Cosmos Hub Tokenomics - Fiscal and Governance Policy (Blockworks Research)

TL:DR - As a follow-up post to our Cosmos Hub Fiscal Policy post here, where we introduce a Dynamic Liquid Staking Tax, Blockworks Research is proposing to change ATOM’s fiscal policy from a static 10% community pool tax to a multi-pronged tax approach that includes a (in addition to the LST tax):

1. Vote Power Tax
2. Dynamic Community Pool Tax

Although outside of our scope for fixing ATOM tokenomics, Blockworks Research also proposes the idea of “Cubic Delegation”, which we go into more detail below. Cubic Delegation has the potential to fix the centralization concerns with respect to onchain governance while ensuring quorum can still be met.

Pitfalls of dPoS with Onchain Governance

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.

Explaining the LSM “Validator Bond”

To help solve the Principle-Agent problem inherent in dPoS systems, a “validator self-bond” was introduced to align node operators and their delegates by showing delegates that their validators have “skin in the game”. Unfortunately, the Cosmos Hub (and other dPoS chains to our knowledge) never instituted a minimum for how much self-bond a validator should have relative to how much stake is delegated to them. In theory, the primary concern is that a malicious validator could convince delegators to delegate, unbond all of their self-bond, then commit an equivocation and get their delegators slashed without incurring any punishment themselves.

These Principle-Agent problems are compounded with liquid staking providers since they have their own governance structures external to PoS chain’s native asset being liquid staked. As a risk-mitigation feature to ensure the safe adoption of liquid staking, the Liquid Staking Module introduced a requirement that validators self-bond ATOM to be eligible for delegations from liquid staking providers and be eligible to mint LSM tokens. This new mechanism called the “validator bond” is distinct from the current self-bond mechanism, but functions similarly.

To quote @zaki_iqlusion in the LSM forum post:

The maximum number of tokens that can be delegated to a validator by a liquid staking provider is equal to the validator-bond multiplied by the “validator-bond factor.” The initial validator bond factor would be set at 250 and can be configurable by Cosmos Hub governance.

With a validator-bond factor of 250, for every one ATOM a validator validator-bonds, that validator is eligible to receive up to two-hundred-and-fifty ATOM delegated from liquid staking providers.

Using a validator bond in the LSM got us thinking…

What other ways can we use this mechanism to fix the Principle-Agent problem as it relates to onchain governance? Can the Cosmos Hub benefit from the current poor distribution of stake while also creating a market mechanism to help fix this issue over time?

Below is a diagram of the current state of the validator/self-bond and the potential future state of how to leverage this validator bond. We explain each in more detail below.

Fixing Stake Distribution through a Vote Power Tax

In addition to a dynamic liquid staking tax, 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 be subject to a higher tax.

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.

Where:

  1. VPi=Total Vote Power of a given validator
  2. VPs=Total Vote Power of the active set
  3. 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).

As an example, lets look at three validators in the set that have more than the median (~0.17% VP today):

  1. Coinbase Custody (7.17% VP, 0.0% self bond ratio): Vote Power Tax = ~7%

  2. Dokia Capital (3.38% VP, ~15% self bonded ratio): Vote Power Tax = ~1.01%

  3. Allnodes (3.56% VP, ~0.0% self bond ratio): Vote Power Tax = ~3.39%

Although Dokia and Allnodes have similar amounts delegated to them, Dokia Capital’s delegates see a ~2.38% lower tax burden (a ~70% decrease) because of the amount Dokia has self-bonded to their validator.

If this Vote Power tax was implemented today, the Cosmos Hub would generate an additional ~900k ATOM in tax revenue in year 1. This revenue could be equally distributed back to all validators as a subsidy (~5k ATOM/validator/yr) to keep all validators afloat as Interchain Security scales, allowing the Cosmos Hub to benefit from the poor stake distribution.

If the community decides to implement this tax, Blockworks Research recommends re-distributing this tax back to validators equally only if the Replicated Security soft opt-out is removed, as this tax alone should allow most validators to scale to ~5 consumer chains without running at a loss (at ATOM current prices).

