[Proposal] "Change fee structure, treasury management, and revenue allocation" from Economics Design

Summary

From the VIP Summary submitted by Economics Design, a DeFi consulting speciality firm retained by Vesper Brewing Co.:

Change the withdrawal fee, performance fee, treasury management strategy, and structure for how protocol revenue is allocated. Optimize the fee structure and revenue allocation to improve long term growth in users and TVL. In its current state, the Vesper treasury is just holding VSP, which entails unnecessary risk through lack of diversification.

Relevant Documents

Formal VIP on GitHub

Medium Post from Economics Design founder Lisa Tan

Presentation Deck

2 Likes

My take always from VIP 9 (v2.0):

  • Raising the yield fees (from 15% to 20%): In order to offset the lower feeā€™s for transferring in-between Vesper pools.

  • Withdrawal Fees:

    1. Between Vesper Pools: 0.3%
    2. Outbound: 0.6%
  • Fee Allocation (Vesper Finances Revenue):

    1. Operations & Marketing = 25% (expected to cover a $4 million dollar budget)
    2. Developer = 5%
    3. Partner Program = 5% (incentives for pools created for partner protocols)
    4. Replenishing VSP reserves = 10% ā†’ 25% (extend the VSP pool incentives past their current expected expiration in August/September 2022. This would gradually increase from 10 to 25 in order to offset the gradual decline in VSP rewards from emissions)
    5. Rewards: 55% ā†’ 40% (gradual transition)
  • Algo Trading Strat: this would determine whether revenue should be diverted into stable coins, or an allocation of other tokens on the platform such as ETH, UNI, LINK, or VSP in order to limit our reserves exposure to any one asset in case of downward price pressure. This will not apply to existing VSP held in the treasury.

my opinion:
I agree with this proposal. I was against the prior proposal (VIP 9 v1.0) that proposed the revenue split, mainly due to price action. While some of us may still be against seeing a loss in our current rewards, I believe itā€™s in the best interest of the DAO that we strive at all cost towards sustainability. This is a good starting point, and would ā€˜set in stoneā€™ our initiative towards being sustainable and would definitely push the project in the direction towards decentralization.

3 Likes

I would like the withdrawal fees to be lower on the conservative pools across the board.

Withdrawal fees make more/less sense depending on the APY of the pool.

We should reduce them by 50% (to 0.3%) at a minimum across the board. Not just for transferring in between pools.

1 Like

As of now, withdrawal fees make up 93% of protocol revenue. Cutting withdrawal fees across the board would have a significant negative impact on revenue that would hinder funding required for long-term growth and vVSP rewards. The step to increase the yield fee should help reduce reliance on withdrawal fees and lowered fees for transfer gives an alternative for those who feel stuck in low-yield pools. Pending the result of this change, further reductions could be implemented in the future. But it is easier to drop fees than raise them, so a gradual change allows for examining the tradeoff in revenue with less risk of going too far.
-Kiefer Zang (Economics Design)

2 Likes

It is true that itā€™s easier to lower fees than to raise them. We can see how this goes. In the future I would like to see them be lowered, but I trust that a lot of science and analysis has been done on this proposal with longevity of the project in mind which is what we all want.

Iā€™ll support it.

1 Like

Great breakdown, thanks.

2 Likes

I support the revenue change and VSP treasury creation. Next step, I would like to see some small portion of treasury funds given to a group of active participants to utilize at their discretion for marketing, etc. Itā€™s important to start to move the needle on decentralization in meaningful, tangible ways. Engagement from our community is how we keep users dedicated to the projects success. I would like to propose a vip that supports this if this passes.

2 Likes

Iā€™m ok with 0.6% and 20% perf fee but I think the withdrawal fees need an if/then switch to keep TVL in vesper ecosystem:-

