AIP - 17: Redirecting a portion of fees to veANGLE holders

[GTRminator]

Hey everyone,

This is a proposal to rebalance the weighting of the protocol’s fees and interest to direct a portion of the protocol fees to veANGLE holders.

Context

In January 2022, the protocol launched a veToken model for ANGLE similar to the Curve system, which incentivizes long term locking of ANGLE tokens for the benefit of being able to vote on protocol changes as well as getting a share of the interest generated by the protocol. This mechanism was designed to align the interests of long term ANGLE holders with the protocol, and to reward that commitment.

Currently, the protocol generates revenue from 2 sources. First, Interest generated from the yield strategies and second, fees from minting/burning agEUR as well as fees from opening/closing perp positions. The current design of the protocol splits the revenue sources in the following manner:

Interest is split 50% to veANGLE holders, 30% to sanToken holders, and 20% to the protocol reserves.
Fees from minting, burning, and perp trading is split 40% to sanToken holders, and 60% to the protocol reserves.

The ANGLE token has seen relentless selling pressure over the last few months, even though over 40% of the circulating supply has been locked as veANGLE. In order to reduce the selling pressure of ANGLE and to improve the long term prospects of the project, we need to find ways to improve the value proposition for investors to buy the token and for farmers to hold instead of dump the token. One way to improve the attractiveness of locking ANGLE for veANGLE, and therefore reducing market sell pressure, is to increase the amount of protocol revenue directed at veANGLE holders.

initially, the majority of fees from minting/burning agEUR and perp trading was directed to the protocol reserves. The reason was to build up a sufficient buffer for the protocol to be able to defend the agEUR peg in case of adverse market conditions or other unforeseen events such as a hack of a protocol vulnerability. This goal has been accomplished, as the protocol excess reserves now stand at roughly 20% of the outstanding agEUR supply.

Additionally, I think it is logically consistent that protocol fees should be shared by those who have committed the most to the long term success of the protocol, e.g. veANGLE holders, vs sanToken holders, who have no such long term commitment to the protocol. In short, sanToken holders deserve a portion of the interest generated by the protocol, but have no logical claim on protocol fees.

Proposal

This proposal seeks to reduce market selling pressure on ANGLE, as well as further align the interests of long term ANGLE holders with the long term success of the protocol by increasing the proportion of protocol revenue directed to veANGLE holders.

To accomplish this, this proposal suggests that we change the revenue split of fees (minting/burning/perps) from:
40% to sanTokens / 60% to protocol reserves
to:
50% veANGLE / 50% protocol reserves

The split of interest revenue shall remain as is.

Implementation

Core developers will modify the appropriate contracts to reroute the protocol fees in the above ratio to the correct contracts.

Value to the protocol

Protocol fees are a significant source of revenue, and if we are successful in bootstrapping new pools paired with EUROC, there will be a substantial increase in revenue directed towards veANGLE holders if this proposal is passed, increasing the positive flywheel effect. This should have a strong impact on market demand for ANGLE, as well as increasing awareness of Angle Protocol within the crypto space.

The shift in the protocol fees will primarily come from sanToken holders, with only a small decrease going to the protocol reserves. This should ensure that the protocol reserves continue to grow at a healthy rate in line with the growth of the protocol. The reduction in revenue to sanToken holders can be offset somewhat by directing some ANGLE rewards via the gauges, ensuring that there are still adequate incentives for sanToken holders, in addition to the interest share they already receive.

Risks

The risks are that the reduction in fees going to the protocol reserves leads to insufficient reserves in the future. Although I believe this is unlikely, since this proposal will only decrease revenue to the protocol reserve by about 15%.

Another risk is that sanToken holders leave the protocol due to lowered APR. I also believe this is unlikely given that they can be incentivized with gauge emissions, and those emissions could theoretically be locked as veANGLE if they chose to, whereby they would then get a portion of those fees back.

[sogipec]

Thanks for the thoughtful proposal here.

