YIP-65: Evolving YFI Tokenomics

Authors: @0xJiji, @banteg, daryllautk, HAtTip3675, @onlylarping, @vany365, @Wot_Is_Goin_On

Summary

Evolve the role YFI plays in Yearn over four distinct phases, cementing the vision of the token as the fundamental foundation of governance.

Status

Voting
This proposal is currently in the voting phase. Cast your vote on Snapshot.
You can learn about our voting rules in YIP-55[1].

Abstract

If adopted, this proposal seeks to:

  1. Direct a portion of YFI that is bought back by the Treasury as a result of BABY [1] as rewards to those YFI token holders who actively participate in Yearn Governance.
  2. Evolve the role YFI plays in Yearn Governance through four distinct components. These build on top of each other and thus come in a particular order:
    • 1: xYFI. Distribute YFI that’s been bought back with Treasury tokens as rewards in a YFI vault.
    • 2: Vote-locked YFI. Introduce ve-style locking of YFI (veYFI) for up to four years (exact max duration tbd), where a longer locking duration gives a greater share of voting power and share of YFI rewards. An early exit from the lock is possible by paying a penalty that is rewarded to the other locked token holders.
    • 3: Vault Gauges + Voting. Introduce vault gauges where vault depositors stake their vault tokens and earn YFI rewards according to their veYFI weight. YFI are allocated to gauges based on weekly governance votes.
    • 4: “Useful work” features. Expand the duties and responsibilities of veYFI voters, and their locked YFI, in exchange for earning additional protocol rewards. Pending the tbc v3 vault design.
  3. Give the mandate to Yearn Developers to roll out the above components at their discretion as and when they become feasible.
  4. Restrict the YFI eligible to vote in Yearn Governance as only those staked in xYFI (from Phase 1 and onwards) or vote-locked in Yearn (from Phase 2 and onwards).

Background

This section covers general information that readers may find useful in order to better understand the motivation and specification of the proposal. Namely the community exploration of various tokenomics ideas, the source and quantity of YFI rewards proposed to be used, some general information about Curve’s ve-mechanic, and what was deliberately kept out of scope of this proposal.

Call for tokenomics ideas

On October 6th 2021 there was a call for tokenomics ideas on Yearn’s governance forum[2] that attracted over 5k views. Over 200 people completed a survey aimed to express the preferences of YFI tokenholders, results of which can be found in the References section.[3] A discord chat gave those interested an opportunity to express their views. There were also multiple independent groups created with high degree of engagement, culminating in a signalling vote[4] which aligns with the components of this proposal. This proposal attempts to bring together the discussions that have occurred over the last 2 months.

State of Holdings & Buybacks

YIP-56: Buyback and Build (BABY)[1] was adopted in early January 2021, instructing Treasury to buy back YFI with excess Yearn treasury tokens. A few weeks later, the yvDAI vault suffered an incident[5] where 11m DAI were lost. The Treasury compensated those affected, making vault depositors whole. Since then, the Treasury has been on a path of recovering from this expense while at the same time saving for a potential market downturn and a prolonged bear market where a runway would be required during periods of lower protocol activity. While the details can be found in the quarterly financial reports[6] the high level summary is that Yearn has gone from $11m in debt in February to $30m+ in liquid assets in November.[7] Despite saving more than $40m, the Treasury has still been able to buy back ~300 YFI year to date at a cost of ~$10m equivalent.[8]

Now that the yvDAI debt has been paid off and the bear market buffer exists, Treasury expects to be able to direct ~$35-$45m to YFI buy-backs on an annual run-rate basis assuming TVL and margins stay constant.

This would equate to 1,100-1,400 YFI per year at a YFI price of ~$30k, or somewhere around 100 YFI/month. These bought back YFI could sustainably be used in a reward system, without impacting overall Treasury health negatively.

