Switch to gYFI time weighted voting power distribution

[Arcturus]

@SheckyGreen

I support any tweaks that reward YFI holders’ commitment to participate in governance, such as this. But seeing how things are unfolding, I’m wondering:

Does it make sense to merge the YFI vault with YFI governance, so YFI holders can do all three simultaneously: 1) stake to vote in governance; 2) stake to earn governance rewards; 3) earn YFI yield with the vault strategies.

Why force YFI holders to choose between the yYFI vault (yield) and governance staking (rewards)? Why not allow us to do it all—to the benefit of the token and the system?

Maybe I’m overlooking a reason why this is not a good idea. Is there a downside, perhaps it introduces more risk?

I explain it here more in-depth Governance, vaults, and yYFI - #5 by Arcturus but essentially staking YFI in a vault exposes it outside of our platform via lending on CREAM. What someone with sufficient funds could do is continuously borrow more YFI to get yYFI, virtually increasing the total tally of votes. In other words “double voting” among other issues described above by @skrypts

[iTo]

And if they vote, the time lock prevents the system from unwinding the positions thus preventing actual owners of YFI from voting at all.

[DaDi]

For those that continue to ask why no voting for vaults, a parallel is to check how stocks voting are done in TradFi. Share owners that open a margin account “loses” the right to vote with their shares, as they don’t actually hold the shares (loaned out). People that short sell the shares (borrowed) also don’t get to vote. Only the buyer that bought this particular share (loaned out -> short sold -> bought) from open market gets to vote.

This is to prevent double voting.

Vault is much like a margin account: when you put your YFI into the vault, you are also giving up your voting rights, as there is no way to stop double voting. In effect, the yYFI token you are holding is a right to claim a YFI, but not an actual YFI (that allows you to vote).

[jiecut]

days locked

So one thing unclear to me and to others. Is this a locker like curve, or is this just an incrementer for how long you’ve been locked. My immediate though was that this was a curve style implementation, but there’s some comments that think it’s an incrementer.

There’s some very different implications for both.

If it’s an incrementer, where gYFI power accrues over time, YFI stakers don’t really have as much at stake. Sure, they’ve been locked into the contract for a longer period of time, but they could leave whenever they want. Also it makes valuing YFI weirder, as new holders will be earning rewards at a much slower rate.

I’m more a fan of a Curve style locker. Higher rewards and voting power, if you actually lock up your capital for a year. These are the long term minded participants, who decide to lock their capital with this project. New holders are composable with long time members, as long as they lock up their capital for the same amount of time.

[skrypts]

One way to handle this is to burn gYFI or have it decay over time on exit.

[SheckyGreen]

Thank you for the replies @Arcturus, @skrypts and others.
I do understand the risks inherent with vaults moving the YFI around different smart contracts, and now I see the double voting dilemma as well. Still learning, thank you.

[banteg]

I’m a fan of locker too. Just clarified with Andre, he means gYFI to be transferrable.

Besides the obvious contradiction (transferrable locked tokens), I think we need to have a deeper look at possible adversarial effects of this change. Vote renting comes to mind first.

[timbit]

Maybe this is when the YFI vault comes into play

• If a new holder’s interest in improving/governing the protocol is strong enough, maybe there’ll be a higher likelihood that they will be willing to stake in gov for one year to receive the full ratio of gov rewards. Could act as a filter to prevent speculators from holding too much governance power and decrease mindless voting power bc I want to claim muh rewards

• If a new holder has little interest in improving/governing the protocol, maybe there’ll be a lower likelihood that they will be willing to wait a year to receive the full ratio of gov rewards. And instead they’d consider the yfi vault because the rewards would be higher for their preferred amount of time that they want to stake their yfi. Maybe the yfi vault will just act as an attractant for speculators / high velocity traders who hold yfi and decrease the amount of yfi staked in rando vegetable factories and overall make the protocol safer

[Daft]

May be 3 options

1. stake - remove any time - earn only gov rewards (voting gets you lock for some time)
2. Vault - remove anytime - earn only vault rewards
3. Stake and locked for long period - earn both gov and vault rewards.

[iTo]

I like this.
Questions:

1. What happens to gYFI when locked period ends?

2. Should it be made such that gYFI must be burned to unlock YFI?

3. If gYFI is burned to unlock, should renewal of locked periods be rewarded with further gYFI to further strengthen their stake.

