Tokenomics Proposal [Community Consideration]

I see neuron sales/transfers kinda working like life insurance policies or annuities sales…where you would have to take a hit or get paid out cents on the dollar to cash out early but it could become an interesting market on an exchange. Unique compared to other tokens.

Example: a neuron with 100 ICP with an 8 year delay.

Why would someone sell? They need emergency liquidity maybe…

Would you pay market price for that? Of course not because you could just do that on a normal exchange. So you would buy at a discount and your return would be the market price after dissolve.

I don’t see the vote buying argument. If you wanted to buy votes go buy up tons of ICP on an exchange and stake but you could get votes at a discount but only if people are willing to sell at a loss.


Here’s one concern I have. If neurons are transferable, in a market downturn many neuron holders would sell their neurons–almost certainly at a discount to the price of the ICP in the neuron. Anyone who wanted to buy ICP to stake in a neuron would almost certainly buy neurons on the secondary market instead of ICP. If long-term investors are all buying neurons instead of ICP in the downturn, the price of ICP could absolutely crater due to the severe supply/demand imbalance.


I am strongly against a secondary market for neurons. No one would be able to anticipate how an actor with hostile takeover intentions could social engineer people into selling their neurons or how the holder would dump the neuron price on such a market in a stressed market scenario. Preferably, introduce an unwinding mechanism where the neuron holder pays a non-linear (based on time to maturity, dissolve time, etc) termination fee in ICP (a % of the ICP locked in the neuron) that goes straight to Dfinity or distributed pro rata to all neuron holders as locked ICP in their configured neurons. Dissolve/unwind the neuron and the remaining ICP after termination fee goes back to the holder which then can be sold on the market.



If I follow this thread of @Alixthe and @wpb I’ve got a few questions on what is discussed there:

Are you talking about simple multisig or much more social recovery?
If you’re talking about social recovery in terms of neuron ownership/transfer you should really be aware that this is an absolute sensitive matter IMO. @Alixthe wrote:

Also, perhaps the Master II could be pre-configured for certain eventualities for example in the event of a death (which needs to be verified by say the beneficiary)

Where does social recovery start and where does it stop? The further it is pushed, the more it gets into the direction of KYC. I think everyone sees this 3 fat marked letters now, and knows how sensitive this could get. As it is formulated above, you put yourself into an oracle problem, which further relies on trusted 3rd parties. The beneficiary can act fraudulant as well, that’s where 3rd parties would come into play. And in the end you’ve suddenly got a neuron which is only transferable if a certain lawyer (completely different from country to country) agrees to this. This wouldn’t go well in cryptoland (see Davos FUD).

If it is only much more an addition/tool for someone it could be a whole different story. The 2 following statements are completely different:

“You’ll get an additional tool, where you can set up social recovery, if you want it.”
“You’ll only be able to transfer a neuron if you set up social recovery.”

One statement might give freedom, the other one might take it.


I think there should be some sort of delay or waiting period (no voting capability) before a transfer can be completed.

I think users should have to express their intent to transfer their neuron by initiating a countdown (weeks or months). During this countdown the neuron should no longer be allowed to vote. once the timer runs out the user should be able to set up a contract …

Your mechanism is interesting. I’m quite on your side with this. However this would already go deeper into implementation.
I think there are still many points which have to be resolved if I read the latest posts. You’ve also written:

It might be helpful if @diegop could provide the team’s rationale for setting it to 1 ICP in the first place.

I’m as well interested, and @cryptoschindler also writes:

I agree on this and would be interested in the Foundations reasoning behind setting the threshhold to 1 ICP


I see neuron sales/transfers kinda working like life insurance policies or annuities sales…where you would have to take a hit or get paid out cents on the dollar to cash out early but it could become an interesting market on an exchange. Unique compared to other tokens.

Absolutely agree on this.


