Possible Optimizations of NNS Tokenomics (Updated!)

[dominicwilliams]

Hello everyone. I want to propose two changes we could make to NNS governance with the aim of optimizing network tokenomics.

These two changes are not slam dunks, but they could provide some significant advantages, and our purpose should be to continue improving tokenomics, in the same way that we constantly make the network faster, more efficient and more capable as we move into the future.

The ideas are standalone, so we could take both, one, or neither!

IMPORTANT/UPDATE: Forum discussions on May 6th updated Proposal 1 below. Some people are not reading through the replies thread! The proposal was modified so that 8 Year neurons would NOT be forcibly reduced into 5 Year neurons (and in fact, would have some adjustability too). In the modified proposal, it would only become impossible to set dissolve delays >5Y in the FUTURE, and delays on existing neurons would only be brought down if their owners wished (at their choice). Please make sure you read all replies and the state of play before assuming. As soon as I’ve got time, I will create proper forum proposals that split 1 & 2 below and which contain all updates and then fresh discussions can begin from there. This will take a week or so due to other commitments. What’s fantastic is that everybody cares so much and we can have these frank discussions about ideas as a community. Big thank you!

PROPOSAL 1: Reducing neuron dissolve delays to reduce inflation

General context: ICP has fought its way through significant sell pressure pretty well. Future step downs in sell pressure are also coming. Our belief is that major VCs who bought in early will be exhausted this year, which will put the network in an enviable position – the other major networks remain mostly held by such parties behind the scenes, often in the range of 75-85% (yes, you read those insane numbers correctly!). ICP holders are already vastly more decentralized, and the more so it becomes, the better.

Nonetheless, finding ways to allow the network to breathe still make a lot of sense, and one of the ways we can address that is by reducing inflation. Today, the primary source of protocol inflation is governance staking. The question is, can this be reduced by updating the protocol, without breaking promises the network has made to those who have staked. The answer is, yes, potentially:

Outline of proposal:

1. Reduce the dissolve delays of all neurons by 5/8th (to 62.5% of what they is today)

2. The age property of neurons would stay the same (age increases over time, benefiting voting power, and is set to zero whenever a neuron is placed into dissolve mode).

3. The minimum dissolve delay at which a neuron can gain maturity by voting, would be reduced to a round 3 months.

4. The overall notional voting reward distributed by the network, which is logically distributed by votes to determine how maturity increases, would be reduced, such that the rate at which a neuron’s maturity grows, given some amount of locked ICP, dissolve delay and age, would remain almost identical to that today.

The main benefit to tokenomics is that this would reduce the amount of new maturity that neurons gain by just under 20%, which is definitely significant (the immediate ICP impact would be less, since many do not use their neuron maturity to create new ICP, but especially over the long-term, this would definitely have an effect).

This would mean that if you have 8 year neurons today, and this change were made, then you would find you now had 5 year neurons, and the maturity would grow at the same rate as for 5 year neurons today (given the stake and age).

Those with 6 month neurons (currently the threshold for gaining maturity by voting), would then have 3.75 year neurons, but they would still gain maturity, since the minimum threshold for gaining maturity would be reduced to 3 months – which will also be useful for those wishing to try out governance staking for the first time.

5 years is still a very considerable time to make a commitment, but it’s not as mind-bending as 8 years, which makes the system more friendly.

But what about the hardcore “8 Year Gang”!! Would this have to become the “5 Year Gang” since they have now have 5 year neurons!!!?

My suggestion is that if you identity as “8 Year Gang” (and all mine are 8 years) then you would continue to identify as before and the moniker says the same.

For example, Porsche is going electric, but the fastest Taycan is still the “Taycan Turbo S” even though electric cars don’t need turbos. The “8 Year Gang” can continue in the same way. It’s really about commitment to the future.

PROPOSAL 2: Swapping locked ICP for unstaked maturity.

You should always get your own tax advice, but, by design maturity is an attribute of a neuron, not a cryptocurrency that lives on a ledger like ICP that can be transferred and sold. Furthermore, the amount of newly minted ICP that a neuron can produce by applying its maturity is indeterminate, and subject to modulation by external factors like price trends, and even updates to the NNS, further indicating that maturity is part of the value of a neuron and has indeterminate value.

