chains
  2. yes if staking in governance causing 7 days lock up, reward hodlers, vs short term opportunity stakers, Claiming rewards renews 7 days lock up
  3. if cancelled 7 days no rewards penalty till lock expires?
  4. yes

I feel that the mismatch is acute now but only temporary. Actually adding more massets, hence more pools, will generate more rewards for mir stakers and less for lp stakers. Mir governance avoids impairment losses and drives all decisions therefore as a position has some privileges already on short term. The general goal could (I think we can agree on it) be to make massets pools very liquid and promote expansion more than equally rewarding mir stakers and people staking in the pools. Imagine to have 50 pools and have mir value doubling. In that scenario, which I consider realistic, Mir staking in governance would avoid impairment losses and the pools would provide less rewards. All without big changes in current structure.

I propose that MIR-UST-LP tokens should have option to vote but not without lock up period (only if they choose to partake in governance)

  • Lock up period of MIR-UST-LP should be 3, or 5 days here once vote is cast
  • Partaking in governance comes at a slight cost here (illiquidity for the 3 or 5 days)

Agree with @blackbird points on the value of current MIR governance structure avoiding impermanent losses. This is more noob / norm / risk adverse / friendly and bodes well to keep this as viable option for the valuable community members who have no interest in taking on the risks associated with yield farming - “inclusive DeFi”.

  • All NEW staked MIR-GOV tokens be subject to an initial 21 day lock up regardless of vote or not.
  • Every Poll voted on adds additional 7 day lock to tokens used in vote (unless the 7 days falls within the initial 21 day lock up period)
  • This would create a small “stake commitment renewal” cost to average voter (7 day illiquidity)
  • After 21 days initial MIR-GOV stake lock up period terminates - but this ensures that there is an associated cost and not only a place to park capital for free APY. -Keeps APY more healthy for serious stakers. Staking % becomes much more meaningful metric.
  • This more traditional staking option in MIR-GOV lends itself to wider audience both in terms of understanding and risk tolerance. Wider audience appeal has obvious benefits.

**My overall belief is that governance is a privilege and should always come at a small cost. By adding lock-up periods to all types of voters (mir-ust poolers, mir-gov stakers) we incentivize the community’s most committed and invested to also have the most influence while discouraging traders from having undue (or rather undeserved) influence.

Finally: 100% of CDP fees should remain unto MIR-GOV stakers only.

1 Like

Currently, there is a limit of 13 mAssets.

I believe that if more mAssets are added in the future, we will be able to get a lot of fees from CDP liquidation.

I think there is a need to distribute guidelines for quick mAsset additions, or to shorten the voting period for adding mAssets.

Also, it would be nice if rewards were possible even when voting.

While i think Seb has some good points , i think we are trying to solve a temporary problem and i do feel this is a valid use of the MIR community pool. MIR gov stakers are getting staking rewards from the fees generated when closing CDPS, and so as the platform grows there will be more value going to MIR gov stakers.

As the platform grows this issue will resolve itself, but there’s a bit of math worth doing to determine at what volume will MIR rewards compete with those of the LP pools. Increased adoption also means more MIR-UST , and UST-mAsset stakers which will bring down the APY for the LP pools.

In the meantime, i think we should target an APY of 100% for MIR gov stakers. While this is still much lower than the 300%, i think it might be enough to convince some of us hardcore reflections to invest in the MIR gov pool.

Another challenge we have is how can we differentiate MIR-UST staking from MIR staking, my view on this is MIR is a 10$ coin and i would rather be fully exposed to MIR and contributing to governance ,vs MIR-UST staking, but im sure not everyone shares this view .

I think a slightly higher APY and shorter voting periods will get us to where we need to be while the platform transitions in to the behemoth it is .

2 Likes

Some fair points from @dokwon @Papi and @nachodon above, I think our main discussion now revolves around the following:

  • 1 - Keeping an option for voters not to be exposed to IL by restricting governance to individual MIR tokens - which means getting rid of MIR-UST LP as a right to vote since we cannot combine MIR-UST LP tokens and MIR tokens when voting: they’re not worth the same weight.
    There may be some misconception around IL though - holding MIR-UST LPs will get you less exposed to the volatility of MIR, so one could argue it’s less risky than being 100% exposed to MIR solo.
    Regardless, this is the main question we need to answer before moving forward with anything

  • 2 - All current rewards except governance will be diluted soon by the upcoming launch of new mAssets and the opening up Mirror on BSC - this means we will see APRs for mAsset pairs go down to the 50s% very soon, and MIR-UST APR go down by a 3rd as well.
    This would make MIR governance staking rewards more competitive vs mAsset LP rewards , but still not comparable to MIR-UST LP rewards (we can do the math @Papi but I believe this would be true for as long as inflation exists on Mirror as-is).
    So my point is still to modify the current inflation schedule to fix this imbalance rather than use the community pool funds (a budget that we’d have to reassess every time we open up to a new blockchain where MIR/UST staking will be available, i.e. BSC soon).
    I agree gov rewards will increase as CDP fees accrue with growth of the protocol, but i’m not worried about governance staking yielding more returns than other pools, only the opposite is an issue.
    I am also aware of the admin key burn on the Ethereum side and understand that inflation schedule changes might be unrealistic - if anything this is something i’d support with a one-off community grant.