  1. if conservative pool has an aggressive pool sibling with same token (or aggressive pool had orbit pool sibling), the withdrawal fee should slide over to next higher pool with no penalty. i.e a withdrawal to aggressive should be no penalty to avoid TVL slipping awayā€¦ Donā€™t give users a reason to moveā€¦withdraw to aggressive with no fee vs withdraw with 0.6% fee makes it harder for user to move TVL elsewhereā€¦ Also, conservative pool users will want to move higher pools within 3 months and 0.3%x4 is 1.2% APY that conservative pools would have failed to achieve leading to IL and TVL moving elsewhere . We will capture higher withdrawal amounts in the long run
  2. If we have not earned the withdrawal fee of 0.6% APY within 6 months in a pool, withdrawal fee should be waived. (I had this happen to me in Eth conservative pool being in that pool for 6 months and getting IL with withdrawal fee. Only for being a vespernaut knowing future plans, I would have moved TVL elsewhere)ā€¦ If we donā€™t earn the withdrawal fee over that period of time, then the strategies are wrong or poorly performing. Iā€™m being generous saying 6 months, it should be lower but lets start there
1 Like

Iā€™m curious what expected impact would be if we removed withdrawal fee for transfers between pools. Retaining tvl and longer users is good strategy, and Iā€™m not sure Iā€™m into paying a total of almost 1% to withdrawal my tokens after switching. We should be rewarding users for switching to higher yielding strata, right? Tvl and revenue increases

1 Like

Proposal looks great - canā€™t wait to start building up that treasury!

Throwing an idea out there to those who want lower withdrawal fees: how about a lower in-platform withdrawal fee for conservative pools? Most of the concern Iā€™ve seen voiced in Discord over the fees references the low APY by DeFi standards that we have on conservative pools. Could look like:

.6% out of platform withdrawal
.4% in platform withdrawal from Aggressive pools
.2% in platform withdrawal from Conservative pools

[Side bar: is everyone aware that there is no withdrawal fee on Earn pools? As Earn product is built out, depositors will have a choice on fee structure between Earn and Grow.]

2 Likes

Yes, that is my point. Assuming an aggressive pool is 6-8% APY , the 20% performance fee cut would generate at least 1.2% APY in fees. By allowing a no-with-fee ā€œwithdrawal/transfer to aggressiveā€ option, we would get the 0.6% withdrawal fee at a higher appreciated amount + earn the 1.2% APY performance fee and happier userā€¦ .

1 Like

I agree easier to lower fees in future but see my points about supporting a transfer from conservative pool to aggressive pool with no withdrawal penalty or in future aggressive to orbit pool. If you did on-chain analytics of how many wallets went from conservative pool to aggressive pool to date, I think it would justify my pointā€¦ Goal should be moving more of the protocol revenue to performance fee vs losing TVL with 93% of current protocol revenue

1 Like

I agree with you on the benefit of having no withdrawal fees for transferring to a more aggressive pool as that allows users to freely maximize their returns while increasing yield fees for the protocol. That was a motivating factor for proposing the fee reduction on transfers.

However, we also wanted to keep the short term negative impact manageable as we make tradeoffs for long term growth. Withdrawal fees are a more immediate (and currently significantly larger) cash inflow to the protocol than yield fees. This proposal already involves a reduction to vVSP rewards, so halving rather than completely removing fees for pool transfers cushions that blow. After the impact of this change is determined and stakers have had time to adjust to the changes, the transfer withdrawal fee can be further reduced to 0 if desired by the community.

2 Likes

You are focused on near term treasury which is fine/important but miss the point about customer TVL retention/acquisition. We are at ATH and our TVL is struggling.

My main point is that withdrawal fee should be dynamic based on performance of the pools.

e.g. I would pay 1.5% withdrawal fee if APY on pool is generating avg APY of 12%. Currently, ETH conservative pool is 0.19%, it requires a user to stay in that pool for 18 months just to break even without IL (even with 0.3%) feeā€¦ That is not acceptable and folks will just revenge leaveā€¦

most defi will evaluate their yield performance (mental withdraw) after 3/4 monthsā€¦ even with flat rate of 0.6% with fee, that is 2.4% effected fee in a 3 month periodā€¦ Maybe only one of the stable coin pools is keeping that pace today. I would strongly urge a rethink

You have more faith than I do. I too would hope that some type of analysis has been done, but unless itā€™s published somewhere, I donā€™t trust that itā€™s been done. If anyone has evidence of what analysis has been done aside from a ā€œthis feels like a good fee structureā€ I think here would be a great place to post it.

Can we expand on ā€œThe algo-trading strategy for the treasury will determineā€ā€¦ what metrics go into this determinationā€¦

FYI, the AMA from Lisa Tan took place on Discord Nov. 11 at 5:00 p.m. U.S. Eastern and starts here:

Let me share more details!

