Reducing minimum staking time for expanding governance participation next steps

On 7th of June 2022 we created a proposal to reduce minimum staking time for expanding governance participation. The community response for the proposal has been great and very engaging. The post received 180 replies with passionate arguments for both for and against the proposal. Overall sentiment from the community was positive and several community members believe this proposal will benefit the ICP ecosystem. This proposal also addresses the benefits of staking of ICP directly on NNS for more effective governance and value creation.

Let’s take a closer look at few suggestions from the community members and our take on them:

  • A community member mentioned that, “Drawback of this proposal is that it would instantly increase the amount of voting power on the network, thus reducing the APY for all stakers”

For the new proposal we propose to change the minimum dissolve delay to 30 days instead of 7 days. Here is the breakdown of staked neurons from 0 to 6 months dissolve delay

Neurons with dissolve delay < 1 month 194
Staked ICP in neurons with dissolve delay < 1 month 37,52,197
Neurons with dissolve delay > 1 month and < 6 months 1420
Staked ICP in neurons with dissolve delay > 1 month and < 6 months 22,754,499
Source: api.internetcomputer.org/api/v3
  • Another post discussed the need for improving the UX and creation of more educational content around staking & governance.

  • This is an important suggestion and we strongly agree to the need for a better UX, the proposal is not between a choice to implement reduced staking or improve UX/educational resources.

  • We are strongly committed to improving the UX around staking and governance. Additionally we do foresee constant improvements in the educational content to help new users onboard effectively.

  • Governance is a shared topic that is core to DFINITY and the wider IC community. We would also like to encourage the community to contribute to ICP by creating content like articles, videos, tutorials, etc around the existing platform

  • A few people discussed that yields for staking are lower than those being offered by centralized exchanges.

  • The idea of the proposal and its intended changes is to drive governance adoption.

  • We are aligned to drive incentives for people who stake for a larger period of time. Hence the voting power will always be lower for short term stakers, when compared to that for people who stake their tokens for 6 months.

  • The risks of having a token in a centralized exchange have played out with recent events in the industry and decentralized network staking would soon become a preferred choice when people evaluate the risk-to-reward ratio

  • A user on twitter suggested to reduce the minimum spawn reward from 1 icp to 0.1 icp.

  • This hasn’t been raised by many people and requires additional engineering effort which may not be necessary.

  • One prominent argument was the alternative proposal from Wenzel Bartlett which proposes to reduce the minimum dissolve delay to 1 month AND increase the maximum dissolve delay bonus to 3.5 keeping all other tokenomics parameters the same. He has provided a helpful spreadsheet to compare the results.

    • We appreciate the effort of Wenzel to provide the research. One key challenge with this alternative proposal is that it will disproportionately penalize stakers with dissolve delay greater than 6 months and less than 5 years. The original proposal reduced the rewards uniformly from 6 months to 8 year. The diagram presented in this comment is helpful in explaining the difference.

There were some concerns from the community about security especially a 51% attack if dissolve delay changes to 7 days. While we think it is not a fair argument (as discussed above), we feel that the new proposed 30 days dissolve delay completely eliminates this issue as well.

The presented proposal is our hypothesis on what could drive the maximum impact. Based on our discussions with various ICP holders, we found that there are several people that are sitting on the sidelines due to the 6 month minimum lockup requirement. We feel shortening the duration will enable them to get some real skin in the game when they start to see accrued rewards. Staking 1 ICP does not have the same impact as the accrued rewards are not that significant in value. The eventual goal is about giving people more flexibility to choose with a lower barrier to entry.

Here is the updated proposal:

Proposal: Reducing minimum staking time for expanding governance participation

TL;DR

  • To encourage additional users participating in the governance process of the Internet Computer we suggest adjusting the rules for the minimum dissolve delay and dissolve delay bonus for neurons.
  • In particular, we suggest that the minimum dissolve delay for voting should be set to 1 month (instead of 6 months) and the dissolve delay bonus should be extended by a linear function with values between 0 and 1.0625 for the time between 0 and 6 months.
  • As a consequence, new users will be able to try out the governance process due to the lower minimum dissolve delay, while still keeping the incentive for users staking for longer periods due to much higher voting rewards for long-term staking.

Background

  • The voting power of a neuron is computed as the product of the staked amount, the dissolve delay bonus and the age bonus.
  • The dissolve delay bonus historically has been a value between 1 and 2 and a linear function of the dissolve delay (which has a maximum value of eight years). The current minimum dissolve delay for making a neuron eligible for voting and receiving voting rewards is 6 months.
  • Currently, a neuron with a dissolve delay of 6 months has a dissolve delay bonus of 1.0625. A neuron with a dissolve delay of 8 years has a dissolve delay bonus of 2.
  • For a detailed description of the Network Nervous System and the determination of voting power please refer to this article.
  • With a lower barrier to entry, we believe that more retail participants will be able to experience the entire workflow of staking ICP, casting votes and accruing voting rewards Our hypothesis is that the proposed change described in the next section will significantly increase the number of community members participating in the IC governance. We also believe that over time some of these community members will start committing to longer staking periods that offer larger voting rewards.

