RPIP-62: Tokenomics Rework Prelude

[haloooloolo]

This is the official discussion thread for RPIP-62: Tokenomics Rework Prelude, also referred to as Saturn 0 or ETH-only minipools.

The proposal is now in a state where discussion on Discord has largely converged, the results of which are reflected in the current draft.

The main ideas are:

• Remove the RPL requirement for new minipools
• Set minipool contract commission to 5% (from 14%, commission of existing pools will be unaffected)
• Provide additional rewards based on RPL stake to temporarily increase effective commission from 10% (no RPL) to 14% (≥10% RPL collateral) as part of the smoothing pool claim
  • This temporary increase will end a few periods after Saturn 1 is released

There is still an open item on how to handle RPL rewards, which we’d like more feedback on. Please see the comment below for a summary of the options and a survey on your preference.

[haloooloolo]

The RPL reward cliff at 10% is still a point of contention. This post should serve to gather feedback for the main RPIP from a wider audience, and conduct an informal poll on how to proceed with the reward cliff.

The following is a brief summary of arguments that came up in previous discussion.

Remove cliff (peak APY between 0% and 15% borrowed):

• Simpler to implement
  • The fewer changes that need to be made to the reward spec, the sooner an implementation can be rolled out
• Simpler for users
  • If we keep the cliff, there will be some cases where users need to split their minipools across two nodes for optimal rewards
  • No need to monitor the node’s collateral level
• Incentives align better to TVL growth
  • Bond reducing to LEB8s will be a no-brainer: they will be more profitable in ETH terms and equally-or-more profitable in RPL terms
• Increased retention
  • Rocket Pool has seen a slow decrease in TVL over the last months
  • Part of this has been from NOs exiting to stay above the cliff
  • More of this has been from full exits. In these cases, the root cause is unclear, but loss of RPL rewards may be a contributing factor

Keep cliff (peak APY between 10% and 15% borrowed):

• Dilution of NOs over 10%
  • Rewards for nodes under 10% collateral lead to lower rewards for nodes over 10%; as most nodes are currently not receiving RPL rewards, the drop in RPL yield will likely be substantial (-30% or more)
• Potential increase in net sell pressure
  • Recent data shows that a significant amount of node operators still add to their collateral to qualify for RPL rewards. Removing the reward cliff would likely also remove most of the remaining reward-based buy pressure.
  • Rewards below 10% borrowed are going to node operators who have shown they do not want to add to their RPL exposure; they are in theory more likely to sell rewards
• Reduced retention
  • Increased sell pressure may cause downward pressure on RPL
  • The expectation of unfavorable changes in RPL price may lead to an increased rate of exits in order to reduce exposure, either in anticipation or as the effect on price begins to materialize

• Keep cliff
• Remove cliff

[drdoofus]

Just going to throw out there that Rocket Lend is a potential solution to the cliff issue and has already been approved for funding by the GMC. Not saying there are no issues with it (for, example, it is not quite ready to launch).

I am sort of putting this here as a placeholder, as I know @ramana is traveling and needs some time to put a response together.

Personally, I would not stop a needed change to the protocol just to save an otherwise worthy project, but I’m not so sure removing the cliff will have the intended effect. It certainly might, but higher RPL stakers will lose out on rewards. Overall, I am slightly on the side of keeping the cliff for now, but not rabidly so.

[Snocones]

Regarding the proposal, I’m in favor of cliff removal.

Regarding Rocket Lend, I’d hope the GMC could work with @ramana and the Rocket Lend team to ensure fair payment is recalculated and provided for their time spent, assuming their solution is made unviable / obsolete by a change in strategy.

I’m all about prioritizing the protocol, but the community is a huge part of that. I would hate to see contributors get stiffed and turn away from us.

[ramana]

Thanks for putting up a message on this Dr Doofus!

I would like to explain my vote, including the impact on Rocket Lend (which is one but not the only factor affecting my vote).

My opinions on the cliff, ignoring Rocket Lend

First, ignoring Rocket Lend entirely, I am not convinced that removing the cliff is good for Rocket Pool pre-Saturn-1. I am sympathetic to the arguments that removing the cliff produces RPL sell pressure (from those NOs that aren’t currently receiving RPL rewards but don’t believe in RPL any more anyway) and removes RPL buy pressure (from those NOs that are topping up to receive RPL rewards only because the cliff incentivises them to, if we ignore the unpredictable but in fact falling RPL price). And I don’t believe it’s obvious that removing the cliff increases RP TVL (is there really an ETH-only NO who cares about some small amount of RPL rewards vs not?). I would say it’s unclear either way what the effect would be, but I lean in favour of the cliff for now.