Below is a sensitivity analysis that shows the approximate Vote Power Tax a validator and their delegates would face today depending on how much is self-bonded to a given validator. With Coinbase Custody and SG-1 at the top today with ~6.5-7% VP each, they would need to self-bond around 20-25% of their entire stake weight to no longer have this tax burden.

While we know a Vote Power tax will likely be a contentious point of discussion, we believe the community must start considering solutions to combat the centralization concerns of the Cosmos Hub.

Cubic Delegation

One of the issues with combining dPoS and onchain governance is the overwhelming amount of power validators have in deciding the outcome of proposals with low individual voter turnout and validators receiving 1:1 vote power of their delegates.

This begs the question…should validators have 1:1 governance vote power of their delegated share? The idea of “quadratic voting” has been written about at length by Vitalik and other leaders in the space, but has seldom been implemented. With quadratic voting, each validator’s voting power equals the square root of the total delegated stake it accumulates (a simple example is shown below).

Today, Axelar uses quadratic voting to help combat the centralization concerns outlined above.

While quadratic voting does help lessen the influence large stakeholders have over swaying the result of a governance vote, we believe creating a mechanism where a validator’s governance vote power is a direct function of their validator bond ratio might better align validators and their delegates while also incentivizing node operators to put up more collateral. The problem with quadratic voting is that it assumes all Vote Power is the same. Going back to the example with the Vote Power Tax, if you assume Dokia Capital and AllNodes have the same total Vote Power, quadratic voting would give them the same weight in governance while not taking into account that a larger share of Dokia’s Vote Power is owned by them.

Below is our proposed formula for calculating validator governance vote power moving forward. Please note that a validator only gets the vote power of the delegates that choose not to vote individually. We propose the same mechanism, but instead of getting 1:1 vote power of delegates that don’t vote, validators should only get a proportion of that vote power since they do not own those assets.

Where:

  1. VPi=Total Vote Power of a given validator
  2. VPsb=Vote Power Self Bonded by a given validator
  3. 40% is the minimum quorum of the Cosmos Hub today

What this means is that for a validator to receive more than 40% of the delegated stake vote power that has not voted, they would need to bond over 6.4% (40%^3) of total delegated stake to their validator.

As an example, assume Validator A has total Vote Power (VPi) of 10 ATOM, of which 2 ATOM are self-bonded (VPsb) and none of their delegates voted:

Today, Validator A would get 10 ATOM of total Vote Power.

Under this new mechanism, Validator A would get 58.4% of their delegate Vote Power ( 20%^(1/3)) plus the 2 ATOMs they own, giving them a Total Governance Vote Power of ~6.67 ATOM (58.4%*(8 ATOM) + 2 ATOM).

We believe this puts more governance weight into direct voters and further incentive aligns validators with their delegates since they are putting up more of their own slashable collateral to secure the network.

While governance was not directly part of our scope, we think this idea (or a version of it) should be considered by the community. The validator bond mechanism imposed by the LSM opens the design space for how to solve the Principal-Agent problem with dPoS in all facets while ensuring the system is not too capital inefficient (forcing validators to put up too much collateral).

Dynamic Community Pool Tax

The Cosmos Hub is overpaying for security. High inflation is NOT a pre-requisite to a large percentage of the supply staked, as shown by other top PoS networks in the space.

The Cosmos Hub has more than enough issuance to pay for security AND fund strategic initiatives.

To properly bolster the community pool and fund initiatives like the ATOM Alignment Treasury, core development work, subDAOs, etc, Blockworks Research is recommending changing the community pool tax from a static percent of inflation to a known schedule that decreases block-by-block until a base tax of 5% is reached over time as inflation is reduced accordingly based on our proposed supply schedule here.

We propose the Community Pool tax be increased from 10% to 20% with a 20% annual decrease that decreases block-by-block until a base tax of 5% is reached.

Below we compare the current community pool tax (blue) to the recommended tax (red).

With this tax schedule, ATOM stakers can expect to see a lower community pool tax than they face today by Q2 2027.