One thing that we forgot to mention in our previous Discord discussions is that no fees from perpetuals are redirected to SLPs: all of them are going straight to the reserves of the protocol.
SLPs only receive fees from users minting and burning stablecoins.

I agree on the need to reduce the portion of the fees going to SLPs at this point. The protocol has given them too much for little value added, except the fact that they helped to grow the TVL of the protocol.

About the redistribution of the fees to veANGLE holders per se, I am still not clear about whether this is a good idea or not. I understand the need to reward people who showed long term commitment to the protocol, the potential for the price upside as well as the opportunity for creating flywheel effects where increased ANGLE price means increased APRs, thus increased TVL and increased liquidity which means higher revenue. I still don’t know though whether this is the most efficient strategy with respect to simply using this protocol owned liquidity to seed other pools (like a EUROC pool when it’s live).
True that both could be done in parallel.

In terms of implementation if we were to do this, I also don’t know if hard coding a percent of fees to go to veANGLE holders is the right way to do this. We need to put more analysis on this, but essentially all fees taken are not “real” protocol revenue as people are arb-ing the protocol when it updates its oracle.
On top of that, this requires changing the StableMaster contract which is kind of a heavy contract, we could find workarounds but this would probably take some time and efforts to build something that is robust.

One way we could implement this quite efficiently (even though it won’t be as neat as having it permissionless in the smart contracts) is doing the redistribution using off-chain computation but verifiable with on-chain data.

Like every week, we could have a script that fetches how much of the fees (or of the real revenue from fees have been made), and then distributes the right amount to veANGLE holders. We’d need to grant some roles to new addresses, but if the script can be ran publicly then it’d still remain transparent.

So to conclude on my point:

• I stand for reducing fees to SLPs
• not 100% convinced of distributing fees to veANGLE holders yet
• implementation is not so straightforward but there are workarounds and this comes with trade-offs

[tuta]

Thanks for the thoughtful discussion @GTRminator, the situation is very well explained. I dig into the situation myself as well, and here is my opinion.

First thoughts

Basically, I think that in the future, protocol revenues should be distributed to:

• operational cost (mostly SLPs and HAs)
• reserves
• remaining profit to veANGLE

Operational cost

I did a Core module cost analysis around mid-April that can be seen here. My takeaways were that 1) we are paying SLPs too much, and 2) a share of interest could be enough and more sustainable to attract HAs than ANGLE rewards.

SLP

Today, we pay SLPs through fees, interest, and ANGLE rewards. I agree that they are paid too much (around x2 what they should imo), and reducing that is a no-brainer.

I also don’t think it makes sense to pay SLPs transaction fees as they play no role whatsoever in that, but I would be in favor of increasing their share of interest to counterbalance this.

HA

Currently, HAs are only paid by their profit, and with ANGLE rewards. This is not sustainable for the price of ANGLE or for the quantity of HAs if the price of EURUSD were to revert.

I think we should stop rewarding HAs with ANGLE tokens, and distribute a portion of the interest generated by the protocol instead, depending on their position size. I did an analysis, and paying only 30% of what has been generated so far would have provided very reasonable APRs with 0 impact on the ANGLE price.

This would create a potentially self-sustainable HA mechanism, with no impact or dependency on the ANGLE price.

I’m very interested to hear the thoughts on that from Perps holders, @GTRminator and others !

Reserves

@GTRminator

initially, the majority of fees from minting/burning agEUR and perp trading was directed to the protocol reserves. The reason was to build up a sufficient buffer for the protocol to be able to defend the agEUR peg in case of adverse market conditions or other unforeseen events such as a hack of a protocol vulnerability. This goal has been accomplished, as the protocol excess reserves now stand at roughly 20% of the outstanding agEUR supply.

I agree with your take here, and I think the interest directed to reserves can be reduced. As a matter of fact, interest represent only 20% of what has been distributed to reserves since the beginning.