Firstly, the goal of the proposal is to shift the protocol towards long-term growth. So what is the trade off here? We are looking at balancing between long-term growth (real productivity growth like what the gov does), short-term growth (instant returns and asset inflation in the short-term) and real value growth to existing users.
We managed the short-term growth so far, with VSP inflation and incentivizing growth to existing users in both returns of native token and VSP.
Now, we are shifting towards a new dimension: long-term growth. That means real productivity growth and returns for the protocol, not just asset inflation (which is not always the most sustainable) or increasing returns to existing users at the expense of the longevity of the protocol.

Next, we need to understand the returns by the protocol so we know where to balance the asset inflation (VSP inflation) and move it towards long-term growth. Digging deeper into the protocol revenue, we realised that most of the fees are from withdrawal. And that is the first step to solve. The withdrawal and yield fee. For this, we looked at historical data of Vesper, analysed wallet address behaviour and also to compare against competitors.

Thirdly, the treasury. You can have the best model but if you donā€™t manage the treasury well, as we see in many countries, you dont get real growth in the protocol/market. The first thing to consider is to make the treasury anti-fragile. And hence we brought in our quant finance guy to build the model and also to analyse the volatility to make sure the treasury is robust against bear markets.

The truth is, it is still a bull market, arguably, despite the few weeks/days in May and Sept. So we donā€™t have enough data to measure against how TVL is affected by ETH prices. We did an analysis on the token TVL against ETH and compared it with other competitors in the space - in the strat and no-strat portfolio/asset mgmt space.

And from there, we modelled what could happen in the long-run and the ML model is built from there

1 Like

the algo looks at hourly historical data and determine the difference between it. The algo will then do a simple decision: either to keep it at the native token because it is within the risk tolerance, or to switch it to a stable coin when the 2nd derivative of the price change is out of our risk scope

pls allow me to share more - we basically had various hypothesis and tested them out

H1: how the withdrawal fee, yield fee or VSP fee affect TVL in the pool?
Turns out, the R-squared analysis on pool APY vs each fee type found that only a moderate amount of the variance in fees can be explained by the APY. That means APY is important but actually less important than what we thought so initially.

However, when comparing R-squared analysis of APY with and without VSP tokens, there is a significant difference. E.g. 0.2802 with VSP and 0.0369 without VSP (!). That is a huge difference. Hence, in the treasury management, we focused on increasing the runway of VSP, since it is 0.28 points more impactful to have VSP

H2: can we just reduce the withdrawal fee to 0%?
We found that 93% of protocol fees have come from withdrawals rather than yield.

This indicates that the withdrawal fee should not be dropped too drastically in order to retain enough revenue to cover the operations/marketing budget requirements. Thatā€™s why we need to keep the withdrawal fee. Now, what about the intra pool transfers? We noted that 3.6% of capital and 10.2% of users are withdrawing and depositing to a new pool in the same day - this is based on same wallet address within a 24h period. It is reasonable to deduce the behaviours of users this way, to understand the ā€œintra-poolā€ transfer.

Lowered withdrawal fee for pool swaps should increase retention to the ecosystem, withdrawal fees currently disincentivize switching pools. Thatā€™s why we need to maintain those numbers - both from quantitative and qualitative perspective.

A higher expectation of the portion of users paying full withdrawal fees does improve revenue projections. With greater buffer for hitting operations budget target, there is more room for lowering pool transfer fee (after seeing real impact of this change at 0.3%).

H3: Tell me more about the competitors

One thing in our analysis is to measure if people are leaving VSP for other similar protocols. Well, itā€™s quite a mixed bag on TVL comparison of the competitors.

H4: However, are people leaving VSP for better APY elsewhere?

There is a very small correlation. one could argue that the other protocols are attracting capital, but the data is not significant enough to conclusively state that. Also see #3 for that.

H5: As for new capital entering the space, is it entering just VSP or other competitors?

Again, it is a mixed bag. The protocols all hold various tokens, and TVL analysis might seem like a good correlation, but the individual assets might vary a lot so it is hard to be conclusive. However, we can see that there is a general positive correlation of assets entering this asset management defi segment. A general positive correlation of all the TVL growth in the space

1 Like