NodeSet Review of RPL Tokenomics

[Wander]

Hello, RP community! In response to the recent tokenomics discussions, NodeSet has prepared a full, 27-page report on the subject. Please see below for an overview and summary.

This report is not yet as complete as we would like, but it appears the conversation is moving at substantial speed and a decision is being encouraged despite the draft status of existing proposals. In particular, we note that this report needs more citations and further editing. We invite the community to help contribute, comment, and continue the research.

We are also drafting RPIPs for our own proposals listed in this report. We encourage everyone to be patient while these ideas are formalized further.

Finally, I want to thank the following folks for their pre-publication comments and editing: @ken @Marceau @samus @rocknet @ib1gymnast @waqwaqattack

---

Version 1.0

Version 1.1 - Improved Example Roadmap section

---

Overview

In this report, we analyze RPL in the context of the current use case and what makes it sustainable. Additionally, we consider how the current indirect model compares to the proposed direct revenue approach. Next, we analyze the proposed tokenomics rework and assess the risks associated with each component individually and in totality. Lastly, we evaluate the impact of these proposed changes on NodeSet’s work before suggesting an alternative proposal.

Abstract

The current tokenomics proposal that has received the most attention fundamentally changes the relationship between RPL and the protocol and, therefore, the relationships between stakeholders in the system. Our findings indicate that the proposed new dynamics may be less healthy than existing dynamics because they create a financialized shareholder class that is economically incentivized to exploit the other actors (NOs, rETH holders) who make the system run, necessarily changing the makeup of these groups over time. In turn, this is likely to eventually lead to systemic instability and a transformation of RP into a protocol dominated by ungovernable large actors.

While RPL exists today as a utility token with an indirect revenue model fundamentally tied to node operation, the new proposals shift it towards an asset that can be priced more concretely by the market through ETH revenue accrual, opening it to increased levels of speculator ownership rather than participant ownership. Although this may make RPL more attractive to some, it does so by taking value from existing and future NOs, and for many scenarios, there is little change in terminal valuation versus existing tokenomics. Potential mitigations are suggested to prevent the optimization of the RPL/ETH ratio from overshadowing these and other serious risks associated with the proposed tokenomics changes.

Additionally, these changes would significantly impact NodeSet’s ability to grow Rocket Pool with Constellation. Altering tokenomics after the upcoming release of Constellation might, at worst, prevent its operation. At best, these changes would reduce its positive impact and delay future development to focus on complying with new changes to RP. Therefore, we suggest a different path.

[rocknet]

Thanks for this. I know a lot of work has been put into this, and has been quite a distraction while you’re trying to finish Constellation to help scale rocket pool. I hope the community reads and considers the perspective shared!

[cnupy]

Great research and you have managed to reflect all my concerns about the current Tokenomics revision proposal. Much to think about.

[yorickdowne]

Thank you. That’s well-written and I understand the concerns re centralization and governance capture.

Saturn 2: Eliminate oDAO via ZK or Optimistic Oracle Solution

Realistically this may be “Reduce oDAO duties”. The oDAO has a duty that is far harder to eliminate, and that is to apply scrutiny to contract upgrades. Only a fully ossified RocketPool could do without that duty, and that’s a long way out.

[Wander]

Realistically this may be “Reduce oDAO duties”. The oDAO has a duty that is far harder to eliminate, and that is to apply scrutiny to contract upgrades.

After Houston, the Security Council may take on this duty. The existing oDAO members (especially you) are well-suited for this, in my opinion.

[Darkmessage]

@yorickdowne

Only a fully ossified RocketPool could do without that duty, and that’s a long way out.

Ok, it’s off-topic but why do people within RP keep talking about ossification as if it was desirable?

Protocol ossification is the loss of flexibility and should be avoided! The word ossification is literally explained as “The process of becoming set in one’s ways or beliefs; rigid conventionality.”.

Are we talking about different things here?

[Wander]

I think this is important to the discussion, but a deeper tangent might necessitate its own thread.

First, ossification is extremely valuable for developers building on top of a protocol as a dependency. Adjusting to new changes (some of which may be controversial) is costly, and without ossification, this process is continual. Without ossification, RP will always be an end-user-facing product – no ETFs will be built on top of an unstable foundation. Even partial ossification (e.g. of a particular interface) is useful. Uniswap v1, v2, v3, etc is an example of this in practice.

Secondly, it’s a valuable security hardening technique, since an unchanging system has fewer chances to be exploited.

As you’ve noted, however, the flexibility of remaining upgradeable is helpful in many ways, and each protocol should choose its path according to its goals. Many old-school RP community members (e.g. myself, Yorick) tend to prefer ossification as we are more technically inclined and want to see secure projects built upon RP.

