Proposal: Revamping Fees

Summary:

A few weeks ago, we (ParaFi Capital, SRS) opened up a discussion in the forum to optimize fees (here). We continue to believe that there is confusion on how fees work among most. Moreover, we believe that the current fee structure of vaults is suboptimal.

Since August 1st, 2020, yearn has generated $2.03M in total fees (withdrawal + harvest). If we were to enact this proposed YIP retroactively, yearn would have generated an estimated $9.66M in total fees ($8.06M in withdrawal fees + $1.6M in harvest fees), or 4.75x more fees than realized. Calculating realized harvest fees is not straightforward. However, for simplicity, we assumed $500M in vault funds, 30% APY, and 5% profit to arrive at $1.6M in harvest fees between 8/1/20 and 10/16/20.

With v2 upon us, the community has the opportunity to streamline fees in a meaningful way, which would multiply the fee potential of the yearn system and stand to benefit all stakeholders. We propose the following:

  1. Withdrawal fee: charge 0.5% on all withdrawals (not just idle funds) (or consider moving to a time-weighted management fee)

  2. Harvest fee: rename this to performance fee and charge a minimum of 5%, with the option to increase/decrease depending on the complexity of the strategy

Abstract:

yearn continues to be among the highest quality robo-advisors for depositors in DeFi. As proposed by this YIP, streamlining fees will introduce more predictability and make it easier for all stakeholders (developers, depositors, and governance members) to understand the yearn system’s earnings potential. More importantly, we believe the majority’s current understanding is that yearn charges 0.5% on all withdrawals and 5% on all vault profits. Therefore, the impact of this YIP is largely uncontroversial and positively asymmetric - we expect a limited impact on withdrawals/AUM growth and considerable upside as it would multiply fees.

In essence, this YIP clarifies the fee mechanics and considerably increases the earnings generation of yearn by potentially more than 5x (retroactively based on our estimates) on withdrawal fees alone.

As it stands, idle funds in vaults are not subject to withdrawal fees - only deployed funds are. We estimate only 0.1% of withdrawal volume is subject to the withdrawal fee. The harvest fee charges 5% only on “yield-generating” aspects on the strategy (a subset of total strategy profits). While conceptually these two make sense, we believe that this can be optimized/simplified to bring more predictability and resources to support yearn’s development (strategy creators earn 10% of harvest fees) and reward active governance YFI holders.

We understand that the harvest fee on profits will likely take more time to implement. Still, as the core devs finalize v2 specs, they should make an effort to allow for more modularity on the % performance (aka harvest) fee that is set per strategy/vault.

As a first step, we propose making an adjustment to the withdrawal fee.

Motivation:

Streamlining fees will multiply the fees accruing to the yearn protocol and bring more clarity and predictability. Removing the ongoing confusion on fees is essential to incentivize strategy developers, depositors, and active YFI governance members. We believe that the changes proposed in this YIP reflect how most of yearn’s users understand fees to work and, therefore, will not compromise the competitiveness of yearn to continue to attract more AUM across vaults.

Specification:

  1. Modify withdrawal fees to apply to all withdrawals (not just idle funds). Alternatively, remove withdrawal fee and move to time-weighted management fee
  2. Rename “harvest fee” to “performance fee” and set a minimum of 5% on all vault profits, with the option to increase/decrease based on the complexity of the strategy
  3. If possible, overhaul existing vaults. If not, modify future vaults and v2 specs

For:

Against:

Poll:

  • In favor of change
  • Against change

0 voters

7 Likes

love it. lets get it done

3 Likes

How is this not just rent extraction?

I’m all for better incentivising strategy designers, but this is a very blunt instrument to solve a currently nonexistent problem.

1 Like

So, from a design perspective we are thinking a lot about this. Firstly though, it is good to have data backing this assessment, that is a very important part of solving this problem.

For example, the Withdrawal fee at first glance seems to be a much larger part of yearn’s revenue model. However, on further inspection there may be a time component to it, as people have gotten more used to the Vaults and how they work, and have withdrawn less and less over time. We have heard of several teams hoping to integrate that have mentioned more permanently keeping funds in the Vault, which would affect withdrawal earnings significantly. Lastly, certain strategy types involving leveraged or locked funds might actually be impossible or difficult to withdraw from, setting a “withdrawal depth” on the amount that can actually be withdrawn in the first place! This makes the fee even less useful.

From an incentives perspective, this also doesn’t make a lot of sense. A withdrawal fee means yearn earns a bunch of fees when people stop using our system, when in reality management is a continual process, and we should seek incentives that align with the long term maintenance and performance of the Vaults. Hedge funds usually use a simple fee system that is something like a 2% yearly fee on AUM (assets in the Vault), and a 20% performance fee. Charging a management fee on AUM would perhaps be the best alignment with proper maintenance/management of yearn’s Vaults by Governance, and it provides for continuous revenue to cover operating costs proportional to the total assets that are being managed (incentivizing the long-term growth of yearn’s AUM).

