Staking - Voting Rewards & Age Bonus Behaviour Question

Question about the Voting Rewards & Age Bonus over time.

I have read the descriptions below and have listend to official R&D videos from way back about voting rewards and also heard it from a lot of people in the space.

From what i understood is:

At this point in time the voting rewards for 8 year undissolved is at 16.1%.
Over time as more people own ICP and the price goes higher, the 16.1% voting rewards will go down to a minimum of 5%. Right? The main question comes in at the next point below.

I also have seen, heard and read, the age bonus will increase over the years. Wich is also clear to me.
The question is, people are saying that the age bonus will roughly cover / even-out the decrease of the 16.1% voting rewards that are going down to 5%. So that at the end, the outcome would be about the same rate as if the 16.1% never went down to 5% because of the increasing age bonus.
So one can get a lifetime yearly return of around 16% when staking for 8 years in a undissolved state?

It would be great to get a answer from a official at Dfinity to confirm this.
To be sure before i create another 8 year neuron.



This is a good question. I’d like to add on that in particular, I’d like an answer on how sustainable that is in the long run. Right now it seems ~1500 ICP or so (~$15,000 @ current prices; $4500 when price was ~$3) locked nets about .5 ICP a day. I’m sure there are fewer wallets with that amount of ICP than there are with it. However at high prices assuming these stakers are still accumulating at 16%, let’s say one ICP is $50. These individuals are receiving ~$25/day, perhaps indefinitely? They’ve made their investment twice over in a given year while the principal sat still and continues doing so. I can’t say whether it is or is not sufficient to pay $4500 and take advantage of the tokenomics in this way indefinitely as that is subjective but I would like to know how sustainable that is as the ICP token goes up in price as I believe the ICP token could be worth @ least $700 one day or more. Perhaps I’m overlooking something or missing something as I’m just getting into totally understanding the ICP tokenomics so I just want to get some additional thoughts on this model going forward for the IC. Thanks.

here is my opinion on this. It’s very sustainable as game theory would dictable most people will not stake never hit disolve delay at some point. Let’s say I am wrong 50% of people decide to do that. You be paying 16% on 50% of the token to not exist.

Awaiting a official answer or description please.
Would love to clearly know how this works, thanks a lot.

Where can i find a clear layout on how it behaves?

  1. The 5% refers to the total inflation from voting rewards for the entire network. It’s currently at 7.25%. You can see that here:

  2. It’s hard to predict because there’s no “final” APY for neurons. It’s partially based on the total inflation rate mentioned above, and also dependent on how much ICP is locked in neurons, and for how long, and how much age bonus they all have. Currently, 48.3% of the supply is staked. If it doesn’t increase much in the future, then the age bonus should be enough (or close to enough) to counteract decreasing rewards. But Dfinity is targeting 90% in the future. At that point 8 year neurons could fall to 10% or less even with max age bonus.

Yes, that is the estimated annualized voting rewards rate for an 8-year neuron with no age bonus as of today.

No, that’s not the best way to think about this. First, the price of ICP has nothing to do with voting rewards. What you are referring to with the 5% value is the voting rewards inflation function, which you can see on the ICP Dashboard ICP Circulation page.

We can see that this theoretical annualized inflation rate is currently 7.25%. That is, the total of all voting rewards for a year is up to 7.25% of the total supply of ICP.

This means that if all ICP was staked with the same neuron dissolve delay and age, then the annualized voting rewards for all neurons would currently be 7.25%. But all ICP is not staked, just 48.3% of the total supply is currently staked. The less ICP that is staked overall, the higher voting rewards will be. The more ICP that is staked, the lower voting rewards with be. So at 48.3% staked, the current estimated annualized voting rewards rate for an 8-year neuron with no age bonus is 16.1%, but this number would be higher if less ICP was staked, and would be lower if more ICP was staked.

I’m not sure what you mean by this. Age is a property of a neuron. If a neuron is not dissolving, its age will increase over time, and its age bonus will increase to a maximum of 25% at an age of 4 years. The maximum age bonus that a neuron can have will be 25% forever, unless this maximum is changed by an NNS proposal. So what’s true is that if you create a non-dissolving neuron today, your age bonus will grow to 25% over 4 years.