This design means that when a neuron gains maturity, this is not realized income (again get your own advice). However, it is hard to argue that when the owner of a neuron uses its maturity to produce new ICP by spawning, then this represents the generation of income, since the newly generated ICP can be transferred and sold, and has a more-or-less determinate value, and in most jurisdictions (unless you live in Monaco, UAE, Puerto Rico, etc) income tax will be due on the new cryptocurrency you have produced.

The maturity of voting neurons increases near continuously, and its nature arguably provides administrative advantages compared to traditional staking, which regularly pays out cryptocurrency, forcing constant accounting. Nonetheless, there are some disadvantages.

Perhaps the largest one is that maturity is an attribute of a neuron, and not a cryptocurrency! This means that it can’t be used in things like DeFi. There is a lot of maturity sitting on neurons that might usefully be deployed if it were unlocked ICP! We might try to fix this.

Outline of proposal: When a neuron has unstaked maturity, then the owner can swap that unstaked maturity for ICP that is locked inside i.e. such that the unstaked maturity becomes locked inside, and the ICP is freed (currently maturity can already be staked, so this is not a big leap).

Example benefits:

1. Unlocking your pre-existing ICP is tax neutral, so the owner could, for example, transfer the newly unlocked pre-existing ICP into a DeFi service and borrow against it. This might unlock more liquidity for DeFi.

2. If a neuron owner needs to realize x fiat gains for a financial purpose, they could sell their newly unlocked pre-existing ICP, paying capital gains, rather than applying maturity to generate new ICP for sale, which is subject to income tax. Since in e.g. in places like the UK, capital gains is 20%, while income tax 50%, this would make it possible to sell less ICP to realize the required fiat gains. This might result in less forced sales.

3. If a neuron owner wishes to have unlocked ICP to swap for some other token, and would otherwise have to apply unlocked neuron maturity to produce new ICP, then benefits as per advantage (2) above.

These two proposals are quite nuanced, and interested to hear thoughts.

[Mico]

In proposal 1 are you saying that for people who have staked for 8 years would endure a 20% hit on the voting rewards they are getting currently?

[Accumulating.icp]

I can’t help but feel like there are better ways to “reduce inflation” than penalizing those who have staked for extended periods of time.

@Accumulating.icp

Lately I’ve been considering various factors that contribute to the sustainability of the supply of $ICP - one of which is staking, and in turn the yield correlated to it.

The current APY Scale can be found below;

6 Months - 8.9%

1 Year - 9.4%

2 Years - 10.5%

3 Years - 11.5%

4 Years - 12.6%

5 Years - 13.6%

6 Years - 14.7%

7 Years - 15.7%

8 Years - 16.7%

Information provided can be cross referenced at:

Governance - ICP Dashboard

This depicts that 6 Month Neurons are incentivized at 53% of the rate of 8 Year Neurons (8.9%/16.7%), while having the staking obligation of 6.25% of an 8 Year Neuron (6 months/96 months).

Additionally, a pattern that can be noticed within the staking reward scale, is that each additional year of staked time, results in about a 1% increase in rewards - however, the base layer of rewards, is 8.9% for only a 6 month timeframe.

For example, the APY scale clearly is not proportionate to begin with, so maybe that’s a better place to start than making such a drastic change that will affect the staking parameters of 56% of staked ICP.

I also have to question, how many voting rewards have to be taken away for inflation to be “acceptable”? This is not the first time we’ve gone in to change the reward scheme for the purpose of altering inflation.

Have you considered exploring the other face of the coin; the deflationary mechanics of the network?

[dominicwilliams]

People who have staked for 8 years, would see their maturity grow more slowly/get fewer rewards. However, if this scheme were adopted, they would only be locked for 5 years, and they’d been getting the same rewards that people locked for 5 years get today. In one sense this is a bonus for them: they rode the wave as 8 year neurons, getting those higher rewards, and now they are only locked for 5 years.

[Mico]

Would the current 8 year stakers get any sort of bonus to offset this reduction in rewards?
Such as the “genesis neuron” bonus?

Or will this be just a straight reduction across the board?