Curve’s ve-mechanism in a nutshell

An incomplete and very high level description of this mechanism is as follows:

  • Users can lock CRV into veCRV for a duration of up to 4 years.
  • Depending on the locking duration, users get different weights linearly: 4 years = 100% weight, 1 year = 25% weight, etc. The longer the lock, the more voting power and influence, and the more rewards can be earned.
  • CRV rewards are distributed through gauges in which users stake their Curve LP tokens. Each gauge gets allocated a different amount CRV to emit, based on weekly voting amongst veCRV holders.
  • Based on their veCRV lock, users can boost their rewards of up to 2.5x when they claim CRV rewards from gauges.
  • Unlike typical ERC-20 tokens, veCRV is non-transferable. This can be bypassed via wrapped derivatives such as the Yearn yveCRV vault or Convex’s cvxCRV product.
  • Each account can only have one single veCRV lock (of a certain duration).
  • As the lock duration decays, so does the influence. If a user initially locked for 4 years for 100% weight, once 1 year has passed, their weight will be 75%. Users therefore have to constantly extend their locks (up to the maximum 4 years) in order to maintain their weight.

Out of scope

The following topics were intentionally left outside of the scope of this proposal. Each one is potentially important and probably would be best served by its own stand alone proposal. They are also independent from what is proposed here, and could be added as extensions that build on top of the concepts covered in this proposal.

  • New YFI Emission. The current proposal assumes no new YFI emission or minting. Emission could be attached to the components introduced in the phases, but is not an explicit requirement at this point.
  • Liquidity Mining Program. A new liquidity mining program, whether funded by Treasury YFI assets or by new emission, is somewhat orthogonal to the topics covered in this proposal, and could be introduced with or without the concepts covered here. There are numerous trade offs to consider, better discussed in a dedicated proposal.
  • Unit bias conversions.
  • Detailed specs of “useful work” features. The v3 vault specification is still in the design and discovery phase. Whilst ideas such as vault insurance, safety modules, and vault allocations are mentioned in this context, it is premature to discuss specifics as such proposals are far from being ready to go live.

Motivation

Benefits

The following are some of the properties where the new tokenomics design is believed to improve over the previous state:

  • Incorporates YFI buybacks. The mandate of YIP-56 is unchanged, the new design builds on top of and integrates the bought back YFI.
  • Is a sustainable ecosystem. The new design does not create a drain on Treasury assets. Instead there are reinforcing flywheel effects where tokenomics rewards drive more TVL, that in turn drive more fees, that in turn drive more YFI buybacks, that is then used to reinforce the tokenomics.
  • Incentivizes a long term view on Yearn. Token holders are motivated to support the protocol over the longer term rather than to speculate in the short term.
  • Disproportionately rewards those most loyal. Weaker conviction holders effectively become diluted over time by the stronger conviction holders.
  • Limits rent-seeking benefits. Over time as the components are introduced, the design avoids holders being rewarded for nothing. Or letting the largest holders accumulate more at the expense of the smaller holders.
  • Makes vaults more competitive. Additional YFI earned from vault gauges are effectively added yield for depositors, in proportion to how dedicated they are in their support.
  • Motivates 3rd party protocols and DAOs to become YFI holders. Yearn products are used as a yield components of a broader DeFi stack, and integrated in wallets and protocols. With this design, they have incentives to direct rewards to vaults and products they are using.
  • A seamless experience for integrators. Participation is optional. This maintains the simplicity integrators have come to appreciate and makes it easy to reason about vault behavior. Only those who are motivated to do so can participate.

Future possibilities

The proposal also opens up doors for potential further improvements that extend or build on top of these concepts, including:

  • yOptions program for contributors, where contributors can acquire YFI at a discount based on how long these are locked into the protocol for.
  • Insurance/Backstop functionality, where vaults could be insured against loss based on the amount of YFI locked, which in turn could be used to draw up stablecoin CDPs to repay users in catastrophic events.
  • New YFI Emission to veYFI holders, where new YFI is minted according to some emission schedule and used to reward veYFI holders based on their lock duration.
  • Partnership program adopting veYFI to boost partner performance, where partners can earn more based on their veYFI size and lock duration.