Preferably introduce an unwinding mechanism where the neuron holder pays a non-linear (based on time to maturity, dissolve time, etc) termination fee in ICP

You’re approach is great the more I think about it. Maybe it’s a correct balance between complete liberal free neuron trading vs restriction and shadow market.
This could be enough to counter liquidity arguments: Paying a fee to unlock early is fair, otherwise I’d have to chose a shadow p2p market and trust on some unknown party. If I really need to liquidate I could, and ICP wouldn’t restricit it. With a fee there could be enough motivation not to do it. However it might become a political tool in the future (one governance vote makes restrictions much harder and the other one makes them loose). Since we all have governance rights that wouldn’t be such a problem (hopefully with enough decentralization).

I see your concern, but IMO that would lead to a healing effect. I’ve written something before which addresses your concern.

The main question we should ask ourselves is: Are we afraid to make some sincere postive longterm changes in tokenomics, because it might have a negative impact on the price over the short term?


No one should ever invest money that one can’t afford to loose in the first place. However, with that said, I completely understand that one’s life situation can go from better to worse quite fast and therefore it would be beneficial to unwind the locked neuron(s) in that case in case of the need of emergency funds. But in order to do so it must come to a cost and I believe it is best if that cost come to such a price that it should be close to ones last resort of options. This is the governance of the whole network and if one decides to unwind the cost of ones decision of doing so should benefit the rest of the neuron holders in the network why I am leaning more towards a pro rata distribution model of the fee paid for unwinding. We have a perfectly capable primary market where ICP is traded, creating one for neurons would most likely make this one very illiquid as well on top of the other concerns I mentioned earlier. Just my 2 cents!

1 Like

It is clear to me that the actual design (by Dominic and Dfinity team) was done with a long term vision rewarding more investors ready to lock their investment on long term. The longer, the better reward. it make so much sense. Obviously, I see that some people would want to have the best long term reward, while being able to sell their neuron quickly. At then end, 100% would go to neurons, the rewards would go down for everyone and the ICP price would also being kept low as the entire supply would be available for sell somehow. Just make absolutely no sense to me and this is not what I have signed up for. It is a hard decision to lock for 1 year and have your investment available in a quiet short period compare to 8 years but if you have any doubt, go for the 1 year. But we have to do it and live with it.
On another related but different subject, we should first vote on how many % of votes should be needed to pass some of the proposals. Not all proposals weight the same. A change proposal that affect others financials should certainly need a 80% or 90% approvals for passing. I don’t understand why or where it comes from that a 50+1 adoption level was adopted for a fit for all proposals.


Very interesting @lastmjs and thanks for the information ~ just read Vitalik’s article with one of his key takeaways summarised below:

To me, the goal of crypto was never to remove the need for all trust. Rather, the goal of crypto is to give people access to cryptographic and economic building blocks that give people more choice in whom to trust, and furthermore allow people to build more constrained forms of trust : giving someone the power to do some things on your behalf without giving them the power to do everything. Viewed in this way, multisig and social recovery are a perfect expression of this principle : each participant has some influence over the ability to accept or reject transactions, but no one can move funds unilaterally.


Hello @Michaelannh ~ thanks for tagging me. Surprised that a simple post can generate so much interest. The beauty of this is right or wrong, ideas are being put forward as there is so much to learn from others.

For me, I am just a layperson throwing out ideas for discussion as I am neither a crypto tokenomics expert nor someone who professes to be and you for one are obviously a learned and well researched person in this field based on your highly analytical and well thought-out posts.

I have not been keeping pace with all the comments here on this forum but have read snippets here and there. With so much interest generated and comments received (93 posts and counting), we should probably start thinking of summarising what has been discussed as follows:

  1. Without over-thinking, what is the number 1 problem (if any) in ICP’s tokenomics that needs to be solved right now?

  2. If this problem has been identified, what are the proposed solution(s) (extra points for proposed solutions from qualified and experienced crypto tokenomics experts) that the community can come up with?

  3. From all the proposed solutions received, what are the pros and cons in implementing each of the proposed solutions together with supporting evidence / calculations / research backing up these proposals?