I am counting on Saturn 1 (and beyond) to deliver a more substantial and long-term rework, so in a way the decision about the cliff doesn’t matter so much. I do believe it is important to grow TVL prior to the Saturn 1 launch, but don’t think the cliff makes much difference to that (whereas ETH-only minipools does make a big difference - as does Rocket Lend’s substitute for ETH-only, for that matter).

My opinions on how Rocket Lend compares with no-cliff

Now about Rocket Lend. I believe removing the cliff would reduce demand from potential Rocket-Lend borrowers, since they no longer have the “unlock RPL rewards” motivation to want to increase their RPL stake. (This is in addition to the proposed ETH-only minipools which themselves remove the “unlock access to rETH commission” motivation to stake RPL, but which are much more clearly valuable as a TVL-growth mechanism.) However, the size of the effect is not clear.

What would the Rocket-Lend-supporting approach look like? I.e., suppose we do not remove the cliff and also get Rocket Lend deployed quickly (main bottleneck is community testing and feedback). In this world, all the ETH-only NOs who don’t care about RPL still come along (they are essentially unaffected relative to the no-cliff world: they just miss some small RPL rewards). Fully collateralised NOs do better because there are more RPL rewards for them (relative to the no-cliff world), and if they have enough RPL they can also earn additional yield on their RPL stack by being a Rocket Lend lender. But the most affected group is semi-collateralised NOs who do care about RPL rewards but do not want to take a large speculative position on RPL. This group with Rocket Lend has the option to take a more neutral position on RPL while still unlocking the rewards, by borrowing via Rocket Lend. In the no-cliff world, they can still avoid going highly long RPL but just get smaller rewards. If they want to increase their rewards they can still use Rocket Lend, but the size of the increase available is dampened by the lack of cliff.

Overall, it is not very clear to me which world is best for RP, and my reasons earlier for preferring to keep the cliff still hold.

Why I care about Rocket Lend besides

Now, what are my reasons for caring about Rocket Lend besides the effects on RP TVL I tried to analyse above? The main ones are these:

• I (and a few others) have already sunk a great deal of time and effort into building Rocket Lend.
• Rocket Lend is approved for GMC funding, and takes no protocol fee for itself. In other words, it is a pDAO-funded public-good project.

Ways forward, and a request for better coordination

I feel that it would be best for the pDAO to be more coordinated in our approach to handling the time prior to Saturn 1 (which we have already voted to approve). A coordinated approach would look like picking one of these options, regarding Rocket Lend:

1. Abandon the grant awarded for Rocket Lend in light of the effort being poured into RPIP-62. This would signal a clear intention from the pDAO to address the missed market-making opportunity between RPL speculators and RPL users entirely within-protocol (mainly by reducing RPL utility overall) and not with a market-making protocol like Rocket Lend. It would be clear to me and the Rocket Lend volunteers that there is no point continuing development and we could focus our efforts elsewhere. (The nicest version of this for us would be some token compensation for time already spent, but I do not mean to imply that these efforts were not made at our own risk.)
2. Instruct me to focus development of Rocket Lend for post-Saturn-1 usage. This would imply, for example, a more relaxed timeline for doing community testing and auditing. Most importantly, it would mean not auditing Rocket Lend until the Saturn 1 implementation is clear enough for us to know the exact contract API for megapools that Rocket Lend would need to operate with. (As it stands, Rocket Lend is built for now, i.e., for minipools, and if audited as implemented would need to be reworked and audited again post-megapools; my assumption is that funding this second audit would be reasonable if Rocket Lend demonstrated substantial pre-Saturn-1 adoption.)
3. Stop undermining Rocket Lend with the current fast timeline to implement RPIP-62 (or at least do not go for the remove-cliff version) and instead focus efforts on getting Rocket Lend deployed as fast (and safely) as possible, to give it a chance to start getting adopted and improving RP TVL. Then use the observed adoption (or lack thereof) to inform the details of an RPIP-62-like change (if there remains enough time before Saturn 1 is expected for it to be worthwhile making such a change).