Risks

  • Governance attacks, where one or several actors accumulate sizable positions of YFI and can control rewards and decisions of the protocol. These risks exist today, and are mitigated somewhat by the limited supply of YFI and how the strong demand for YFI amongst Yearn contributors makes such attacks costly.
  • Not enough rewards to make locking attractive, where vaults may not generate enough tokens to the Treasury to buy back enough YFI to motivate YFI holders to lock into veYFI. This has somewhat of a balancing effect, where as demand for locking decreases, so does the share of the rewards for those who actually do lock. If it’s determined that the equilibrium does not lead to enough YFI being locked, additional YFI could be minted and rewarded to veYFI holders as previously mentioned in “Future possibilities”.
  • YFI liquidity dries up. Currently YFI is traded on multiple centralized and decentralized exchanges. As demand for using YFI elsewhere grows, there may be a lack of YFI/ETH LP supply in liquidity pools and lack of interest in general YFI market making, leading to YFI becoming more illiquid. In such an event, additional incentives may be required in order to ensure a healthy liquidity exists for trading in and out of YFI. The Treasury may also explore owning some of this liquidity outright.

Alternatives considered

  • OlympusPRO style designs
  • Voters voting on fees
  • YFI as a fee discount
  • Distribute treasury YFI

Specification

1. Spend bought back YFI in tokenomics rewards

  • Use a portion of bought back YFI from BABY to finance a new tokenomics program.
  • The 6,666 YFI that were minted as part of YIP-57: Funding Yearn’s Future[9] are not used for this purpose.

2. Evolve YFI through four components

In required order:

2.1 xYFI

  • A typical Yearn vault with YFI as a want token (potentially repurposing the existing yvYFI vault).
  • Receives bought back YFI as rewards.
  • Does not have any locking requirement.
  • Does not charge fees.

xYFI diagram

2.2 Vote-locked YFI

  • Locking similar to the ve-style program of Curve.
  • YFI can be locked up to 4 years into veYFI, which is non-transferable.
  • The maximum lock duration is still tbd, but will be in the range of min 1 year, max 4 years.
  • Locking duration gives the same linear weights, so if max duration is 4 years, this is 100%, and 2 years = 50% etc.
  • Weights decay as the remaining lock duration decreases, and can be extended up to the max lock duration.
  • Replaces xYFI, where a user must have a veYFI lock in order to continue to earn rewards. No lock leads to no rewards. Maximum lock, continuously renewed, maximizes rewards.
  • It’s possible to exit the lock early, in exchange for paying a penalty that gets allocated to the other veYFI holders.
  • Penalty size may be fixed (i.e. 50%), or may be depending on the remaining lock duration.

veYFI diagram

2.3 Vault gauges + Voting

  • Vault gauges allow vault depositors to stake their vault tokens and earn YFI rewards according to their veYFI weight.
  • YFI are allocated to gauges based on weekly governance votes. Each gauge can get a different amount of bought back YFI to emit.
  • Based on their veYFI lock, users can boost their rewards of up to 2.5x proportional to the amount of vault tokens deposited, when they claim YFI rewards from gauges. The greater the amount of veYFI, the more vault deposits can be boosted for the user.
  • Inspired by Andre Cronje’s initial design of Fixed Forex[10], in order for gauge rewards to be claimed, the user must have a veYFI lock. Depending on their lock duration, they are entitled to a different share of gauge rewards: if max lock = 4 years, and user is locked for 4 years, they are entitled to 100% of their rewards, if user is locked for 2 years = 50% of rewards, if user has no lock = 0% of their rewards. The difference is paid as penalty to veYFI holders, as an additional source of yield.

gauges diagram

2.4 “Useful work” features

  • This is driven by requirements of the yearn v3 vault design (still tbd).
  • veYFI holders would be able to do useful work for the protocol and earn greater rewards as a result.
  • Examples of useful work could be, but does not necessarily have to be: Configuring vault parameters, vault fees, strategy allocations, providing vault insurance, protocol backstop, and many other additional functionality.
  • No useful work will be introduced without consulting YFI voters beforehand through YIPs.