4. If 3 is implemented then gYFI received must = gYFI required to unlock original YFI staked.

5. I very much like keeping rewards and governance separate. This does however eliminate vote renting no?

[ETH_Maximalist]

100% agree with this. My biggest concern is that YFI on CEXs (looking at you Binance) can come in and quickly skew votes and rake in the rewards.
voting YFI should be locked in for the long haul, and rewards scale commensurately.

[TheSouthSeaCompany]

If gYFI is transferrable, then I don’t see how it would achieve this stated goal:

@andre.cronje

distribute more voting power to long term minded decision makers

It would be analogous to creating different classes of shares with different voting rights, like Alphabet currently has, except that we’d create 365 different classes in one go.

A well capitalized actor – let’s call them Nancy – could make a tender offer for only “fully cooked” gYFI tokens (with at least 365 days on the clock), offering to pay slightly above current market value of YFI, creating an incentive for arbitrageurs to sit on YFI tokens for the purpose of resale instead of voting. Nancy or her proxies could then create a new, lucrative opportunity to stake YFI tokens outside the treasury, say “YAM 3”, with the objective of attracting many token holders to unstake from the gYFI wrapper.

With the above combination, Nancy would have a shot at gaining significant voting share or even control of governance while spending only a fraction of the cost it would otherwise take to gain that power.

In my opinion there could be a better alternative where:

• YFI is locked (there is no transferable gYFI) to prevent the problems above, and
• the voting power growth function is logarithmic instead of linear, to let new stakers to gain voting power a bit faster at the beginning

I think the second point would make governance more welcoming to newcomers while maintaining the basic idea of rewarding “seniority”.

[Jecherio]

Because it would be missing the point, if you stake in a vault, you not deciding the vote yourself anymore, thats andres choice then.

[zeframlou]

Fantastic points, totally agree that voting power should not be tokenized. It would introduce a multitude of risks without benefiting the goal of giving voters with more conviction more voting power. It’s important to note that we would need mechanisms for removing “dead” voters who stake but don’t participate, for instance if someone didn’t participate in the last 3 votes they get unstaked.

This proposal is generally in the right direction. It defends our system against whale takeover/dominance. It also allows smaller YFI holders more of a voice, which reduces voter apathy and incentivizes more community participation.

Increasing voting power logarithmically is a cool idea. It resembles conviction voting, where your voting power increases exponentially when you stake your tokens, and decays exponentially when your token is unstaked. Using CV would probably be the best solution, though the complexity of the calculations introduces risks, so I think it’s ok if we just used linear growth for now.

[Dark]

Won’t unstaking people for not voting just make people vote for whatever, just so they don’t get unstaked? Also I think gYFI should NOT be transferable. If binance then wants to lock its tokens, there is a high possibility that it would do with yfi what happened with steem and have too much locked, thus not allowing for withdrawals. Having transferable gYFI sounds like a mess of vote buying and selling voting power. I support gYFI only if it is non transferable and you have to lock your yfi away for a set time like crv tokens. What is the vote multiplyer for locking for a year?

[uplink147]

Feel 1/365 is too heavily weighted. 1/100 might provide a better balance. 1/100 allows newcomers to have a meaningful contribution sooner, while still providing protection against malicious actors.

[banteg]

I think it works not by accumulating, but by decreasing. For example, you create a year-long lock with 365 YFI, so you have 365 gYFI voting power, the next day your vote weighs 364 gYFI and so on. I don’t see a problem with extending or automatically extending the locking duration (e.g. you want to be staked for 30 days before you decide to unwind).

For some reference code see NuCypher’s StakingEscrow which implements restaking and winddown, and Curve DAO’s VotingEscrow.

[cleopatra]

I agree with the motivations and aim.

However I’m worried about the implementation. If gYFI is transferable there will be all kinds of secondary market shenanigans.

[crypt0tinker]

Guys do we have to create yet another token for this ? Can we use a score system like icon does, but sticking to the normal YFI token ? multiplying tokens like this makes YFI harder to adopt by newcomers…

[ETH_Maximalist]

Agreed. I don’t actually want to tokenize governance. Once locked, we shouldn’t be able to transfer the tokens. What’s the purpose of locking otherwise?