Additionally, when applying this new tax schedule with our recommended supply schedule, it frontloads the community pool by about 18 months as seen below.

As the Cosmos Hub’s revenue grows through the economic relationships its fostering with its consumer chains, the need for as high a community pool tax reduces. Ultimately, to pull forward the community pool tax, which we believe will be critically important to properly right size the Cosmos Hub’s treasury to jumpstart the ATOM Economic Zone, fund core development, and subDAOs, we believe it is the fiscally responsible thing to do to lower the tax on the backend.

Fiscal Policy (End State)

Below is an overly simplistic view of what the fiscal policy and budget distribution of the Cosmos Hub could look like depending on what the community approves if/when these ideas move to an onchain governance vote.

If you recall in our post here, this is what the current state of the Cosmos Hub’s fiscal policy/flow of funds is:

We believe that by bucketing the community pool tax into smaller buckets (with the community deciding how much is directed towards each bucket), it should allow for greater accountability and better budgeting of the Cosmos Hub’s capital. This will allow the Cosmos Hub to be a true Port City that strategically allocates its capital to welcome the brightest development teams in the ecosystem that want to align and build on behalf of the Hub, foster economic growth of its consumer chains and the wider Interchain through the use of Protocol-Owned-Liquidity, and funds sub-communities (subDAOs) with specific mandates to help better the wider Hub community.

While what we presented above is not final, we hope this post furthers the discussion about the future of how the Cosmos Hub (and its community) operates in the future.

We will have one final post next week that directly links to all 3 separate forum posts and will also include a customizable model we created to allow the community to ultimately decide the parameters it wants to select if it does decide to move forward with any of the ideas proposed by Blockworks Research.

23 Likes

Huge but what is the process discussion? then draft proposal in the forum then vote?

How long does it take? I know it will depend on vote and what direction hub want to take but we need to have a roadmap or date for tokenomics change.

1 Like

How do you plan to deal with sybils which will try to undermine their fair share of tax?
Ex - Binance Node and Binance Staking
stake.fish and grant.fish

1 Like

The extent of this research is staggering. We will take the necessary time to read it carefully and propose feedback with as much depth as you putted into this post. We would just note that the quality of discussions in the Hub have improved significantly over the last few months and thank you for being part of the persons who greatly contributed to this.

4 Likes

can validators liquid stake their self-bonded stake?

The VP tax gives CEXs an advantage/exemption because they custody the majority of their delegations and can simply self-bond their customers staked assets.

Why are you recreating the same corrupt financial systems that the cosmos is intended to free us from? If I wanted to pay > 20% tax to support a welfare system I would just stay in the Dollar.

Firstly, I’d like to start off by applauding the effort that was put into this; this is a great piece of work and I do hope that this results in spearheading proposals to implement these changes. In a separate manner, the mechanisms proposed above will still add substantial value, however (and by way of a precusor), I am highly in favour of implementing this holistically as there are a lot of underlying benefits that would emerge therefrom.

dPoS Pitfalls re. Centralisation

All in all, the underlying mechanics that are implemented so as to cater for the principle-agent issue (which is, as stated, a susbstantial centralising factor in all dPoS blockchains), must strike a balance between the following principles of decentralisation (which is a very wide spectrum):

  • Incentivising users to disburse delegations to varied validators so as to i) Incentivise new validators to participate in consensus and ii) reduce reliance (or rather, over-reliance) on limited few validator operators (thus contributing to the progressive decentralisation of the Cosmos Hub (disbursement, new entrants, more validators in consensus = progressive decentralisation of the stack’s consensus infrastructure)

  • Ensuring the adequate safeguarding of delegators’ stake by indirectly ensuring validator accountability through enforceable self-bond mechanisms so as to i) incentivise users to delegate to validators and increase the chain’s security (less malicious validator events = an indirect incentive to delegate due to peace of mind), and ii) ensure that validators have a vested interest in ensuring smooth operations due to them having ‘skin in the game’ as stated above. Similarly to the first point, less malicious validators → Chain has a good reputation → delegators are indirectly incentivised to purchase ATOM and delegate to Hub’s validators due to peace of mind → more delegations, more security → more active delegators, validators see a business opportunity and want to run validators → more validators = progressive decentralisation.