3 Give Yearn Developers the mandate to roll out the above components

  • Components can be introduced based on the requirements as set out here, with minor modifications as deemed necessary.
  • Exact setting of parameters such as max lock duration and penalty system is left to be set at the discretion of the implementers.
  • There is no requirement to add all four components, but components need to be in the order outlined as they build on each other.
  • Timelines are at the discretion of the contributors doing the implementation work.

4 Restrict the YFI eligible to vote

  • Once xYFI is introduced, only YFI staked is eligible to vote in Yearn Governance.
  • Once veYFI is introduced, only veYFI is accepted voting power in Yearn Governance.

Signalling vote (closed)

  • Yes, I support this proposal.
  • No, I’m against this proposal.

0 voters

Non-binding poll

Binding vote

:ballot_box: Cast your vote on snapshot

References

  1. YIP-56: Buyback and Build
  2. Call for Ideas: YFI Tokenomics Revamp
  3. Yearn Tokenomics
  4. Snapshot
  5. yearn-security/2021-02-04.md at master · yearn/yearn-security · GitHub
  6. yearn-pm/financials/reports at master · yearn/yearn-pm · GitHub
  7. https://twitter.com/bantg/status/1461910717494398983
  8. https://www.yfistats.com/financials/YFIBuybacks.html (Page 2)
  9. YIP-57: Funding Yearn's Future
  10. Fair launches, decentralized collaboration, and Fixed Forex | by Andre Cronje | Medium

Change log

  • Dec 19: Add an author
  • Dec 23: Add snapshot link + update URLs to diagrams
19 Likes

Amazing multi-dimensional approach imo, great work. Strongly in favor edit: of the proposal and its in-scope! The BABY proposal should give us enough buffer to distribute and fund this for a long time. I see many buybacks possibly coming in.

6 Likes

Very excited for what’s coming in YFI. Strong support for the proposal.

5 Likes

Fantastic and perfectly articulated.

I support.

2 Likes

I am in favour of the proposal, however I do believe directing the majority of protocol earnings to token holders is not in our best interest currently. I believe this is a path we should consider in the future, but not currently. Instead I believe we should be expanding our treasury for darker days to come as well as be prepared to fund more development and vault strategists as this is our bread and butter. If we are to direct a considerable percentage of earnings, acquiring protocol owned liquidity for YFI-ETH should go hand in hand with that (which we could also build up slowly over time using earnings instead of minting new tokens/use our current treasury to incentivise bonds).

To be clear, I am in favour of evolving tokenomics into a voting escrow type of system and even pay out some earnings (only if gauges would be activated!). As mentioned in the proposal, gauges will allow for YFI to be used on a protocol level as we have seen with Convex and Curve and put YFI on treasuries as we are currently witnessing with KeeperDAO, Wonderland,… However, paying out the majority of earnings seems unnecessary to me at this point in time as we can use these funds to grow much more / protect ourselves better by keeping them in the protocol. Using this time of growth and expansion in the space wisely is what will set us apart from other projects when an eventual winter will come. At some point in the future, when our long term future (and DeFi as a whole) is much more certain, we can still direct more and more earnings to token holders.

Therefore, I support this proposal but I think an important parameter (that is not being discussed here currently) is how much earnings would we be paying out.

7 Likes

I think if Yearn Developers are given a mandate to roll out the components described in this proposal, some indicative timeline will have to be communicated once Snapshot is positive.
And should we not first test out the xYFI model once implemented before going forward and evolving a model without testing it for a significant period?

3 Likes

This is an amazing proposal! Maximum support!