[Accumulating.icp]

@dominicwilliams

Our belief is that major VCs who bought in early will be exhausted this year, which will put the network in an enviable position

Please don’t take this as me voiding my previous comments, but why is this relevant to the lowering of APYs? If you believe they’re contributing to sell pressure, wouldn’t it be more logical to lower APYs while they have stakes active, rather than after they’ve exited?

[Accumulating.icp]

I’m also curious if you think the continuous intentional “avoidance” of taxable events will lead to longterm legal problems.

@Accumulating.icp

The Evasion of Taxable Events

While it may not have been intended, this mechanism directly evades taxable events.

This is due to the fact that the Neuron Fund can only draw from liquid Maturity, that could be spawned, but instead is utilized under the guise of “Maturity” - which is conveniently not tracked within either Blockchain Explorer (ICScan or IC Dashboard), or the NNS itself.

The lack of transparency presents two issues - being the inability gauge how much maturity was used in the sale, and how much is received from off-loading the position. Which in turn leaves all maturity to fall under the umbrella of a singular taxable event, being the spawning of staking rewards.

Utilizing the USA as the example jurisdiction (however, many jurisdictions apply similar laws), given the abundance of open court cases against DFINITY within the region, the following taxable events are directly evaded;

1. Spawning of ICP Rewards
2. Conversion of $ICP to Token
3. Spawning of rewards from SNS Tokens
4. Conversion from Token to ICP

I’d also be willing to make the arguement that Maturity Modulation only exists to avoid Switzerlands “Wealth Tax”, but that is another topic.

@dominicwilliams

PROPOSAL 2: Swapping locked ICP for unstaked maturity.

You should always get your own tax advice, but, by design maturity is an attribute of a neuron, not a cryptocurrency that lives on a ledger like ICP that can be transferred and sold. Furthermore, the amount of newly minted ICP that a neuron can produce by applying its maturity is indeterminate, and subject to modulation by external factors like price trends, and even updates to the NNS, further indicating that maturity is part of the value of a neuron and has indeterminate value.

This design means that when a neuron gains maturity, this is not realized income (again get your own advice). However, it is hard to argue that when the owner of a neuron uses its maturity to produce new ICP by spawning, then this represents the generation of income, since the newly generated ICP can be transferred and sold, and has a more-or-less determinate value, and in most jurisdictions (unless you live in Monaco, UAE, Puerto Rico, etc) income tax will be due on the new cryptocurrency you have produced.

The maturity of voting neurons increases near continuously, and its nature arguably provides administrative advantages compared to traditional staking, which regularly pays out cryptocurrency, forcing constant accounting. Nonetheless, there are some disadvantages.

Perhaps the largest one is that maturity is an attribute of a neuron, and not a cryptocurrency! This means that it can’t be used in things like DeFi. There is a lot of maturity sitting on neurons that might usefully be deployed if it were unlocked ICP! We might try to fix this.

Outline of proposal: When a neuron has unstaked maturity, then the owner can swap that unstaked maturity for ICP that is locked inside i.e. such that the unstaked maturity becomes locked inside, and the ICP is freed (currently maturity can already be staked, so this is not a big leap).

Example benefits:

1. Unlocking your pre-existing ICP is tax neutral, so the owner could, for example, transfer the newly unlocked pre-existing ICP into a DeFi service and borrow against it. This might unlock more liquidity for DeFi.
2. If a neuron owner needs to realize x fiat gains for a financial purpose, they could sell their newly unlocked pre-existing ICP, paying capital gains, rather than applying maturity to generate new ICP for sale, which is subject to income tax. Since in e.g. in places like the UK, capital gains is 20%, while income tax 50%, this would make it possible to sell less ICP to realize the required fiat gains. This might result in less forced sales.
3. If a neuron owner wishes to have unlocked ICP to swap for some other token, and would otherwise have to apply unlocked neuron maturity to produce new ICP, then benefits as per advantage (2) above.

[dominicwilliams]

Firstly about deflationary mechanics, that’s down to compute. The more compute that’s happening, the more ICP is being converted to cycles and burned. That’s why it’s exciting to see web3 services seeking ways to achieve hardcore mass adoption, such as Yral, and compute intensive applications, like AI, on the network. The more cycles being burned the better.

Secondly, about the fairness of voting rewards, the voting power of a neuron is neuron_stake * dissolve_delay_bonus * age_bonus, and your “share” maturity gains are proportional to your voting power/votes performed (as per Staking and voting rewards | Internet Computer).