Comments on the proposed solutions

I believe solutions posed are very beneficial in attempting to achieve delegator disbursement among a wider range of validators and thus changing the current status quo of prospective delegators only considering the top few validators as suitable option for their ATOM. As highlighted this will also potentially (probably imo) lead to validators having more skin in the game and thus, overall, this will be a great value-add in progressively decentralising the CosmosHub (it serves to contribute to the principles of decentralisation enlisted above). In addition, the cubic delegation model will also potentially lead to further validators having skin in the game and thus, all in all, they are net positive for the Cosmos Hub.

Concern

My only concern here is whether this can be gamed (so to speak) by validators and actually result in fewer delegators entering the ecosystem and thus, doing the opposite of progressively decentralising Cosmos Hub. Let me elaborate a bit; Were the proposed solutions to be implemented (as I stated, I am in favour of the solutions being implemented holistically by the way), could this potentially lead to validators reaching a collective decision to reduce reward rate distribution to delegators and maintain a larger % of revenues to themselves? Theoretically, validators would do this so that they can finance their ‘skin in the game’ in this manner via rewards they accrue.

Naturally, if validators emit less rewards to delegators and keep the larger % of rewards for themselves so as to self-bond in an increasing manner over time and get the benefits linked to having a higher self-bond it could lead to:

  • Less users purchasing and subsequently delegating ATOM due to the lower reward rates;

  • As a consequence of the above, decreasing the security of the Cosmos Hub;

  • As a consequence of the above, less (specifically, less delegators), having a decreased number of validators running on Cosmos Hub due to the decrease in delegators;

  • As a consequence of the above, actually decreasing the overall decentralisation of the Hub.

Conclusion

All in all, a great read and well done to the team for coming up with this new policy, I hope to see it being discussed avidly by the community and hopefully, we can look forward to it being implemented so as to see its effects on the Hub!

Great job guys!

Kind regards,
IL

2 Likes

I guess because like with every other system, those who maintain, work for and bring growth and innovation to it have to be paid.

But unlike the US welfare system, tax money spent on the Cosmos Hub system should directly benefit every $ATOM holder, rather than being wasted on unproductive expenses like funding military operations, secret agencies and feeding a corrupt and obsolete boreoarctic apparatus.

In crypto, we the plebs, actually have a say on what happens with the taxpayers money and this proposal even strengthens this point by taking away voting power from the top of the validator set and distributes it more evenly thorough the active set. The same is true with validator earnings.

Thus, to me it makes sense to initially increase the community pool tax to 20%, in order to bootstrap the growth of the AEZ, incentive align dev teams working on and for the Hub, and to fund novel primitives like the Atom Alignment Treasury, the DRIP Module and various SubDAOs.

However, I agree, there should be checks and balances in place, so that ATOM holders can at any time revoke or amend the funding amount that goes out to certain parties, if they are deemed untrustworthy or don’t comply with their mandate.

5 Likes

thank you so much for saying this, it’s why I’m here

1 Like

First of, thanks for all the thought and work put into these posts, great to see the Blockworks team getting involved in Cosmos governance like this!

I found the liquid staking tax an intriguing deterrant to concerns about overreliance on liquid staking to consider but do also feel that the issue with these sort of models generally lies in workarounds that are available for custodial actors and larger operators.

Especially the concern around custodial actors/CEXes being able to circumvent liquid staking limitations and via self-staking customer tokens additionally allowing them to earn higher yield than independent validators with the Vote Power Tax might have the opposite effect of what the goal of this mechanism is. It might be that centralized players will be legally constraint to do something like this (seems likely at least in the US given the actions against centralized staking products recently), but other jurisdictions might not impose such regulations.

With respect to the validator self-bond, one thing to think about is the potential for operators to borrow ATOM to self-stake to maximize their capacity. A financially savvy and well-connected operator will be able to structure deals/partnerships with larger token holders and circumvent these restrictions.