3 Likes

A synthesis of the most popular tokenomics ideas, while leaving the door open for future innovation. Love it - 100% yes.

4 Likes

maybe thats something gov can vote on. set a floor amount of ETH/BTC/stables for treasury to be holding before we pay out to xYFI or veYFI. We currently have around $30m but if gov votes for it to be more than that then rewards can be paused til treasury is full.

4 Likes

veCRV’s non-transferability is bypassed via wrapped derivatives. Won’t veYFI be immediately affected the same way after release? There has to be a better way to implement the timelock.

1 Like

The behavior you claim it limits it will actually only contribute to in this case because only wealthier holders have the long time preference to remain illiquid.

Using the ve-mechanism ends up creating multiple classes of YFI holders, which is that really supporting Yearn’s inclusive nature?

Yearn can accomplish something similar with likely better economics using a simpler structure. Instead of using fees to buy Yearn at any price and then partitioning distribution among classes, it would create Uniswap range orders below the market price. Shorter term holders will sell bringing the price of YFI down and the range order would take those tokens out of circulation rewarding longer term holders. It would do this even more so because it’s buyback yield will be higher than a mechanism which buys at any price.

5 Likes

As it currently stands, the YIPs process is unclear: must we submit via the proper github repo or is forum discussion and snapshot voting sufficient? I only ask as so that it is clear the proper steps/workflow for this stipulation.

This actually will create a more lucrative instrument than YFI itself. Now the inevitable “…but what about past contributors…” question, ser.

I would like to see existing ecosystem partners more active, as it looks from the outside its seems entirely superficial.

mmmmmmmmmmmmmmmmmmmmmmm

2 Likes

Honestly, in my opinion, while I agree with some of your points, I also think the DAO hoarding tokens in the Treasury is redundant and pointless if we do not drive more value into the hands of active YFI token holders.

The DAO is in a healthy position, and would be even healthier if we tripled TVL and YFI was trading at $70-100k right? The superpower we have is fair and even distribution of the value capture, and this has to be represented by the token value (which it currently is not, and has been discussed at length in other posts).

I am mindful of a bear market…but some evidence would suggest we are already in a bear market since the summer - and look at our healthy balance sheet.

We can’t predict the market, and we certainly cannot predict how the composable, multi-chain DeFi ecosystem plays out and how all our partners interact with us. We have to focus on what we see and hear right now - and right now. I for one do not anticipate a return to a multi-year bear market - too much has changed - we are in the post-Dotcom bubble period and I feel have crossed the S-curve.

What YFI needs now is be taken seriously as a productive asset that rewards participation and engagement, integration and also rewards the input of Big Brains that write the strategies we all benefit from.

We are Great White Sharks in DeFi and I think we should not be afraid of the market - sharks need to swim forward or they die.

5 Likes

The key aspect, imo, of the veTKN model that makes it powerful can be seen in curve, Iron Bank (CREAM), and Fixed Forex and that is two fold. First is the payout in stables that don’t force lockers to choose to dump or not, and these stables being ibTKNs so claimers need not rush to claim.

I would consider how to include both or at least the ibTKN factor.

1 Like

Thank you for your feedback.

I don’t follow–Why would only wealthy holders have the preference not to sell their YFI? You can be cash-poor, and still take a long time preference on your YFI holdings, as long as that doesn’t end up being in direct conflict with basic needs such as food and having a roof over your head.

Either way, the bullet you quoted refers to an outcome where vault gauges requires you to actually use the product (use yearn vaults) in order to earn YFI rewards, rather than just staking your YFI to earn more YFI (which in a way is just those who already have YFI getting more YFI). The former aligns better with what the protocol ought to be rewarding, imo.

If by “multiple classes” you mean two: those who time-lock and commit to holding for a set duration, and those who do not, retaining the option to sell at any given time, then yes that’s right: We propose to reward the former group rather than the latter as their views align better with the long term interests of the protocol. How is this in conflict with being inclusive?