Currently dissolve_delay_bonus starts at 1 for 6 months, and ends at 2 for 8 years, so 6 months gets 1, and 8 years gets 2, i.e. 8 years only gets double what 6 months gets. Personally I would agree that this should be radically adjusted.

What I would prefer: if we updated so 3 months minimum, and 5 years max, then probably 5 years should be receiving 2.5X+ what 3 months gets.

The challenge is how to change that. Somebody who is currently locked at 1 year would see their rewards change downwards, and might scream the change is unfair since the dashboard told them 1 year gives you something else – albeit the guarantees provided by the NNS are weaker than “you have an ICP on the ledger,” which is of the immutable cryptocurrency kind.

Here’s what I would suggest might be done: adjust the way the dissolve_delay_bonus works for new staking, but not for neurons that are already staked. So if there’s a 2 year neuron, it continues to have the same voting power before (unless it was Seed/ECT and is in the process of dissolving, since this cohort didn’t stake on the basis they would get some specific voting power, and no promise is being broken).

New staking would work according to the new scheme however.

[Accumulating.icp]

@dominicwilliams

Firstly about deflationary mechanics, that’s down to compute. The more compute that’s happening, the more ICP is being converted to cycles and burned. That’s why it’s exciting to see web3 services seeking ways to achieve hardcore mass adoption, such as Yral, and compute intensive applications, like AI, on the network. The more cycles being burned the better.

I am aware, but like staking, there are parameters that can be adjusted to offset inflation, if it is a concern. This in turn makes the Internet Computer more expensive to use, but this isn’t unprecedented; inefficiencies have been defined & corrected in the past.

For example, you can refer to Node Provider payouts in relation to Cycles burnt, and it’s easy to determine where inflation runs rampant.

@Accumulating.icp

TLDR:

- Nodes on the Internet Computer operate at a 100x+ inefficiency, with self imposed limitations of 300GB while having the capacity to store 30+ TB

-The Internet Computer burns .84% of what is rewarded to Node Providers on a monthly basis via Cycles Burnt

-Of the 36 Subnets, with a total storage capacity of 10.8 TiB, only 3.55 TiB is utilized, representing an inefficiency of 70%+

On September 11th 2023, I published the article “Node Provider Inflation Spiral” within $ICP forums, of which, can be referenced below:

Node Provider Inflation Spiral Please note that aside from raw data, this article is purely speculative with the intention of provoking thought & conversation. This article covers hypothetical scenarios at predetermined price points extrapolating from current metrics to observe potential futures. This article is not financial advice, nor should the scenarios within be considered as fact. This stems from a conversation with a fellow ecosystem member regarding how Node Providers are rewarded for…

This conversation sparked great interest, earning acknowledgment from members of the DFINITY Foundation as a valid concern that should be addressed. Subsequently, this sparked the interest of multiple Node Providers, most notably, @DavidM & @dfisher ,
who voiced multiple valid points & concerns, from not only the perspective of a Node Provider, but from the perspective of an individual attempting to safeguard the Internet Computer Network.

With this being said, a perspective recurringly raised by vocal node providers, is the importance of not correlating a subnet & its nodes’ reward rate, to total subnet useage.

This concern was raised due to the fact that Node Providers can not pick the subnet they’re in, and therefor are consequenced with the potential of contributing to an underused subnet - earning lesser rewards in contrast to the previous system.

This again, sparked my interest:

if Node Providers can not contribute to the network at scale without fear of contributing to an unused subnet, is the network over burdened with Nodes, or are Cycles not being burnt proportional to the true cost of running the network?

The first question to address is whether unused subnets should be compensated - taking us down a difficult road to navigate.

On one hand, if a Subnet is not being used, it is not contributing to the cycle burnt rate, and therefor only has the potential to increase inflation under the current reward scheme.

On the other hand, we have to consider why this inflation trade off is made. The Internet Computer is completely reliant on Subnets of Nodes to scale - if left without room to grow, there will come a time in which the network has to scramble to provide Nodes in time for dApps, which can take months, given Nodes are hosted in Data Centers.

As such, the line of determination in regards to what Nodes should and should not be compensated is cloudier than presumed.