Generally speaking, I think this is a great research direction and discussion to have. In the end it might be a desirable tradeoff for ATOM/Cosmos Hub to implement some of these concepts, if the community is aligned on that the benefits outweigh the downsides, which can be hard to measure/tell in advance.

I think one feature that hasn’t been discussed in while on the Cosmos Hub (to my knowledge) is quadratic slashing, i.e. the idea of increasing slashing penalties for correlated faults. To me this feature always seems to nicely incentivize decentralization with less potential for workarounds (vs e.g. the Vote Power Tax and/or self-bond ratios discussed in this post).

1 Like

This is an interesting idea and I hope it will be implemented, thank you, however regarding the other idea for the community pool tax, it was already increased from 2% to 10%, I think increasing it to 20% is excessive. But why would the top validators vote yes to approve this Vote power tax proposal? If the proposal is approved it is unlikely that delegators of the larger validators would be willing to pay this new tax, which would be like a higher commission. They would redelegate quickly to smaller validators, so the ~900k ATOM in tax revenue in year 1 mentioned, assumes no delegator would redelegate and they will pay this tax, which is relatively unlikely. Also, according to the formula, in the edge case that all validators have the same voting power there wouldn’t be any vote power tax? Overall great idea, similar ideas have been discussed over the years but nothing implemented so far, this idea is basically to incentivize delegators to increase decentralization, where the lowest overall tax would be paid when there is an equal stake distribution across all validators. Again, this idea is great and also that this new tax revenue will go to validators to support with consumer chain costs. The issue is, for this proposal to pass the largest validators would need to vote yes, and although increasing decentralization is positive, they might not be willing to vote yes and risk losing voting power because of this new tax so further discussion is required here.
Also, about the self-staked ATOM in the equation, it would be more complex because some validators might be self-staking from another account to their validator so you wouldn’t be able to see so clearly the self-staked amount of each validator, each would need to provide this information and there should be some methodology to verify and confirm this.

The US welfare system uses this exact same argument to justify it’s corrupt and obsolete bureaucratic apparatus.

Is this a joke? are you saying that you actually believe that validators voting to raise the minimum commission or raise CP tax was based on the will of the users? madness.

lol, very surprising that validators would be in favor of raising the amount of capital they can take from users to kickback to themselves and promise users future profit and growth trickling back to them through the efforts of the AAT, DRIP, and various corrupt and obsolete bureaucratic apparatus.

validators have gone unchecked for to long, their conflict of interest within governance has become obvious and the ongoing corruption repels outside investment from brining capital onto the chain.

This research seems to align with principles reminiscent of wealth redistribution, a concept often associated with communism, where resources are taken from the affluent and distributed to the poor. The concern is that this approach discourage high-performing top validators and encourage smaller ones, some of which might not offer the same level of service quality.

Furthermore, this incentivization might not motivate smaller validators to improve their performance, as they would receive more stake regardless. Meanwhile, larger validators might become disheartened due to reduced profits and consider redirecting their efforts to more democratic projects.

As rightly mentioned, custody validators would have an advantage, as this scheme is unlikely to significantly affect their profitability, and they can self-stake a substantial portion of their holdings, unlike others.
While self-bonding dependence may have its merits, the proposed proportion appears impractical.

Regarding increasing CP tax, perhaps a more prudent initial step would be to judiciously manage the existing treasury and ensure that expenditures are both equitable and beneficial. What’s the point of funding CP without a proven mechanism for its proper utilization?

In general it’s not a good idea to set Hub on these rails amidst of the bear market to cause even bigger frustration for stakers with less rewards (min -3% from current) & validators struggling to keep afloat. Such tax could work well when set at the launch when all nodes are barely in the same state & could accumulate self stake gradually.

The potential consequences of implementing it now? The Hub, delegators, and validators all suffer losses with no clear benefits.