For the rest it looks like we’re aligned:

  • 3 - No cancellation
  • 4 - Lockup period for voting (the rolling lockup proposal from @nachodon is interesting but MIP 30 will already pass with 7 days lockup, so it looks like this will be a “test in prod” approach)

I concur. The high APR is only temporary as DeFi farmers are still skeptical on mAssets. It will naturally decrease as confidence grows (+more mAssets are introduced) and LPs come.

Regarding governance, an ideal solution IMO would be to reward MIR stakers for voting, and also letting MIR-UST-LP token holders to vote (with a lower voting weight and reward)

The misconception you speak of is real but volatility is something that most crypto users (especially DeFi users) are well acclimated to and accepting of. And fixing peoples misconceptions never comes easy. Besides- the volatility in mir should smooth out a bit as time goes on and inflation lessens.

That being said I can see the value and rational in your idea to move all governance to the LPs - especially since we can’t separate lp from solo mir staking. I could vote for this certainly (especially the weighted design) since it does seem to offer more influence to those providing more value to the platform compared to a passive staker who does nothing, in this case, to further security (as in the case of Luna stakers).

However- while my view on this is flexible I do find it interesting that several have already expressed desire to keep governance separate from the mir pool. This could be the market speaking - so to speak. I think this alone would be interesting to put to a vote. After this is decided we can add further measures.

But first I’d suggest attempting to get a bit more people weighing in here before putting to a vote. There might be better ideas out there.

And the 7 day lock must be out in place before I vote on this! Oh—and btw as far as I’m concerned the community pool should be balls to the wall to grow this thing into the beast it deserves to be. Nothing else.

5 Likes

Personally I believe that only MIR staked for governance should allow to vote.
About IL. I see it a bit different from you in our context. If we mention impairment losses on MIR we assume that we are long on it and keeping it passively could be better than joining a pool. So if you are long MIR staking it for governance could be better once the lp pool is no more subsidized and you can have voting privilege together with the best investment solution according to your view. Actually the calculation is a bit more complex: if the volume of an LP is 100% of its liquidity the APR is a bit more than 100% per year and, unless I expect the token of the pair I am long to go up 500% I don’t care too much of ILs…
In our case I expect governance staking to become competitive with LP participation on long term. In that case I think that again the market will self regulate with more people staking MIR for governance reducing its APR.
My general view is that before acting on the system we should let it adjust naturally. The main reason to act on the rules is if we see a lack of growth and we can convince ourselves that the cause of it are the current rules.

So it seems that we’re getting closer to an agreement :

1 - Keep governance rewards exclusively for MIR staking (no governance rights for MIR-UST LPs)

2 - Increase governance rewards (we should do a text poll on this, it seems like it will be either by diverting a portion of the existing inflation into MIR staking, or via using the community funds)

3 - No vote cancellation

4 - Lockup period for voting (MIP 30 - 7 days lockup)

@Sebnondzee - I think the changes should look like:

1 - Keep governance rewards exclusively for MIR staking (no governance rights for MIR-UST LPs)

2 - Put increase of governance rewards to a TextProposal --> if this proposal passes, then we can create a new contract that provides more rewards --> do a follow up community pool spend proposal thereafter

3 - Voting & staking changes

  • Voting period is now 7 days (MIP 30 - 7 days lockup)
  • Add a unstaking period of 7 days
  • If unstaking occurs and the user has voted on pending proposals that are still in voting phase, said votes are not included in the final tally of the vote

The reason for suggesting the above changes in 3 is to remove the penalty for users that have staked tokens from voting.

1 Like

Just to clarify :

2 -> Put increase of governance rewards to a TextProposal
We want to poll the community on how to best increase rewards into governance staking right ?
And ask if this should be done by:

  • Either changing the current weights of existing inflation
  • Or adding up to inflation by spending from the community pool

If yes then I’m ok and I also agree with the path you propose once the Text Proposal vote is cast

3 -> Voting and staking changes
Ok very fair, as long as I’m able to see votes according to “stakers only” (Total votes minus votes currently in unstaking period) at any time when looking at a proposal (not only in the final tally).

This will need its own governance proposal change as well right ?

If we’re all aligned then let’s put the proposals out this week !

2 Likes