  4. Put the best proposal to be voted-upon by the community.

  5. Repeat step 1.

A summary of the above should be pinned to this main post for the benefit of others who have not kept up with all the posts here (apologies if I am somehow reiterating what @lastmjs may have mentioned previously but I guess this should be high on the agenda).

On a side note, I have also noticed a few posts about issues with transferring of neurons. Not sure if this is the same but Harrison Hines mentioned in his recent community conversation that “neurons as NFTs” was on Fleek’s roadmap as per the following link:

Wonder how this is going to play out.



I highly agree with the eventual importance of having different vote threshold levels for different types of proposals.


Another challenge with neuron transfer: wouldn’t this very much complicate liquid democracy?

Once neurons start following a broader variety of other neurons, wouldn’t neuron transfer make it much more of a challenge to ensure that you’re still following who you want to, given that the neuron you’re following may at any arbitrary time be sold to another principal? Many of the guarantees of liquid democracy would seem to break down.

This could also very much compound the security risks here.

Malicious actors technically need only a significant proportion of the liquid voting power on the IC at a given moment to determine NNS outcomes. Some neurons can have small ICP stakes, but a ton of voting power from followers and their followers. So it could be relatively very cheap to buy at-least-momentary control of the IC if neurons can be transferred - nothing remotely close to the cost of “50% of the ICP + 1 vote,” as is more commonly discussed.

Unless these challenges can be addressed, supporting neuron transfer looks prohibitively risky.


I love this. Personally, I am somewhat confident that I could create a small group of 3-5 people that would be one of the most unbreakable networks ever to exist, considering how strong my trust is in them. I would trust them with my funds, if not my life. I hope many of us have people like this in our lives.


I think these are really good points.

If neurons can be sold, neurons basically just become speculative ICP futures (with certain optionality attached, and bonus-adjusted for dissolve delay and age bonuses). The incentives for ICP holders to steward the IC’s development for its long-term success mostly evaporate.

Combined with the fact that liquid democracy could become unworkable with neuron transfer (per post above), how would governance on the IC function anymore?


so in a free market parties can decide on rules and boundaries. a free market isn’t willy nilly anything goes as participants are free to make rules and come to agreement among participants; you are free to not participate and create your own market with no rules at all and see if anyone wants to be a part of it as well. the rules of the dfinity market are such that people who funded this project and gave it money knew that it would be a staking type system, and in general staking type systems involve locking up coins which generally entails not being able to sell that account. given that no one was forced to give the foundation money it seems obvious to me that the free market of people who gave money went along with this idea.

i don’t know how many have set their dissolve delay to 8 years now but those who have basically have consented to the current arrangement. you can talk about free market all you want but you must recognize that in a free market actors are free to collude and agree on rules/contracts

regarding the minimums have you considered the “nothing at stake issue” this is a major issue and well i do not know the internal functions of the dfinity system i can tell you that perhaps you should consider the idea that another reason the minimum exists as such is because of scaling issues around the ledger that need to be worked on. so the solution isn’t let’s make it harder for investors but lets make the system function better to provide more access. and to your point about decentralization if we were to move to a % based system it would mean that less new coins are available to be sold to the market thus thwarting the ability of new stakers to come on board.

and btw the minimum on Eth is 32 Eth, which is quite a lot, and so yes you actually get more for having a high stake vs having only 16 eth and needing to participate in a pool. By comparison waiting for 1 ICP ($55) to be generated on a daily basis given that the current yield is about 25%+ (if locked for a long time) requires about 1470 ICP which currently has a market value of $81k, whereas 32 Eth is over 100k; so i’m not sure how you can claim that the other systems somehow offer something better to the small investor.

finally as far as payments for following as you say they should collect fees but they can also choose not to collect fees (remember that free market!). an entity that wants more votes would obviously have lower fees or even no fees. in the case of dfinity foundation most of the important proposals come from it at this point and it has a sizable war chest so one can imagine that they value having more followers and voting power at this point rather than charging some fee.

your concerns are well places but I think in the case of the 1 icp minimum you are attributing to some sort of poor design decision to what is more easily explained by technical scaling issues that in the future may improve. and with regard to neurons you have to acknowledge that while it’s true that someone locked into an 8 year dissolve may sell their vote it would be highly irrational for them to sell their vote for a change that would have a negative impact 8 years down the line, whereas if he were free to sell his neuron whole the system losses the extra protection of participants who must without doubt consider what are the long term consequences of their actions.



