CHIPs signaling phase: Vote Power Tax

This signaling proposal and the research prior to it were funded by the AADAO. The initial post outlining the concept and purpose of a “Vote Power tax” can be found here. In the initial post, we also propose a dynamic community pool tax, which is outside of the scope of this signaling proposal.

Author: EffortCapital | Blockworks Research

Summary

The Vote Power Tax is an economic solution to multiple issues the Cosmos Hub is currently facing:

  1. Concerns around validator economics as it relates to Interchain Security
  2. Poor distribution of stake

This tax would be equally distributed back to all validators as a subsidy (~6k+ ATOM/validator/yr) to keep all validators afloat as Interchain Security scales. At current ATOM prices, this tax alone should allow most validators to scale to ~6+ consumer chains assuming $600/mo/consumer chain OpEx. By making the Cosmos Hub more self-sustaining during this bootstrapping phase of ICS, it would lower the costs of security on consumer chains and potentially incentivize more chains to join the AEZ.

The ICS Economics Problem

While ICS is a great solution for bilateral alignment between the Cosmos Hub and Consumer Chains, the issues around ICS economics are well known. What must be understood is that the open questions around shared security economics is not specific to the Cosmos Hub, but is an unsolved problem facing every shared security provider. The level of adoption for an overwhelming majority of chains is not enough to sustain a decentralized validator set without meaningful inflationary rewards. This forces chains that want to align with the Hub via ICS to (1) have a token at launch, and/or (2) overpay for security that they may not initially need.

These issues might make teams question the value proposition of ICS altogether, and there have already been two examples of the issues around ICS economics:

Recently, a proposal passed to allocate 1.8M NTRN from the Cosmos community pool to the Hub validators actively validating Neutron, after frustration arose from validators running Neutron at a loss. While this is a good temporary solution, many may argue the best long-term use of the Hub’s NTRN (and other assets in the future) is to deploy this liquidity instead of using it to fund validator operations. In fact, the community has already signaled support for this type of initiative with the passing of Prop 864. In our opinion, it is imperative that the symbiotic relationship between Neutron and the Hub remains strong for the market to deem ICS a success with other future consumer chains potentially watching how this relationship unfolds.

Additionally, the Cosmos Hub community has already seen ICS economics deter promising consumer chains from joining. In November 2023, Noble delayed its plans to leverage Replicated Security (ICS v1) and wait for Partial Set Security (ICS v2) to launch. In the forum post, Jelena Djuric, Co-Founder of Noble, wrote

“…given Noble’s fee model under current volumes and near term projections, Noble’s integration as a consumer chain would mean that ICS validators would be operating at a loss out of the gate…We think the risk is high that either Noble is not accepted to replicated security at this time or, potentially even worse, Noble joins ICS and is viewed as a failure due to validators not being able to cover their costs.”

This begs the question, how should the Cosmos Hub navigate the onboarding of new consumer chains? If consumer chains need to worry about economic sustainability, shouldn’t they just become a sovereign appchain using their native token for security?

The Cosmos Hub needs to ultimately decide what its relationship to consumer chains are. RMIT framed this problem well in this forum post here. Is the Cosmos Hub taking a venture bet on consumer chains, or is it a purely infrastructure offering? If the former, the Cosmos Hub needs to structure its ICS deals accordingly and heavily subsidize security costs for chains like Noble, Neutron, and Stride. If the latter, then the Cosmos Hub needs to identify its edge in a shared security market that is becoming rapidly commoditized with multiple shared security solutions coming to market.

Poor Distribution of Stake

Historically, the Cosmos Hub’s Nakaomoto Coefficient (number of validators required to reach 33% VP) has oscillated around 7 and 8, putting not only the liveness of the Cosmos Hub in the hands of a select few, but governance as well. This issue is not unique to the Cosmos Hub, but is inherent in dPoS systems as token holders vote with their feet via direct delegation. Although this incentivizes users to choose perceptually “good operators,” it is a major centralizing force since users typically select the largest operators in the set.

If the Cosmos Hub is to be seen as a credible long-term infrastructure provider, we believe a market mechanism must be implemented to help combat these centralization concerns, while also ensuring large validators are truly aligned with the success of the Cosmos Hub.

There are also some potential outstanding questions as it relates to the evolution of ICS and distribution of stake. Will Partial Set Security lead to further centralization of stake? In our opinion, the best thing for the Cosmos Hub is to ensure all validators in the active set have the financial ability to potentially opt into any consumer chain they want to support by elevating the baseline welfare of the entire active set. The higher the baseline welfare, the more consumer chains the active validator set can secure, which will create strong network effects for the Cosmos Hub.