[samus]

@Wander

a decision is being encouraged despite the draft status of existing proposals

Sorry for confusion around the poll. My intent was to drive awareness and discussion. I wouldn’t say it’s decision time yet, and for those who are undecided you could even vote on How much have you read about the proposal?, but not on What is your opinion on the proposal?

Looking forward to diving into the report and discussing ideas further.

[Wander]

Thanks for the clarification, @samus! Looking forward to your thoughts. The report may be in a slightly more raw state due to our misunderstanding of the polls, but I’m sure the discussion will help refine it further.

[sckuzzle]

@Wander

…for many scenarios, there is little change in terminal valuation versus existing tokenomics.

Going to strongly disagree with this. Currently we are barring a huge number of potential node operators from joining who are RPL-averse. By allowing eth-only pools, we both allow all of the RPL-averse operators and also capture the value from them joining as if they had bought 80% of the RPL requirement (RPL-share / 14, or ~11/14).

changing the buy/burn mechanic to a simple ETH flow claim mechanic, with jurisdiction-specific tax concerns solved by a simple value-accrual wrapper token contract.

Could you expand on this? Everything I suggested was buy-burn because it was the only way I could find to main it capital gains. A claimable ETH flow is, to my understanding, ordinary income rather than capital gains.

@Wander

they create a financialized shareholder class that is economically incentivized to exploit the other actors (NOs, rETH holders) who make the system run

This is a potential problem, although I’m not decided on how large a problem it is. Does adding a capped financial incentive to stake RPL alongside running minipools potentially solve this problem? Something along the lines of “staked RPL (up to 50% value of nETH) gains double weight when collecting commission”. This way, If so, how much extra weight would it need?

If only a sliver of protocol revenue is available (half of the Node Operator revenue slice of the pie), then the community has effectively killed any incentive to invest heavily into building on RP moving forward.

The protocol can still generate revenue by minting RPL. The combination of burning RPL to increase its value and printing some for revenue both allows for a stable revenue stream to cover costs and accruing all surplus value into rewarding RPL-holders.

By opening the floodgates of RP to ETH-only minipools, especially alongside sub-linear bonding, the protocol could easily be overwhelmed by large operators clogging the validator queue, making it very difficult to access available rETH deposits.

Agreed that this is a potentially large problem. I think we need to both be careful with how much potential NO demand we open ourselves to and also ensure that we implement UVC to be automatically controlled to rapidly respond to changing market conditions, as pDAO votes are far far too slow.

Separately, I agree that equal access to LEBs is important and we should avoid sub-linear bonding.

(UVC…) At any time, the pDAO could opt to eliminate or redirect even more of the
revenue generated by Constellation for its users.

This is already a risk, even without UVC. By not implementing UVC, we are effectively subsidizing one group over the other. If your business requires enshrined subsidies from the protocol, it shouldn’t exist (sorry).

[samus]

I haven’t had time yet but I plan to read through the report and respond specifically to where I disagree. But I wanted to go ahead and respond to this since the question was also coming up in the discord…

concerns re centralization and governance capture

and quoting the NodeSet report:

Over the last year, NodeSet has spoken with numerous commercial node operators, including some of the largest in the industry. Many such organizations are interested in taking advantage of RP’s leveraged model to generate incremental returns, and the only thing holding them back has been RPL, which they cannot carry on their balance sheets due to volatility or legal concerns. Independent operators do not have this issue, so the current tokenomics helps keep RP decentralized and distributed. Today, large institutions may engage in bespoke RPL lending agreements. Still, there is limited RPL to be lent, limiting their potential individual impact on the protocol and preserving one of RP’s most significant differentiators: a decentralized operator base.

I disagree that our proposal introduces centralization/governance concerns.

It sounds like the assumption is: “By keeping the RPL requirement, we can prevent centralization since centralized actors won’t buy RPL”

Even if we take that assumption as true for a minute, my response would be:
Under our proposal:

• A large wave of (centralized) Eth Only NOs join
  • They have zero governance power (they don’t own any RPL)
• If we hit our self limit of 22% and still have outsized NO demand we can lower the “node_operator_share” close to 0% and instead make most protocol revenue go to “voter_share” which requires RPL
  • Now we have solved the problem in the same way that the original assumption suggested (require RPL to earn rewards, and centralized actors won’t buy RPL, so they are incentivized to leave and go elsewhere)

If you have qualms with “node_operator_share” being too high (too attractive to EthOnly Node Operators), we can start by giving “node_operator_share” a lower cut of commissions and instead give that to “voter_share” which requires RPL at stake, which resolves the issue under the original assumption.