You’re right. That’s what the whole discussion is about at the moment.
I’d personally like looser conditions at many points. Others don’t. So we agree on frameworks and the majority (Dfinity) decides. If I think that some points aren’t worth the risk, I could just move to another crypto project.
And that’s the point: I’ve provided so many written lines in here, which show how other projects are governed, and what Dfinity/we as a community could learn from them to improve our tokenomics.
Another of your lines:

No, staking type projects don’t require locking up tokens in general. See Cardano for an example. Others have less restricitve rules. One of them is Polkadot (28d unlocking period). Yes, stakers agree to the terms and conditions Dfinity sets, but my point is that others have far better conditions (IMO). You can neglect that fact but in the end, competitors don’t sleep and that’s exactly where pressure starts. Why should someone stake for Dfinity if he has far better conditions somewhere else? For the tech?

To which extent does that compensate better conditions? If you want to sell leasing contracts for a Lamborghini, and prevent people to get out of it, that’s ok, but that’s not my approach. Before doing so, I’d ask people why they need a Lamborghini and if it is a good investment? If they can’t say anything I’d ask them if they really needed a leasing contract and if they could just buy a cheaper car instead. They could also pay a few hundred bucks to drive a Lambo on a race track for an hour if they really needed it for their pleasure. Now translate that to ICP vs Eth/Ada/Dot.

Furthermore you’re already able to sell your Neurons via the shadow market. So that argument is also not valid atm (see discussion above):

And now to your next point:

I’ve already made some points in another comment:

Your next statement:

Here’s an interesting comment about this.

As you can see there is a huge gap. Either you’re in long term (8 years) or you are more on the “I stake for 1-3 years and look on how the project has developed by then. Maybe I’ll stake longer, if I like what I see.” Long term could be people with high interest in the projetct/or long-term commitment (Developer/Dfinity wallets themselves)
That’s what the free market tells me.

As you can see above, the free market gives you information about the commitment. So what does it tell you that so many actors don’t stake longer than 1-2y and that there is nearly no one staking for 5,6,7y?

I’ve considered it. A simple yes/no answer + a reasonable explanation from a dev would have been enough. So why is there neither an yes/no answer or a reasonable dev explanation?

Actually exchanges offer staking services. You don’t need 16 Eth there…
However you’ve got a valid point, if you throw in eth/dot vs icp. Their capital requirements in minimum stake benefits large holders (32 eth/40dot). At Polkadot they’ve decreased (~200Dot to 40Dot), however if the price rises it gets countered again. In the end it’s a problem and they’ve to work on that.
As you’ve pointed out, they might benefit large holders there, and ICP benefits them with the 1ICP treshold.
However we should ask ourselves if both systems have their flaws and if there are others which are a bit better? Cardano doesn’t have requirements like that, as far as I know. I also think that Eth will bring down the 32 Eth after the merge. Dot is also decreasing it. => They improve their tokenomics.

We could just play it down and watch them pass ICP. Not my choice.

I get your concern. You’re basically saying that price dumping in voting fees might counteract the decentralication which was initially intended. So it might become a tool which adds to unfairness like the 1ICP treshold.