Proposal

  • To encourage additional users participating in the governance process we suggest that the minimum dissolve delay for voting should be set to 1 month instead of 6 months.
  • Given that we would like to keep the current incentive to stake for longer time periods greater than 6 months, the dissolve delay bonus between 0 and 6 months should be much lower compared to longer periods. As a simple solution we suggest that the dissolve delay bonus should be extended by a linear function with values between 0 and 1.0625 for the time between 0 and 6 months. From 6 months to 8 years the dissolve delay function should be unchanged The new function is illustrated in the following chart.

  • For a newly staked neuron, this would result in APYs depicted in the following chart. The values for 1 month are new (previously the APY was 0% as no rewards were given) and the values for 6 months and 8 years are unchanged.

Impact

  • The overall average APY would change from 15.72% to 15.46%.
  • Approx 2m out of 116mn daily rewards (i.e. 1.7%) would be allocated to neurons with dissolve delay < 6 months. → The overall impact is relatively small.

Assumptions:

  • An additional 27mn (24mn dissolving, 3mn non-dissolving) of staked ICP becomes eligible for voting/rewards
  • This results in an additional voting power of 14mn. This is based on neuron data from the dashboard API, ignoring any age bonus (which would only affect the small porting of non-dissolving neurons).

Security

With respect to a 51% attack on the voting power of the internet computer, we believe that the proposed change is not relevant. If an attacker has 51% of the voting power, then also the minimum dissolve delay can be adjusted. Thus an attacker would likely stake for 8 years (instead of staking for a shorter period) receive a dissolve delay bonus of 2 and then attempt to change the rules.

With respect to the implementation, we consider this to be a security critical feature as it has an impact on voting rewards and thus the minting of ICP. Therefore it is planned to be reviewed with DFINITY’s security team.

Alternatives considered

  • With respect to extending the function for the dissolve delay bonus to the time between 0 and 6 months, we also considered concave and convex alternatives. A concave (respectively convex) approach would provide higher (respectively lower) voting power and APY compared to the proposed linear function. Given that the precise choice turned out to be irrelevant for a 51% attack as outlined in the section Security, we concluded to use the simplest approach via a linear function. We also considered extending the same function but ruled out as it will offer significant awards to people staking for very short duration such as 1 month which is counterintuitive.
  • Alternative Proposal proposed by Wenzel Bartlett which reduces the minimum dissolve delay to pay voting rewards to 1 month AND increases the maximum dissolve delay bonus to 3.5. %APY comparison for original and alternative proposals:

Drawbacks

  • Existing stakers staking rewards are decreasing albeit slightly.

About Me

I am Nikhil Ranjan working as a Senior Developer Relations Engineer at DFINITY. I am focusing on DEFI and Tokenization on ICP.

7 Likes

I would like to see the option in the Following list to always reject proposals from a certain group. We have a couple in that list I would always reject but they are always adopted, in my opinion, because they are followed for the rewards and that is a security issue for me.

No response, of cause not, when spammers would have to pay 10icp for a rejection, then no more spam and false rewards, democracy in action and you say Dfinity has too much power.

We have submitted the NNS proposal Internet Computer Network Status

I somehow missed this a month ago or so. I don’t really want less skin in the game from anyone helping make decisions.

I’d disagree that additional participation for participation’s sake is a misguided pursuit. The magic in this whole thing is that if you make bad decisions you have to live with them. This nerfs that mechanism and gives leakage to potential black swans where people without long-term skin in the game have an outsized effect over governance.

I understand that right now there may not be enough voting power to affect things one way or the other, but trends and balances change. I’d like to see an expiration date for this if it is really for attracting users. Say 3 years? Maybe 5?

3 Likes

Hi all,

we would like to give the following update on this proposal:

The proposal was submitted by DFINITY as an idea to foster decentralization by attracting additional users to the IC.

Given the syndication and the interim voting results, the majority of the community is not in favor of this proposal. While there was consensus that it would be desirable to reduce entry barriers to the governance system, the community was not convinced that the concrete proposal of reducing the minimum dissolve delay would bring the desired effect (while still having some drawbacks).

As a result of this feedback, DFINITY voted no on this proposal and intends to work with the community on alternatives.

At the meta level, even if this proposal came from DFINITY, DFINITY sees this process working as it should: an entity proposed an idea (in this case DFINITY), and the community voted that it was not mature enough and provided constructive feedback.

4 Likes