[zk_sjp]

Hi, I’m new here! About me:

• I direct some portion of my investing towards crypto. Of that pie, some goes to Ethereum staking.
• In early 2023, I considered solo, RP or staking
• Chose RP as a good mix of less risky than solo (SmartNode and support), better returns, opportunity to learn and wanting to help decentralize Ethereum; in that order.
• Created 1x16EB minipool in Mar 2023
• I wasn’t thrilled about the RPL risk but thought “well there’ll be buy pressure from folks needing to keep their nodes topped up”, so it shouldn’t drop too much.
• I topped up the node a few times but then it felt like I was “chasing RPL down” and any returns from the node got eaten up in hardware costs and having to keep chasing RPL. I was spending more on RPL than I was earning.

I think that makes me a “Lisa” – although I can’t remember which proposal gave those personas.

I considered leaving, but I saw enough activity about reworking the tokenomics that it was enough for me to stick around. I trust the community to propose, debate and choose the best outcome. I agree with epineph here – holding RPL is a bet on the future, not on the current tokenomics persisting forever.

With my Lisa hat, what I want to see is:

• Never be in a position to have to “chase RPL down” in price (topping up node). “RPL Rewards for Undercollateralized Nodes” from NodeSet paper achieves this, as does ETH-only nodes.
• Deploy more validators with less RPL and less ETH (in that order). Whether that’s Bond Curves or reduced collateral, I don’t know yet. 70% RPL to ETH ratio (in LEB4s) is not for me. I’m here for the ETH, not the RPL, even if I like RPL. I wouldn’t go to LEB4s without sublinear bonding or reduced RPL collateral.
• Reduce gas if I start investing more (Megapools sounds great)
• I like proposals that reward small NOs and thus decentralization, but that’s less important than the previous two.
• I like proposals like Constellation where I earn more for the same operational overhead.

The great flattening shows “something needs to be done” to attract more folks like me or convince me to go from LEB16s to LEB8/4/2.

I think Valdorff’s proposal could do it, but equally the KISS one in NodeSet would too. Simpler is usually better, especially for smart contracts.

Tokenomics

I’m not an expert on this, but some thoughts after reading NodeSet paper and others here:

• I doubt RPIP-30 affected the price much. This is crypto. Narrative and macro overwhelms utility. Looking at Lido, the price action was more about the liquid staking narrative IMO. For me, it’s more incentive to buy in order to support a 2xLEB8s vs 1xLEB16, as I know I’ll get proportionally more that overcollaterallized folks.

• NodeSet’s rationale for the centralizing effect claim of ETH-only nodes and lower bonds are very concerning. Would like to see Valdorff, Samus and others’ thoughts on that.

• I agree accruing fees to the token risks having it classed as a security, at least in the U.S. That could be a…large distraction.

tldr; Current tokenomics don’t encourage new NOs or existing small NOs to convert to LEB8s, let alone LEB4s. The great flattening supports this. Something needs to change. Do we go big bang per ala SamDorff? Or a more cautious path like NodeSet proposes? Not sure yet. Following with interest.

Thanks everyone for such thoughtful discourse so far!

[Valdorff]

@zk_sjp

NodeSet’s rationale for the centralizing effect claim of ETH-only nodes and lower bonds are very concerning

Several things:

• Most of the claims are assertion based. Eg, that large operators are more willing to wait in queues with idle ETH losing them money than small hobbyists. That’s far from clear to me.
• One exception is the sublinear bond curve, which does mathematically favor larger stakers.
  • That said, there were 2 math errors around sublinear bonds, making the effect far larger than reality - see cell C27 in Rhetoric response - Google Sheets
• It’s important to differentiate centralizing RP and centralizing Ethereum. Fact is, over 50% of staked ETH is run by NOs with zero of their own ETH in the validators. A moderate hit on the RP side is worth it to improve the Ethereum side imo.

See also: Samus’ post above. If someone put in infinite ETH to be an NO gobbling up all possible rETH, we could reduce NO share, including to 0%. This fact alone should prevent it from being needed.

[Wander]

Thanks everyone for sharing your thoughts! I’ll address a few things, but I want to leave room for others to comment as well.

---

@sckuzzle:

Could you expand on this? Everything I suggested was buy-burn because it was the only way I could find to main it capital gains. A claimable ETH flow is, to my understanding, ordinary income rather than capital gains.

I’m not an accountant and this is a global protocol, so I can’t speak to specifics here. Personally, I don’t think technical decisions should be made to accommodate tax laws in certain jurisdictions since those can change (another example of change = bad for building stuff, as mentioned above). That said, a simple wrapper token contract can be made which accrues value at the rate of rewards, like rETH. You deposit your RPL, you get rRPL, and the rRPL increases in value (i.e. capital gains).

Does adding a capped financial incentive to stake RPL alongside running minipools potentially solve this problem?

It helps, but doesn’t solve it. Generally speaking, less rewards to uninvolved speculators = less incentive to exploit.

The protocol can still generate revenue by minting RPL.