There’s a lot of other factors to consider here, but thinking from an economical perspective is important to ensure we design a system that benefits all stakeholders, as you have said. I hope this comment is helpful for others, this is something we’d like to have figured out relatively soon, but obviously requires some discussion with the broader community.

27 Likes

Very good points. I actually think a time-weighted management fee is a better construct that a withdrawal fee. Especially as some depositors may shift between vaults and so a management fee is likely more impactful and fair.

I’m pro 5%-10% fee off of profits, but not okay with a withdrawal fee. Yearn should get no benefits unless clients get benefits. We should only earn money based off what how much we earn for the client.

8 Likes

Follow up on Doggie reply.

Said this couple of times already:

  • get exact numbers, first
  • model possible performance, second
  • propose, third

some proposals sit on forum without comments cause they need more thoughts.
Here is basic list of questions to cover:

  • Why do you want to keep 0.5% on withdrawals?
  • Why don’t you replace 0.5% withdrawals fee with higher harvest\performance fee?
  • Why do you want to charge smaller holders and get rid of free withdrawal buffer?
  • Do you understand that 0.5% enforced on all withdrawals hurts composability? On-top builders signaled that.
16 Likes

I’m for revamping fees, voted in favor
I’m against the 0.5% withdrawal fee, it hurts small users or people who just jump from protocol to protocol (degens or gamblers?) We should have a time weighted option for that.

3 Likes

In addition to Revamping the fees. Can we be more transparent we the stats of the Vaults? Show more information like a total deposit in each Vault, token minted if any, and the total amount of fees collected in each Vault.

2 Likes

Interesting timing for high volatility?

The outrageous returns in the past are not be indicative of the future. TVL has suffered already greatly due to lower returns. The fees are being paid by the investors. Once they realize higher fees reduce their returns even more, they may go elsewhere.

2 Likes

I am in favour of this! The 0.5% fee currently is to prevent abuse, and I would also try and keep it that way.
Having 5% performance fees was proposed quite some time ago on this forum. What ever happened to that proposal?

2 Likes

Not really
Today DAILY_GAS_USED is capped, so the only way DAILY_PAID_FEES can go up is if AVG_GAS_PRICE increase. What’s important is that today the demand to keep high and so DAILY_PAID_FEES will increase much more while keeping AVG_GAS_PRICE high.
Did Coinbase MakerDao Comp low fee was interesting for the market? No
The higher fee the better, trust me.

Very good points @milkyklim.

Once we have more granularity on harvest fees (with the subgraph deployed), we’ll definitely model it out.
Re withdrawal fee, I agree. We can move to a time-weighted management fee, which @fubuloubu proposed above. I think that is a better model. I don’t see withdrawal fees being a big part of fees going forward and it does hurt composability.

All in all, open to ideas to optimize fees going forward. Importantly, v2 will allow for more flexibility and setting fees based on the complexity of the strategy, which is I think the most important change.

6 Likes

I agree that a transparent and sustainable fee structure should be put in place soon. Voted for a change.

My favorite is still a performance fee because it is the fairest to everyone. A management AUM fee puts disproportional burdens on the low return / low complexity strategies.

Example: 1% AUM fee vs. 10% performance fee

  • Strategy 1: 4% APY, low risk
  • Strategy 2: 20% APY, higher complexity / risk

Using a management fee the conservative strategy would pay 25% of profits to YFI while Strategy 2 only pays 5% of profits. This effectively disincentivizes safer investments.
With a flat X% performance fee the playing field is leveled.

4 Likes

Interesting proposal.

I agree think that structuring and clarifying the fee structure is very important.

As others said, the withdrawal fee is something to think about. I’m personnally against, and think people should be able to opt out as they wish without additional cost.

Implementing a small management fee with a more significant portion directed to the strategists could incentive them more to maintain their strategy.
They could also be less inclined to seek high yield, more risky strategies, as part of their revenue depends on AUM and not performance.

1 Like

This is what I came up with assuming no withdrawal fee and only a performance fee.

3 Likes

Thanks. That table showcases what really drives YFI returns and what we must focus on:

  • AUM
  • APY of vaults

Both come hand in hand if we come up with good strategies, side products (insurance) and reasonable fees.

4 Likes

there is a few stats here https://feel-the-yearn.app/vaults. So you’re looking for amount of fees collected, anything else that would be good to add?

2 Likes

Ah ah !
The 0.5% fee is to earn money, not to prevent abuse.
Please be honest…

This fee doesn’t prevent abuse, it prevents deposit.

3 Likes