Just a few thoughts:

Why do you assume everyone would follow Dfinity it they had alternatives? Of course most devs and Dfinity are practically in one boat, but in spite of that, I’d prefer voting for a dev directly.
Let’s take @lastmjs for an example. I listened to his podcasts, read his stuff on Twitter, etc…
If he’d set up a voting neuron and he’d take 5% from followees , I’d do it, because I’d know that he’d vote in my interest (since I agree to his views). Would you vote for a party/politician you don’t agree with, just because they offer you more (I don’t want to say that I don’t agree with Dfinity)?

I know many people in this world do so, and that’s a huge problem. Many vote for something in the short term without the long-term view. Do voting fees contribute to this misbehaviour or not? Let’s draw a scenario: Party A collects fees, has a high amount of followers and starts fee dumping so its follower get more. More followers will add their votes to A: A gains more and more power and suddenly makes a vote which is absolutely harmful for everyone. Maybe they don’t get it in the short term but suddenly ICP turns into something we know from 1930-1945.
IMO fees are not the problem if we fear centralization. Fees are only an instrument for bad actors to help them fullfill their wishes. Fees wouldn’t contribute to that as long as:

  • Huge stakes have got sincere interest in decentralization, doing the right thing and
  • Actors don’t purely put their votes into something out of short term capitalistic interest

If both points are invalid, we’ve got a problem. We’ve got 2 security levels.

IMO the liquidity inequality via the 1ICP treshold can add much more to centralization than price dumping of a voting fee. Scenario here:

Party A (huge stake) waits for a sudden price increase and liquidates the huge stack, generated via the 1 ICP treshold.
The market dumps and party A buys back the liquidated ICP for a cheaper price. Stakers with a low stake can’t do anything in the meantime. They can’t interact with the market and counter A’s effort to centralize. Repeat this steps a few times and suddenly you’ve got an instance with absolute power, without ever relying on other voters. You’ve only got 1 security level:

  • Party A has got sincere interest in decentralization

If we’d like to further introduce additional mechanisms for decentralization Dfinity could also integrate a minimum fee on voting neurons to avoid price dumping on voting fees. This would benefit lower neurons thus lead to more decentralization. Now you’ve got 3 security levels instead of one.
=> Big player could shut down the minium fee via a vote (they’d require a majority for this). Afterwards the majority of capitalistic players would have to vote for him. And in the 3rd step his malicious interest has to take over. However price dumping in voting fees isn’t even necessary to gain the majority in power. You could even do that without it. As you see the voting fee could be a governance tool. If centralization is too much you could incentivize voting for small neurons via adjusting the fee (maybe Gaussian distribution?).

I sincerely think that the community should start to make a load of scenarios on what might actually happen.

I just want to say i think this is a really great back and forth you have going. As someone who is observing the conversation I do take issue with a couple of your points:

  1. Why do you think we should try to learn anything from the tokenomic models of ADA/ETH/DOT? Is it because they’ve been around longer? Or are you saying their % staked is higher an therefore they are a better model to follow? Proof-of-Stake networks are still very new and I don’t think we can consider these models to be proven effective just because they seem appealing today.

That is definitely an interesting point. However, there is a growing number of people staking ICP for 8 years. Most of which aren’t dissolving. I don’t have a reference now but I will find one later and update. What this tells me is that we might consider changing the max lock-up to something like 3-4 years. To me the extra 4 years seems more irritating than beneficial. Would be great if @diegop could get someone from Dfinity to comment on why 8 was chosen.

I know we have talked about this topic quite a bit in the ICP Maximalist telegram channel. If you’re interested in taking part in a (respectful) discussion you are more than welcome to join the conversation. See @icpmaximalist on Twitter for telegram link.


Great podcasts I must say

1 Like

Fair question. I actually don’t know their percentage numbers. If it comes to that eth is actually still pretty bad (~7 Mio Eth on ~116 Mio Eth in circulation). Lock up till after the merge (unknown) and already down at 5% annual return. Quite some risk for that return.
Cardano leads the statistics here I think. I’d have to research what the percentage staked is at the moment.