Simultaneously funding Rocket Lend and voting in a no-cliff RPIP-62 with a fast implementation horizon is, I believe, an uncoordinated set of actions by the pDAO and we can do better by coordinating. So my vote on the poll as-is is for keeping the cliff, but my real request is that we figure out what we want to do along the lines of one of the options above (or perhaps another coherent option that I’ve failed to describe).

[ramana]

Update on my opinions about the cliff ignoring Rocket Lend

The best argument for removing the cliff I have seen comes from @knoshua on Discord:

the cliff was a bad design. People thought it would increase RPL value (if you fall below the cliff, you need to buy RPL) but at the time people overlooked that it is more likely to have the opposite effect (if you fall below the cliff, you can also exit minipools to get above the cliff again. this doesn’t require buying of RPL and makes the protocol shrink). While the pDAO didn’t realize it initially, I’d say this flaw has been common knowledge and part of consensus for a while now.
I’d argue that people will trust a protocol/dao more that is able to recognize mistakes and fix them rather than one that wants to hang on to them.

(and see the following little bit of discussion after Discord)

I would like to see pro-keep-cliff responses to this to get a fuller picture, but maybe this could lead me to change my vote.

I would still very much like a clear direction for Rocket Lend as described in my previous messages, and this would be even more important if we decided to go with no cliff asap.

[sfbahr.eth]

IMO we already removed the cliff for voting and it feels like keeping the rewards cliff only for validators without it as a requirement for new validators is inconsistent.

I think the whole RPL collateralization concept needs a new consistent story for why it’s valuable. This optional collateralization just for RPL rewards doesn’t seem convincing.

[ramana]

I’m keeping my vote as is for now since although I buy knoshua’s point I think it’s better to make the cliff removal change - as we already plan to - with the full Saturn upgrade that also includes other compensating changes.

[ramana]

@knoshua changed my mind in Discord with this:

you are not arguing against ETH only LEB8s, just the cliff part from what I can tell. Your claim is that in a world with ETH only and cliff, there will be more demand from potential rocket lend borrowers than in a world with ETH only and no cliff. Just skimmed your forum post and I don’t think you are actually explaining that claim, you just say you believe it’s the case.
In both scenarios, the only reason to use rocket lend is to boost rewards, both in the form of increased commission and RPL rewards. Borrowing with rocket lend only makes sense as long as the rewards boost is greater than the interest paid on RPL. Interest is fixed and known in advance. boosted rewards are variable and unknown. There is a risk that boosted rewards drop below interest rate, making the rocket lend setup unprofitable for the borrower. In order to get out of the RL position, the borrower needs to exit minipools, unstake RPL, repay and create new minipools.
A rational potential borrower would consider the risk of this happening and weight the reward of boosted rewards - borrowing cost vs. the risk of having losses from needing to exit.
The cliff makes RPL rewards less predictable. Other people dropping below or getting above the cliff affects your rewards indirectly. You dropping below the cliff affects your rewards massively and very directly. There is no guarantee that you will be able to borrow more RPL to get above cliff again. When ratio is falling, lots of people may be looking to top up and RL lender availability may dry up.
This increased unpredictability of boosted rewards makes RL less attractive and I believe it would lead to lower demand.

incidentally I think this shows that the impact on Rocket Lend is a big factor for me and/or that I’m not very confident about my impact-on-RPL-price or impact-on-RP-TVL claims

[epineph]

My thoughts on the cliff.

So I’ll lead this with a “I don’t feel strongly about this” and also “when voting to change tokenomics, I should feel pretty strongly.”

1. The decision to remove or keep the cliff is primarily about apportioning rewards; there is no increased pie available, just a method of distribution.

The intuitive result is that INCREASING rewards for undercollateralized (A) and DECREASING rewards for collateralized (B) will cause more A to stay and more B to leave. I think the burden of proof is on cliff-removal proponents to show that A will stay without B leaving in equal proportions.

(On the other hand, ETH-only minipools grow the pie.)

2. In reality, i think MORE collateralized will leave than undercollateralized will stay by dropping the cliff:

i) B is at baseline increased risk of leaving.