Redistributing revenue to veANGLE

Then, the remainder of the revenue should be distributed, directly or indirectly, to veANGLE holders. By indirectly, I mean that this revenue could also be used as reserve, not to back agEUR in case of emergency, but to grow treasury for other purposes like PoL or buying tokens useful for directing liquidity incentives. This would be another debate though.

Conclusion

In the end, I think transaction fees from the Core module, and interest from the Borrowing module should be shared evenly between veANGLE and backing reserves.

Additionally, a portion of interest (10-20%) should still be distributed to veANGLE holders, which could be seen as a kind of “management fee” from the DAO.

This is how this new distribution would have looked like applied to the past 8 months:

Remarks

Impact on SLPs

Currently, non-ANGLE APY for SLPs represents around 80% of their APY. A 60% reduction in revenue should reduce the yield for SLPs of around 50%, and their quantity from $35 to ~$17M.

We should also note that this reduction in APY could be balanced by an increase in ANGLE price, increasing overall APY for SLPs.

Aligning collateral parameters

The difference in the different distribution parameters by collateral in the Core module adds a lot of confusion and complexity when trying to compute revenue. I would be fore aligning all those parameters for each collateral.

A heavy rework

As @sogipec points out, all of this represents a big rework and will be time consuming if we want to do it properly.

Personally, I think it’s worth it to push to have a protocol as self-sustainable as possible in terms of operational cost + revenue distribution, though I don’t realize the amount of work it really requires.

It could also be a good narrative for the protocol as a whole when the market starts behaving more positively.

[GTRminator]

Thanks for the fast reply!

Here are some of my thoughts on your response.

@sogipec

One thing that we forgot to mention in our previous Discord discussions is that no fees from perpetuals are redirected to SLPs: all of them are going straight to the reserves of the protocol.
SLPs only receive fees from users minting and burning stablecoins.

I did not realize that the SLP’s did not get perp fees, and that they only go to the protocol reserves. Thanks for clearing that up.

@sogipec

I still don’t know though whether this is the most efficient strategy with respect to simply using this protocol owned liquidity to seed other pools (like a EUROC pool when it’s live).
True that both could be done in parallel.

Indeed, I think the best way forward is to do both in parallel. I am a huge proponent of using excess reserves for things like POL that help move the protocol forward. I just think the incentives for veANGLE are currently lacking, and that we can accomplish both without much sacrifice other than taking revenue from the overpaid SLP’s.

@sogipec

We need to put more analysis on this, but essentially all fees taken are not “real” protocol revenue as people are arb-ing the protocol when it updates its oracle.

I realize this, as you adequately educated me on this nuance in the forums . But the split that I laid out (which is of course just a starting point for discussion) only reduces the protocol share of minting/burning fees from 60% to 50%. That small of a reduction surely would not lead to a deficit due to the “fake fees” from the arbing. I can’t comment on the technical difficulty of making a change like this since I am not a dev, but I think it’s important enough for the protocol to revamp the revenue distribution that we should attempt it.

@sogipec

One way we could implement this quite efficiently (even though it won’t be as neat as having it permissionless in the smart contracts) is doing the redistribution using off-chain computation but verifiable with on-chain data.

I would not be opposed to this as a short term fix, since I know your hands are full right now with EUROC integration. So maybe the full smart contract rewrite can wait until we get the former sorted out and live.

@sogipec

• I stand for reducing fees to SLPs
• not 100% convinced of distributing fees to veANGLE holders yet
• implementation is not so straightforward but there are workarounds and this comes with trade-offs

I hope we can convince you of the benefits of increasing veANGLE revenue, without overly sacrificing protocol reserves. After all, veANGLE holders are ultimately the owners of the DAO, and should have at least a partial claim on all of the protocol revenue streams.

[sogipec]

Thanks for your reply! Your points are all pretty clear to me!