However, something that the Internet Computer Network as a whole has agreed on, is that Node Providers providing degraded, or lesser services, are deserving of reward slashings. I believe a similar system could be translated to unused Nodes & Subnets, although that’s not what we’re here to discuss.

From here, we can begin to determine the disparity in total Node Provider Reward Distributions & Cycles Burnt, in contrast to the network state overtime.

Referencing the ICP Burn Chart, it can be noted that the Internet Computer has a cumulative burn rate of 136822 ICP - of which, 136,215 ICP come from transaction burns (presumably burning conversion transactions), while 607 ICP comes from transaction fees.

Next, we can reference the Cycle Burn Chart for a more accurate gauge of the cumulative monthly burn of ICP over the last 3 months:

The cycle burn chart indicates that on average, 5.1B Cycles were burnt each second over the last 3 months, which equates to 440.64T Cycles burnt a day.

For simplicities sake, we’ll convert this to an ICP amount before extrapolating to a monthly basis.

This can be done by first dividing the daily cycle burn rate by the SDR exchange rate to determine the daily fiat expenditure of the network.

440.64T Cycles / 1T = 440.64 SDR (• 1.31 = 581.17 USD$)

This can then be extrapolated to an ICP amount by utilizing a 90 day average of the ICP token price, of 3.5$.

581.17 USD$ / 3.5$ = 166.0485 ICP

Now that we have determined the network burns 166 ICP / day across a 3 month average, we can extrapolate this to estimate a monthly network burn of roughly 4957.45 ICP.

In contrast, the previous article “Node Provider Inflation Spiral” depicted that the Internet Computer Protocol minted 500k+ (on the lower end) ICP each month, during this timeframe, reaching an all time high distribution of nearly ~600k ICP tokens last month.

Utilizing Augusts Node Provider Reward Distribution data, we can determine that of the 583,577 ICP minted to compensate Node Providers, only .84% of it was correspondingly burnt via Cycles, showcasing a disparity of 99.16% between Node Provider payouts & ICP burnt via cycles.

Turning our attention back to Network State, it can be observed within the following article that Gen1 Node Machines have a storage capacity of 28.8 TB & 30.72 TB

How To Become An Internet Computer Node Operator: Beginner's Guide

Alternatively, Gen2 Node Machines have a storage capacity of 32 TB.

DRAFT PROPOSAL Below is an example of a proposal we (the DFINITY R&D team) intend to submit to the NNS in a few days/weeks, depending on community feedback as well as questions and answers. We want to hear what you all think, including any wording changes we need to make Summary TLDR: As part of the work to further decentralize the infrastructure layer, we want to submit a new motion proposal to introduce a new type of hardware spec (and its corresponding remuneration) for nodes on the IC. W…

In contrast to their abundant storage potential, according to public documentation & official forums, subnets are seemingly limited to a capacity of 300GB.

With this information laid out, we can continue to the Network State:

As of present, the current Network State is 3.55 TiB of a total 10.8 TiB, across 36 subnets - which equates to roughly .098 TiB per Subnet (the math is not so simple, but this is useful to depict network load), or roughly 30% of each Subnets maximum capacity in correspondence to current self imposed limitations.

This goes to show that the issue is not so simple as “are we over paying or over onboarding Node Providers?” - it’s a combination of both, with Cycles burning less than 1% of what Node Providers are compensated monthly, while Nodes are operating at seemingly a 70% inefficiency.

@dominicwilliams

The challenge is how to change that. Somebody who is currently locked at 1 year would see their rewards change downwards, and might scream the change is unfair since the dashboard told them 1 year gives you something else

I completely agree, and that goes both ways here - when we talk about reducing stake parameters by 3 years & an APY reduction of -20%, we run into the issues of the dashboard promising x, while the NNS is consistently making changes to effect this.

I feel like the solution is much more nuanced than simply reducing stake timeframe & APY in an effort to curb inflation.

[dominicwilliams]

“Legacy” VCs who got into the ecosystem years ago don’t stake ICP. They cycle assets to pay out to LPs of their earlier funds. So I don’t think maturity calculations will make any difference to their behaviors (if that’s what you meant). Most of them are underwater in their 2021-2 investments (think of the huge sales that NEAR Protocol made, for example), and cycling profitable 2018 token holdings to demonstrate to their LPs they are capable of making returns, so they can raise new funds, makes sense for them. That’s just the way it is.