Let’s say you have Amy who bought in at 10% and is now 2% collateral, Jerry who bought in at 10% and has been topping up every month and is now at 9%, and Linda who bought in at 50% max, and is now at 10%. You can look at various metrics:

Has lost the most in ETH terms: Linda(9.6E) > Jerry(?5? ETH) > Amy(1.9E)

Has the most to lose in a volatile token: Linda(2.4E) > Jerry(2.16E) > Amy(0.48E)

Would lose the least ETH rewards by switching to solo (Linda (~9%) > Jerry (~12%) > Amy (34%)

Really, Linda and then Jerry have the most reason to be frustrated, have the most at risk in absolute terms, and lose the least proportionally by leaving rocket pool. So from my standpoint, if you had to choose i would say that you should focus more rewards on Jerry and Linda because structurally they are the most likely to leave; however, removing the cliff means that their rewards will decrease by ~40% or so, going at least in part to folks like Amy who are structurally more likely to remain NOs.

The MAJOR reason that i think that we are talking about decreasing rewards to these overcollateralized individuals is that we sense they are more RPL or RP aligned with higher loyalty or bullishness.

But like the story of the mutinous crew, it's important to realize how loyalty is earned and that it can be un-earned.

After a long voyage, a captain was confronted by several of his crewmembers. They were angry about being away from home and angry about their low wages. The captain saw he had a problem on his hands, but several other crew members stepped up beside him and said “captain you’ve led us well, we’ll stick with you through whatever comes.” The captain thought long and hard, and then said to the mutineers: “gentlemen, you can have the wages of these fine men beside me.” And that captain was never heard from again.

• not a real story

ii) Overcollateralized/topper uppres have been counting on the increased rewards, whereas folks who are significantly uncollateralized have not been counting on the increased rewards.

There may be people who are more likely to become NOs with the knowledge that they still get rewards if ratio drops, and if we had designed tokenomics like that at the beginning we might have a larger group of NOs, but that kind of incentivized demand takes time to develop and we just don’t have much time between now and saturn 1.

iii) The thing that is most likely to cause folks to leave is the precipitous decline of RPL ratio, not the loss of rewards.

Loss of rewards is ~6% per year, while RPL ratio has been declining >10% per month. It doesn’t take a lot of decreased buy pressure or decreased buy pressure to effectively lose an entire year of inflationary rewards. To me there is no doubt that decreasing rewards to those who are net positive stakers and increasing to those who have not purchased more RPL to stake will shift the balance towards increased selling, which I suspect will be worse for undercollateralized NOs in the medium term than receiving no rewards.

The loss of ratio there is likely only temporary volatility until underlying value is achieved with saturn upgrades (so as a long-term investor, i don’t personally care and even would appreciate a steep drop in ratio), but this is a REALLY important year to have our GMC and IMC significantly handicapped from incentives, as well as the team operating on a markedly reduced budget.

4. In general, I don't think there are going to be a lot of new ETH-only minipools that also care about RPL rewards over a 6-9 month timeframe

for instance, folks at 7% bonded could get 25% APR by topping off, but don’t (likely because they feel that RPL will fall further); if this is the case it is hard to believe that there are people who will buy RPL for 5% inflationary rewards. Most new buyers will be drawn to the speculative aspect of Saturn.

5. Removing the cliff makes explanations and specs significantly simpler, and may save time and allow roll out sooner

I suspect that if the amount of effort needed to achieve the transition to cliff-removal had gotten put into specs it would not actually delay much, but this is still the most potent reason to remove the cliff IMO.

6. Bond reductions are a case where the cliff is indeed extremely counterproductive.

RPIP 30 already targets folks to bond reduce, and if folks have not bond reduced in the last 16 months it seems to me unlikely that they will bond reduce purely for 20% more ETH rewards over a 6-9 month timeframe.

Overall I’m on the fence about the whole thing, and happy to go with no-cliff or cliff, however the pDAO votes; this is my attempt to give the argument in favor of keeping the cliff, even if it was not a great decision to enact at the beginning.

[NeverAnIsland]

Let me try to address some of your concerns.

Why would you top up if you’re not RPL-bullish? The only reason to top up is when you think your 8% RPL APR gives you more than RPL decline makes you lose. In the present downtrend thinking of even a 8% annual RPL decline is bullish.
Following this logic we may conclude that people who don’t top up are uncertain of RPL future and those who are bearish have exited – there’s hardly many things that can justify loosing more money.