The reality is the Cosmos Hub’s security budget is more than enough to scale its ICS offering, but the poor distribution of stake has created a huge wealth gap in the active set. Chorus One did a great job quantifying the validator revenue gap in a quarterly research report here.

The Solution

Blockworks Research proposes a Vote Power tax that disincentivizes delegating to validators higher up in the active set to help balance out the stake distribution issues over time. Those who delegate to validators with more Vote Power would potentially subject to a higher tax depending on how much the validator has self bonded. This tax would be equally distributed back to all validators as a subsidy (6k+ ATOM/validator/yr) to keep all validators afloat as Interchain Security scales. At current ATOM prices, this tax alone should allow most validators to scale to ~6+ consumer chains assuming $600/mo/consumer chain OpEx.

One concern with introducing a Vote Power tax is that not all Vote Power is created equal. Some validators own a large share of their delegated stake and don’t face the same Principle-Agent problem that validators with low ownership have. What we don’t want to do is tax validators that have skin in the game.

Below is our proposed formula for how to calculate the Vote Power tax.

VPi=Total Vote Power of a given validator

VPs=Total Vote Power of the active set

VPsb=Vote Power Self Bonded by a given validator

In simple terms, this Vote Power tax takes the difference from the median Vote Power in the active set and the Vote Power of that given validator. It also gives the validator a chance to lower the Vote Power tax on their delegates by increasing the amount of self-bond to their validator. It is important to note that anyone who delegates to a validator with less Vote Power than the median of the active set would not face this additional tax burden (i.e the bottom 90 validators in the active set today).

Below is a chart that reflects how the VP Tax would affect the staking yields for delegates of the top 25 validators in the active set. Because the tax tries to normalize Vote Power at the median, the bottom half of the validator set (validators 91-180) would not face this tax:

The Limitations of the VP Tax (and Potential Solutions)

This solution is not perfect due to the possibilities of sybiling. Technically, a validator could split up its stake to lower its tax burden or self-bond its delegated stake. The only way a validator could do this is if it had full custody of the delegated portion of its stake either as a “whale” validator or an Exchange like Coinbase, Kraken, or Binance.

In the scenario of an Exchange validator, we believe it might make sense for the community to adopt a VP Tax Exclusion List of validators that cannot bypass the VP Tax regardless of self-bond amount. The only way an Exchange validator would be able to lower its tax burden is if it could prove the amount of ATOM owned on its balance sheet. We believe there are also reputational concerns that would deter a major exchange from attempting to sybil and believe this concern is not as credible.

While this solution may seem less than ideal, the reality is there are no solutions to prevent offchain actors like CEXs from bypassing onchain rules. The Cosmos Hub community has actually already set a precedent here for passing rules that don’t apply to CEXs - the LSM 25% cap only applies to onchain LST providers like Stride, Quicksliver, and Persistence. If given enough stake, Coinbase would be able to create their own cbATOM LST and bypass the LSM cap themselves.

In the scenario of a whale validator, it doesn’t make much sense for them to split up their stake to reduce their tax burden since they could just self-bond to their validator to do this. While a whale validator could still split up their stake to receive the additional 6k ATOM/validator/mo subsidy, we believe itd be easy to identify which whale validators are attempting to game the system and also believe an optimistic approach (assuming validators in the active set want whats best for the Cosmos Hub) is the right course of action.

Funding the Development

As this is only a signaling proposal, another proposal may need to be made to fund the development of the VP Tax in the event this proposal passes. Ideally the passing of this proposal is used to influence the funding strategy of AADAO or the development timeline of Informal, as the core development team of the Cosmos Hub.

YES - You agree that the Cosmos Hub should implement a Vote Power Tax to increase the baseline welfare of the active validator set and bootstrap ICS adoption.

NO - You do not agree with the implementation of a Vote Power Tax.

NO WITH VETO - You consider this proposal (1) to be spam, i.e., irrelevant to Cosmos Hub, (2) disproportionately infringes on minority interests, or (3) violates or encourages violation of the rules of engagement as currently set out by Cosmos Hub governance. If the number of ‘NoWithVeto’ votes is greater than a third of total votes, the proposal is rejected and the deposits are burned.

ABSTAIN - You wish to contribute to quorum but you formally decline to vote either for or against the proposal.