This is true, but as we mention in the report, doing this will significantly reduce the advantage in terminal tokenomics vs the existing situation. If you account for pDAO/oDAO funding, it could be nearly identical.

If your business requires enshrined subsidies from the protocol, it shouldn’t exist (sorry).

To be clear, NodeSet as a business is in no danger, even if tokenomics were changed tomorrow completely disable Constellation somehow We have dozens of other opportunities for us to invest development resources, with more appearing all the time. RP is at the top of the list right now because we came from this community, and we wrote this report because we think Constellation’s positive effects on RP may be hindered.

---

@samus The centralization concern isn’t for governance, it’s for key-man risk and consequently rETH safety. If Coinbase has 90% of validators, then rETH is essentially cbETH with some extra complexity (risk). Because RP is permissionless, there’s no way to prevent this outcome unless the portion of ETH-only minipools is limited (Lido is going with 10% for these reasons).

This has historical precedent: at one point, DAI became basically USDC with extra steps, and MakerDAO has floundered ever since. It would be a shame to see this happen to RP.

---

@Valdorff

If someone put in infinite ETH to be an NO gobbling up all possible rETH, we could reduce NO share, including to 0%.

Governance can be slow, and damage continues during debate. And wouldn’t setting this value to 0% effectively disable the incentive for all ETH-only minipools? Wouldn’t it just be better to set up sensible TVL limits in advance so that this scenario doesn’t occur in the first place?

[Valdorff]

@Wander

Wouldn’t it just be better to set up sensible TVL limits in advance so that this scenario doesn’t occur in the first place?

You have, as with most of your claims, zero evidence that this is realistic. The extremely unlikely scenario doesn’t need to be dealt with because (a) it is extremely unlikely and (b) has a solution which makes it unprofitable for the abuser. Why do you believe a big institution is more willing to sit with ETH idling in a queue than a hobbyist? Why do you think they’re willing to risk governance reacting (as we would)? Pretend we’re slow AF and it takes us 3 months. They would probably net lose based on gas alone. Pure FUD.

[Wander]

Why do you believe a big institution is more willing to sit with ETH idling in a queue than a hobbyist?

They don’t need to sit in the queue with all their ETH. They can simply deposit 1000 ETH at a time to wait for rETH, re-upping when necessary. Or they could even purchase the rETH themselves to expand their reach further, gaining liquidity on some of their ETH without a need to pay any commission.

Why do you think they’re willing to risk governance reacting (as we would)?

Because they have something to gain (extra APR) and nothing to lose – even with the pDAO setting ETH-only NO portion to 0, they still have solo staking APR.

Pure FUD.

I’ve already asked you once before in another thread to please keep the conversation professional. I won’t respond to you anymore if you aren’t respectful.

[NeverAnIsland]

@Wander

Governance can be slow, and damage continues during debate. And wouldn’t setting this value to 0% effectively disable the incentive for all ETH-only minipools? Wouldn’t it just be better to set up sensible TVL limits in advance so that this scenario doesn’t occur in the first place?

Concentration won’t happen momentarily either. I guess, we should define CA (concentrated actor) TVL limit in RPIP for pDAO to start formal discussion in case it’s reached. This would serve as a reference and a warning for CAs.

[rocknet]

I guess, we should define CA (concentrated actor) TVL limit in RPIP for pDAO to start formal discussion in case it’s reached. This would serve as a reference and a warning for CAs.

You can’t determine so-called concentrated actors, right? Sybil is fairly easy, and in fact easiest for industrial staking providers that already have pipelines and distributed resources. They could look like any other set of new whales with a few hundred validators each node. Do you mean limit ETH-only validators as a whole?

[rocknet]

Has the idea of allowing ETH-only validators only as add-ons to existing nodes ever come up?

• User sets up a minipool with RPL bond.
• For each x/y combo of ETH bond and RPL (choose your adventure parameters based on research), user can create one new ETH-only minipool (probably LEB4 by the time this happens)

This has the effect of lowering the RPL collateral requirement. Of course it benefits only those with more than 1 minipool worth of deposits available, but it doesn’t entirely change the nature of the protocol. As mentioned in my writeup, I advocate moving smaller (and fewer) levers, and observing.

[drdoofus]

Not saying it’s a bad idea or that it won’t work, but part of RPs issue getting Node Operators has been that a number of people out there just do not want RPL and won’t start a node as long as they need it. We’ve seen a lot of angst in all social media over this issue. I feel like we better listen to what people are telling us over and over. RPL buy in requirements hurt us for a segment of crypto community.

As other platforms come online that don’t require anything but ETH, the fear is that new NOs will flock to it. Once some new protocol (or Lido CSM) gets branded as the great new thing, we could lose out permanently. Why would they ever move to RP from where they first land?

For the people who like RPL, I do like your proposal, though. I would certainly love to both have RPL pools and add on non-RPL pools.