I understand this as an idea for how the Treasury best could buy back YFI from the open market. Sounds like something worth while to consider. This is however off-topic to the proposal in this thread, which is about what to do with the YFI once it has been bought back, no matter the method used.

3 Likes

My hunch is that for people where YFI makes up a larger percentage of their net worth they will like to access appreciated value to finally pay for life decisions deferred such as the purchase of a house, where as wealthier holders either have already made those decisions or have a broader asset base to choose from.

If I understand correctly there are more than two as you can lock for different durations and receive a different proportion of protocol cash flows. Token holders that lock for 4 years will receive different amount of cash flows than token holders that lock for 2 years.

4 years is a fairly long time in crypto given the pace of innovation. At some point there maybe projects that locked YFI holder will want to fund and where the yield of an appreciated YFI is not as attractive at that moment. They will be forced to sell at a discount to market price and when YFI looks attractive again they will require a stepper discount than would normally be required to compensate for the lock up risk. Ultimately the convoluted nature of multiple classes and lock ups will create a higher cost of capital for YFI.

Studies have been done (Ibbotson) on the liquidity premium, if you take away liquidity you should expect a higher cost of capital.

1 Like

If your hunch is correct, it sounds to me like those holders would naturally lock YFI up for shorter duration, and retain some of their optionality. There’s also nothing that prevents wrapped derivatives of veYFI (such as a cvxYFI for instance) from being created that would allow lockers to enter/exit easily, or even be used as collateral in lending protocols, accessing portions of the value without having to liquidate positions.

Thing is, nobody is forcing YFI holders to lock up for the max duration. We’re just designing a system that intentionally rewards those who do so the most. In other words, in compensation for taking the risk you outline, holders get outsized rewards compared to those who do not.

Holders have to weigh their personal circumstances and preferences, and choose the option that’s right for them.

There’s no such thing as a free lunch in this system. To earn rewards, you pay with optionality. This is a good thing I think.

The creation of veYFI would effectively be the equivalent of sYFI (staked YFI) with a more complicated structure to get access to the same cash flows.

If at any point in those 4 years token holders disagree with the direction that Yearn is taking they are being punished for exit. I think such actions will result in a higher cost of capital for Yearn.

Of course this is all academic, it would be useful if somehow these two models could run be run along side each other to see which produces better outcomes. Perhaps a temporary ve-mechansim and sYFI, for say 3-6 months at which point the results of both will be compared.

1 Like

I can’t do anything with my yfi because of gas fees though.

2 Likes

I do not quite see how paying out most revenue to YFI token holders equates to driving more value into their hands. As I see it, distributing this revenue dilutes our project to some extent (we do not increase the amount of reserves held by the protocol and therefore do not increase the amount of value one token represents. Our growth would merely be in terms of TVL). While some costs could certainly be very useful and benefit the protocol more than the costs, I do not currently believe distributing most of our revenue is one of them. Instead, I would rather see these funds in the treasury (and/or spent on development).

In theory, if most of the revenue were to be distributed, this should not actually increase the token value as mentioned earlier. What should increase the token value is spending our revenue on buybacks of YFI tokens. However, it is not necessary for us to then distribute all these tokens to holders as this would partially off-set the buying in the first place by people who decide to sell their newly distributed tokens. Therefore I do think it would be in our best interest to keep most of these assets (whether it be in YFI or other tokens) in our treasury for now. This would actually increase the value in the hands of YFI token holders rather than dilute the token.

I believe this would actually be the most offensive position we could take, as we further grow the amount of assets to deploy in battle and reinvest our revenue. In a later stage, we could still move into a more defensive position by distributing more and more revenue.

To emphasise: I think it would be beneficial to, as you said it, make YFI a productive asset (through the veYFI system) to some extent. I just do not think it would make sense to distribute the majority of our revenue currently. This is not an all-or-nothing decision and I think we should discuss the in-betweens.

5 Likes