So are there really many people who would sell their rewards? This must be bearish operators and I doubt we have many of them at this point.

[epineph]

Of this was directed at me, that did not allay my concerns.

Let’s examine your 8% NO. By topping up, he would get 11.3/0.2 = 56% Apr on this amount if ratio didn’t fall, and obviously if ratio rose would also benefit greatly on that amount. The reason that NO is not topping up is because he either 1) feels ratio will drop further or b) feels overcommitted from a RPL risk standpoint.

If that NO is not buying RPL to top up at 56% Apr (by exiting minipools to sell ETH → RPL or scrounging the couch cushions or borrowing from their grandma), why would we figure that giving them RPL at 6% Apr (likely that they now also have to pay taxes on) would encourage them to keep it?

[NeverAnIsland]

I think that NOs from the first group – that feel that ratio will drop further are actually two subgroups: A) “ratio will drop and never return” and B) “ratio will drop and return”. My point is that NOs from group A are mostly extinct by now, and NOs from group B have no reason not to keep rewards if given (let’s put taxes aside).

[Hodja]

I have spent a total about 5 minutes thinking about this, no analysis, and tbh these are not opinions coming from any deep thoughts, and I want to state it anyway.

I think simplicity makes for good tokenomics. The less artificially imposed rules on the market the better. If this was the initial design and a choice between cliff or no cliff, I would vote for no cliff. Just keep it simple.

Having said that I don’t really have strong feelings either way and just leaving it up to the gigabrains.

[leighm.eth]

What Hodja has written is exactly what I wrote in trading. We need to keep it as simple as possible. The less we have to explain to a new NO the better. Also, isn’t the cliff going away in Saturn 1 anyways, and if so, might as well get rid of it now, simplify the deployment of Saturn 0 and potentially get it to mainnet faster. If we get a bunch of new NO over the next 6 months, with 0 RPL staked, doesn’t that pay off for NOs who are staking RPL when Saturn 1 is released?
Thanks!

[aJoshR]

I believe that Rocketpool’s north-star should be to increase TVL of ETH staked.

Increasing TVL grows the community, rETH supply, node operator demand, and future prospects for RPL price.

In short. We all win if TVL grows.

Removing the cliff is the best way to increase TVL.

[Kevster.eth]

I was more in favor of keeping the cliff, even though I am massively below the cliff and not going to be above anytime soon (I value my ETH in minipools more than RPL rewards). Such a big change feels unnecesary when there’s something else in the works already.

But reading through the comments in this thread made me agree that removing the cliff could be better for the protocol (whose goal is to increase ETH staked via a decentralized system rather than CEX/Lido). Having the option to make more minipools sooner than Saturn 1 sounds, in fact really good.

[haloooloolo]

With discussion on the cliff seeming to subside and over 70 responses on the poll after 8 days, I feel comfortable moving on to a sentiment poll. As usual, this will lead to a Snapshot vote in 14 days if there is promising community sentiment.

While the current poll shows a preference towards removing the cliff, the gap is not wide enough to remove any doubt as to which variant would win in an actual Snapshot vote. Therefore, a full vote on the RPIP will use ranked choice voting with options for variant A (with cliff) and variant B (without cliff) alongside the standard option to vote against the proposal as a whole.

The RPIP is close to its final form now, with no expected changes other than potential clarifications and minor error corrections. Since the creation of this discussion thread last week, there have been three additions:

• Exclusion of 16 ETH minipools from bonus commission to further incentivize bond reductions
• Specification of the Snapshot vote format (as described above)
• Instructions to update the RPIP after the vote to reflect the chosen variant

RPIP-62 Sentiment Poll

• Support moving to vote; I think this proposal is great
• Support moving to vote; I think this is good enough
• Oppose moving to vote; I have a specific issue I’m mentioning in the comments below
• Oppose moving to vote; other
• Undecided; I have a specific question I’d like clarified in the comments below
• Undecided; other

[tektek]

One thing I want to add in favor of removing the cliff, is that if I have 5% RPL collateral, buying RPL to increase to 6% for example have no incentive, and doubling your RPL position to go to 10% sounds like a big investment, so I think that removing the cliff might not be as bad for RPL as people suggest

[Jota]

As far as i remember, the cliff exists to assure that in case anything bad happens, RPL can be used to cover losses, is this not an issue anymore?