As pointed out earlier, comparing 16.1% (the current voting rewards of a neuron with certain properties) to 5% (the future inflation from voting rewards of all neurons) is comparing apples and oranges. It’s true that voting rewards inflation decreases over time, from 10% to 5%, over the first 8 years since Genesis. This has likely led to voting rewards slowly decreasing over time for many neurons. And this decrease has likely been somewhat mitigated for neurons that have had an increasing age bonus. But we can’t predict what percentage of the total supply will be staked in the future, so it’s impossible to know how much voting rewards will be in the future.

No, you absolutely should not count on this.

While we can’t predict how much of the total supply will be staked in the future, we can calculate what voting rewards would be for different scenarios. Say that in 6+ years (when voting rewards inflation is fixed at 5%) the exact same percentage of the total supply is staked as today, 48.3%. Here are two examples of annualized estimated rewards:

  • Neuron with 8-year dissolve delay and no age bonus: 11.1%
  • Neuron with 8-year dissolve delay and maximum age bonus: 13.9%

If less than 48.3% of the total supply is staked, those numbers would go up. If more is staked, those numbers would go down.


Here are the same two examples for the 90% staked scenario in 6+ years:

  • Neuron with 8-year dissolve delay and no age bonus: 6.0%
  • Neuron with 8-year dissolve delay and maximum age bonus: 7.5%

Disclaimer: Everyone should do their own calculations. I did these calculations quickly and can’t guarantee that they are correct. I’m also not speaking for DFINITY, I just saw the thread and wanted to help answer the questions.

Thank you for the detailed and clear answer wow, now i understand.
Its a very good deal after all.

Thank you, i truly appreciate it!

1 Like

That 25% age bonus, is 25% of what exactly?
25% + of the voting rewards, or 25% + of the ammount of ICP thats locked?
The first one i guess.

It also depends on the distribution of dissolve delays and age among neurons, so view the numbers I provided as just a rough estimate.

1 Like

Yes for sure. I calculate with looking at the above mentioned 90% scenario. Better to have a pleasant surprise at the end. 90% would that even be possible for the whole system to operate like that?
That would be pretty high…

Could you shortly explain the age bonus of a not dissolving neuron.
The maximum age bonus will be 25% after 4 years. But 25% of what exactly?

25% of the current 16.1% voting rewards? If in this example the voting rewards would be at todays rate. 25% of the 16.1% voting rewards, would mean 16.1% + (25%) 4.02% = 20.12% ?

Sorry if i make it confusing.

The help text on the ICP Dashboard Neuron page (example neuron) provides a basic explanation of all neuron properties.

Here is the help text for age bonus and voting power (emphasis added):

Age Bonus

A boost to voting power of up to 25% and a linear function of the age of the neuron (capped at 4 years). A neuron with an age of at least 4 years has a 25% age bonus. A neuron with an age of 2 years has a 12.5% age bonus.

Voting Power

The voting power of a given neuron, and its relative claim to voting rewards, are calculated according to the quantity of staked ICP, the length of the configured dissolve delay, and the neuron’s age. The voting power of a neuron at a given point in time is computed as follows:

voting_power = neuron_stake * dissolve_delay_bonus * age_bonus

For example, for a neuron with 100 staked ICP, a dissolve delay bonus of 50%, and an age bonus of 10%, the voting power is 165, calculated as follows:

voting_power = 100 * 1.5 * 1.1

Voting rewards for a given reward period (typically one day) are divided among neurons according to the relative voting power that those neurons contributed to the proposals included in the reward period. Since a neuron cannot vote (or earn rewards for voting) when its dissolve delay falls below 6 months, such neurons effectively have no voting power.

So age bonus is a boost to voting power, and a neuron’s voting power relative to the total voting power of all neurons is what determines that neuron’s rewards.

For example, if the voting power of a neuron with no age bonus was 1 and the total voting power of all neurons was 1000, that neuron would receive 1/1000 of the total voting rewards. If that neuron instead had an age bonus of 25%, its voting power would be 1.25 and it would receive 1.25/1000 of the total voting rewards.

1 Like

Ok thanks for providing such a well explained answer. Very interesting and serious stuff.
Thanks a lot for the effort, appreciate it!

1 Like