8 Likes

Very interesting proposal! I need some more time to digest this and will give some more comprehensive thoughts later, but I have one question that comes to mind immediately:

Does VPsb include tokens that are Validator Bonded for the purposes of meeting minimum LSM requirements? These are two separate bonding types, as far as I know, meaning that the same tokens can’t be both self-bonded and validator-bonded.

Given that the validator is (in almost all cases) putting up both self-bond and validator-bond, not including validator-bonded tokens as part of VPsb creates an incentive for validators higher up in the set to self-bond vs validator-bond. This is potentially harmful to both LSPs and the validators themselves, and we may see movement by validators at the top of the set to shift validator-bonded tokens to a self-bond instead (which could impede things like copy-staking for the Hub or even cap out LSM delegations).

If validator-bonded tokens are not included in VPsb, my one small note at this stage would be to consider including it for those reasons!

This is a very great point especially with PSS upcoming in the v17 upgrade.

The Vote Power Tax is a really powerful idea, many have tried over the years to solve the centralization issue on PoS blockchains without much success. However, the VP tax looks like it could work and if it does work and the decentralization of the Cosmos Hub greatly increases, then its value proposition also, since Neutron, Stride and all consumer chains will not only benefit from billions $ of security but a security which is more reliable and resilient given the constant increase in decentralization of VP thanks to the VP Tax idea and formula and this would then attract many new consumer chains to the Cosmos Hub in a positive feedback loop.

The implementation following this signaling proposal should be straightforward. That formula would take the different updated inputs of the validators above the median (Voting power, self-stake) to calculate the tax of each validator and the collected tax would then be equally shared amongst all validators with a frequency to be decided.

This proposal is a no brainer to 1) increase significantly the security of the Cosmos Hub with a more decentralised validator set 2) attract more consumer chains given the higher security offered 3) incentivize all validators to run all consumer chains, essentially providing 100% security to all PSS consumer chains. The only ones who would oppose this proposal are the few largest greedy validators thinking only short-term about themselves rather than the best for the Cosmos Hub which actually would also be the best for them in the long-term

We are definitely supporting this proposition and hope to see it funded as soon as possible.
In order to provide constructive feedback, we would note these points:

  1. We hope to see this VP tax deployed along with the “cubic delegation” proposition you also suggested. These two make a good pair, and we invite you to start the process for this one in parallel.

    1. When you say, “The reality is the Cosmos Hub’s security budget is more than enough to scale its ICS offering,” well, this couldn’t be further from the truth. These models only account for infrastructure costs and are widely underestimated for a proper professional deployment, compounded by the fact that operators’ costs to run them properly are much higher than just server costs. We are tired of statements suggesting validators are overpaid; it’s simply not true. Still, the vote power tax is a necessary reform to better balance stake distribution. We just think misleading statements to present the reform could have been avoided!
    1. To address the limitations presented by the possibility of sybiling by either dividing into multiple smaller validators and/or self-bonding custodian funds for big exchange validators: we find both to be highly unlikely given the regulatory scrutiny. It would clearly constitute an accounting violation to self-bond user funds, while splitting validators could be easily tracked on-chain. In the event of such misbehavior, we could consider actions, but we really doubt this will ever be attempted.
  2. We also think that an additional mechanism to safeguard sufficient decentralization should be presented. We propose to introduce an additional factor in the equation:

VotePowerTax = max (VPi - Median[VPs] - (VPsb^2/VPi), (100*VPi/VPs - 4)^2, 0%)

Adding this second factor will effectively cause the tax to be applied to validators above 4% VP and scale exponentially, without the chance to counter with self-bond. We propose setting it to 4% as this represents the level beyond which the Nakamoto coefficient falls below 8 (the minimum acceptable threshold for us). This offers an additional layer of security with negligible extra computational complexity. With this formula, a validator with 14% VP would effectively suffer a 100% tax, and all its rewards would be sent to other validators. This kind of extreme concentration should never happen, and if it does, this total tax should quickly force stake to redelegate and decentralize.

To give a practical example: in the current supply distribution, only Coinbase & SG-1 would be affected, respectively taking a 19.8% and 7.45% tax, thereby lowering their APR to 9.18% and 12.58%. This approach does not put them out of business while still bringing extra revenue to support smaller validators.


We are eager to hear the author @effortcapital and community feedback on this upgrade proposition. If there is interest, we are also willing to create a spreadsheet mathematical model to showcase the effect it would have in specific scenarios.

  • I like the idea
  • I like the idea but 4% is too low
  • I like the idea but 4% is too high
  • I don’t like the idea
  • other reasons (please answer with comment)
0 voters

Thanks for reading,
Govmos
pro-delegators-sign

2 Likes