1 Like

You seem to be associating large validator=top performance/uptime, this is totally not the case. Look at the data, some of the biggest validators have terrible performance, missing a lot of blocks and always late for upgrades, not to mention zero participation in governance or here in the forum. Many small validators have very high performance, for example we are a small validator and currently only Polychain missed fewer blocks than us in the last 3 months, 21 missed by us versus 18 by Polychain.

Again, you are assuming big validator=top performance, and because of this top performance they got large delegation. This is absolutely incorrect, as mentioned above many small validators have much better performance than many big validators, and certainly the reason delegators choose a validator is clearly not because of performance or governance participation.

Custodian validators is not that they ‘can self-stake’ a substantial amount, they ‘self-stake’ all, or did you see Kraken, Binance or others staking some of their custody tokens to other validators? Moreover, there are disadvantages also to custodian validators, for example, there is a FINMA discussion in Switzerland about custodian validators potentially needing a banking licence, in contrast non-custodial validators, since they don’t custody any tokens, are not affect by these regulations in Switzerland or other jurisdictions

I disagree, the potential consequences of the vote power tax implemented are increasing the decentralization of the Cosmos Hub which is very positive especially now that replicated security is offered to consumer chains and providing the funds to smaller validators to run many consumer chains and scale the AEZ model, which is also very positive.

If delegators discover they get rewards in range 10-15% apr instead of 19% added to the local ATL atom price, they barely will not be happy. Who will care about decentralization when the chain will lose its audience & thus market cap?
Again, it’s not the right time for such changes during the bear. And even when the market eventually turns green, it’s crucial to carefully consider all elements to achieve a fair balance among participants, rather than creating a situation where one group benefits at the expense of another, making centralization just change the mask.

It seems you are not understanding how the Vote Power Tax works. This tax is for the largest validators as described in the equation in the proposal. When active delegators of the largest validators become aware about this new tax and the lower APR in the largest validators, they will redelegate to smaller validators for higher APR, and this dynamics over time will lead to a much better decentralization in the Cosmos Hub. The delegators who are not very active will be paying this tax until they are aware and then they can choose to support decentralization and earn higher APR, or keep delegating to large validators and pay the tax for this.

1 Like

Governance is ATOM’s unique feature, quadratic voting protects minority rights and prevents whale manipulation.

Anyway LP tax 20% would affect the whole set, and lower apr by 1,9%.
Bottom validators under the median, not all of them of course, but apparantly a substantial part, have a small public presence and poor experience in cosmos. Quality of such might be questionable when the whole entity consists of two people literally, with no backups, without cosmovisor, while replicated security adds comlexity to the maintenance process.
And delegators will have a choice between proven quality & higher yields with higher slash risk. Such experiments could be trialed in better market conditions with gradual raising the tax to have a chance to assess the short term effect & make minor corrections if needed. In the current version the difference between top & bottom is extremely high.

I mentioned before that we don’t support this further increase of the CP tax from 10% to 20%, why are you mentioning this if what I’m discussing with you is the vote power tax, not the CP tax increase?

The median of 0.17% VP corresponds to the rank ~88/180. There are validators above the median and below the median with great performance and adding a lof of value to the Cosmos Hub. You are suggesting that most validators above the median are great, and most validators below the median are poor quality, this not true. What do you mean by ‘poor experience in Cosmos’? If you take objective data such as uptime, governance participation and more as the definition of ‘proven quality’, you will see that your statement ‘most validators above the median have proven quality, and most validators below the median don’t have proven quality’ is incorrect. Your argument for not implementing the vote power tax is ‘the biggest validators are high quality, the smaller validators are bad quality’ which is not true.

Delegators will have a choice between:
-Maintain the centralization of the Cosmos Hub by delegating to the largest validators and pay a tax for this
-Support decentralization by delegating to smaller validators and don’t pay the tax, and no higher slash risk since they can choose the small validators with great performance

somehow slash events were more frequent for smaller ones. And I’m not saying that all top are good & all bottom are bad, but main trend takes place.

Some of stakers don’t even realize aspects of validators, apr and centralization. They may discover their earnings suddenly fell by 30% in pair of poor price and just leave.