I mentioned them only because 1) it’s important to realize that there have been significant sources of sell pressure outside of inflation, so no point getting overly obsessed by inflation when thinking about tokenomics, but 2) anyway, it is also true that ICP has had to fight through sell pressure, and as it dries up, why not press home the advantage by making tweaks to tighten up tokenomics.

[jamesbeadle]

Literally based an entire business on staking my SNS DAO’s ICP treasury, so no, please don’t give me a 20% pay cut.

[System1]

Please, people have made long term plans with the staking rewards and I’m one of them. Whatever you do, don’t alter the tokenomics and reward. We knew the risks and we went for 8year non desolving , please don’t alter it if it will change the rewards. Thanks

[Mico]

I would vote to approve this proposal if it only altered new staking neurons and gave the current 8yr neurons a vp bonus similar to genesis neurons that will allow them to retain the current vp schedule

[dctech]

quote “Those with 6 month neurons (currently the threshold for gaining maturity by voting), would then have 3.75 year neurons”. seems typo. it should be 3.75 month. @dominicwilliams

[Roman]

Dear Dominic,

As always, read your thoughts is a real pleasure. Thanks a lot for this.

Simple question : does Dfinity have any visibility of the behavior of the restakers from the two extrema (6 months – 8years).

My point is : if the main part of the stakers getting the biggest rates of rewards restake all their rewards, the inflation is not a problem since it does not bring any sell pressure.

For example, as an 8 years locked neurons owner myself, I have never sold any icp. I restake all my rewards. All of it. Since the beginning. If the main part of the 8 years stakers do the same, it means the main part the inflation due to them is kind of neutralized.

My fear is that by these reductions, we create more inflation since :
– those currently selling pressuring would continue selling ;
– more short-term stakers would be in invited in the game, potentially with the same behavior : selling ;
– by reducing the rewards of long term stakers, we make long-term stakers want to start to stake less or even unstake.

But I start from the assumptions that the long-term stakers restake, while the short-term stakers sell. Caricatural of course. But it is the limit case from which we need to start in order to check if these changes would be optimal or not even in that case. Before checking what is really the case.

Basically, we need to make those who restake, restake more and more, and make those who sell, sell less and less.

Even if staking is a risk because the protocol is by design evolving (tokenomics included), so that we can not complain by telling “I had been promised x %, I want them” ; this seems very good (and necessary imho) to give confidence to past and future investors :

@dominicwilliams

Here’s what I would suggest might be done: adjust the way the dissolve_delay_bonus works for new staking, but not for neurons that are already staked. So if there’s a 2 year neuron, it continues to have the same voting power before (unless it was Seed/ECT and is in the process of dissolving, since this cohort didn’t stake on the basis they would get some specific voting power, and no promise is being broken).

New staking would work according to the new scheme however.

Regards

[ScouseWill79]

I think i’d just sell my II and get out completely, this isn’t what I believed and became sold into ICP for 8 years never dissolving

[Jdcv97]

This is not a “bonus” for us dominic, I invested 100k to buy 10.000 icp and locking it for 8years for a 16% understanding will decrease over time while the age bonus still got me in the same 16% after 4 years no dissolving.

I’m an investor not a gambler and this will affect my personal investment ROI in more than a 20%, the 20% reduction is from the 16% leaving the utility at 12% a year, and now decreasing from a 12% huge difference.

I wouldn’t locked for so many time if the main reason of locking that is get a revenue for doing so will be changed so easy. What is the second proposal about, it will leave the rewards as are today if adopted, no modification? There’s another proposal about reducing voting rewards if you don’t activate a new kind of “key” making impossible to sell the neuron, if you activate they key you have same rewards, if you don’t you cannot sell, what is all this about man? So were the tokenomics designed bad that now is a worry that
Affecting people’s investments? I love icp and am all in here, but you are affecting your first supporters because of the VC’s this proposals are not about inflation but to stop them from selling, then rebought their shares, but don’t affect us.

[dominicwilliams]

Yup was typo thanks. Posts hv to b 20 characters, so blah!