To be fully transparent, hands are mostly full at the moment with cross-chain deployment and preparing for our agEUR savings rate setup (which will involve EUROC).
Not the point of the proposal, but in terms of EUROC, there is a gauge proposal on Curve for a pool with agEUR, EUROC and cEUR. I think we should do one in parallel for just an agEUR/EUROC pool, just waiting for the Celo proposal to pass to push on this. This new one will allow us to double down on our AMOs, but we’re not in a hurry as EUROC has trouble scaling at the moment with all the € fud.

As for the rest, I really get the point of this proposal! If we go with the off-chain but verifiable with on-chain data way, there’s still a bit of work to be done, but that’s less an overhead than the other solutions!

[GTRminator]

@tuta

I did a Core module cost analysis around mid-April that can be seen here . My takeaways were that 1) we are paying SLPs too much, and 2) a share of interest could be enough and more sustainable to attract HAs than ANGLE rewards.

Thanks for this, it was a very interesting analysis! And I had not thought of replacing HA’s ANGLE rewards with a portion of protocol interest. I will think about this some more, but it is a very compelling idea. One knock against it is that the ANGLE rewards wouldn’t go away, they would just be redirected to other gauges, and therefore this change wouldn’t likely reduce ANGLE selling pressure (although it would likely make for more consistent and predictable rewards for HA’s, which itself is a desirable attribute).

@tuta

I think we should stop rewarding HAs with ANGLE tokens, and distribute a portion of the interest generated by the protocol instead, depending on their position size. I did an analysis, and paying only 30% of what has been generated so far would have provided very reasonable APRs with 0 impact on the ANGLE price.

This would create a potentially self-sustainable HA mechanism, with no impact or dependency on the ANGLE price.

I think this is worth trying out. From my point of view as a HA, it is only worth it to me to take on the risk if I can offset the high fees of opening/closing perps with positive carry. So as long as the interest to HA’s can at least mostly offset the fees, I am on board with this change. If we are going to change the perp module like this, we should also look to see if there is any way to reduce the perp fees without opening the protocol up to predatory arbitrage.

@tuta

In the end, I think transaction fees from the Core module, and interest from the Borrowing module should be shared evenly between veANGLE and backing reserves.

Additionally, a portion of interest (10-20%) should still be distributed to veANGLE holders, which could be seen as a kind of “management fee” from the DAO.

I pretty much agree 100% with what you have outlined, and I think the detailed conclusion you presented very much accomplishes what I was hoping to get done. Thank you for taking the time to put together such a comprehensive and thoughtful response! @sogipec , what are your thoughts on @Tuta’s proposal?

I know it will be quite a bit of work to implement, but I think it’s important, as it addresses several things at once:
-Improving the incentive structure for long term holding vs dumping
-Improving the value proposition for new entrants to the Angle ecosystem
-Stabilizing and right-sizing the incentives for hedging agents
-Reducing rewards for over-incentivized sanToken holders

if we can agree on the implementation structure of this proposal, and assuming the vote passes, could we then put it next up on the to-do list for after the multi-chain deployment and EUROC pools/integration?

[GTRminator]

@sogipec

I think we should do one in parallel for just an agEUR/EUROC pool, just waiting for the Celo proposal to pass to push on this.

I 100% agree here.

Is there somewhere that I can read about the agEUR savings rate setup that you are working on?

[TGT]

40/60 change to 50/50 is a good start. It shows that token holders are taken care of - something important for trust/brand/community building in a bear market. In traditional terms, we call it intrinsic value. We hope Angle could be the defi token with the most intrinsic value.

[sogipec]

Nope haven’t written up anything about this yet! We’re developing the contract, basically it’s going to be an ERC4626 where you can get a boost from owning veANGLE tokens, and this contract can be connected to multiple yield strategies

[GTRminator]

Do we have any movement on this yet? It seems there is very strong support for this proposal, specifically @tuta’s suggested implementation, which I wholeheartedly support. I think it’s important to do this simultaneously with the new Curve pool that will be launching with EUROC and cEUR.