You’re absolutely right with the argument. Other approaches are still young and no one knows which one is the best atm. That’s why I try to give an overview and ask questions. Time will show which one has got the better tokenomics/governance. However if I compare some of them with ICP and watch the main arguments on why ICP is governed that way, I’ll always get the same impression:

We are far more concerned about a 51% attack than any PoS competitior. Much of our governance/tokenomics is founded on that. 6month lock up, discussion about neurons, etc.

And that’s my main point: I can’t predict the future, but this discrepancy is concerning me. If I look at Polkadot and Gavin Wood I’d be really cautious if I’d claim they’ll fail because they don’t have a 6month lock-up. They have good developers and tech as well.
So why do they think this can be done with 28day lock-up and we think we need 6 months lock-up?
Or why does Cardano think they don’t need a lock-up at all? By the way, all mentioned projects have got a higher market cap than ICP atm. So their holders have got far more to lose than we have, and yet we’re far more concerned about a 51% attack than they’re. Why?

Either they’ll fail miserably with their approach and ICP is right with the lock-up to ensure long-term governance, or in the end they’ll succeed and pressure on ICP will increase steadily.

I really don’t have answers to that, and no one has given me a satisfying one yet. Is it only because we are far more cautious? However if we bring in that argument I could also counter it again with centralization arguments: The longer ICP/Dfinity stays too cautious/restrictive/protective and fears giving up governance rights, the less likely its decentralization will become. Who takes part in the governance if it stays too protective? This could be a huge side effect.

And that’s why I think we should learn from other projects, or at least have a discussion about their approaches vs ICP approach.

That would be an interesting approach.

Thanks for your invitation to Telegram. I’ve already written with @ayjayem about it. I’m not the biggest fan of this messenger. Maybe there will be an own format anyway? There was a Townhall meeting on discord as far as I’ve seen.

1 Like

Hey @MichaelAnnh @LightningLad91

I definitely would not want to change the max dissolve delay to make it shorter.

The max neuron dissolve delay of 8 years is important because it discourages unlocking the neuron. It is more attractive to leave the neuron locked with an 8 year dissolve delay and liquidate ICP rewards indefinitely than it is to unlock a neuron with an 8 year dissolve delay so you can liquidate the total ICP in 8 years. Longer dissolve delays makes long term governance more attractive.

The long dissolve delay also creates deflationary pressure for ICP. Eventually we will see somewhere around 90% staking in 8 year neurons, which I predict people will chose not to reverse. Decentralization will occur because eventually it will make more sense for individuals to liquidate rewards than compound rewards.


Is it really the number 8 that discourages unlocking the neuron or the promise of max rewards that discourages this?

I agree we need a long term dissolve delay for max rewards but what I think we’re all trying to figure out is if 8 is really the right number. What is the major advantage of 8? 4 years seems like a significantly long enough period of time to feel the affects of any bad voting history so why do we push it out to 8? Perhaps there is a reason for the 8 year lock-up but that isn’t apparent to me.

If we want to promote more long term staking I think reducing the max delay might help. Some people may want max rewards but aren’t willing to bite the bullet for 8 years. I think this is reflected by the lack of neurons being locked up for 4, 5,6,7 years. We could also make the argument that lowering the range from 8 years to 4 years would increase the rewards for smaller lock-ups, promoting even more staking. I want to reach the 90% staked goal as well but I don’t think we need to shoot for all neurons being locked for max delay. I think we should promote any level of commitment from 6mos+ and if the 4,5,6,7 year range is wasting away then why not re-work the system so everyone benefits?

This all really just comes down to why 8 years? Why was that number chosen. I really hope someone from Dfinity can chime in and provide this context for our consideration.


Thank you for providing this perspective. I agree that we should continue to monitor the situation and adjust accordingly. I’m just not convinced we’ve given it enough time to determine if a drastically smaller lockup (< 6 Mos.) is warranted. I do think we should try to address the large gap in staking between 3 years and 8 years.

The great news is that the flexibility of the NNS allows us to make these changes when needed.