DAO's Departments with delegated governance

**DRAFT v0.1** - Beware: This document is still in the proposal stage and DOES NOT represent how PieDAO governance works right now. It is likely to change in the near future.

Governance

DAOs are designed to disintermediate the hierarchic structure of organizations through their transparency, effectively giving direct control to shareholders. Key features are transparency, the immutability of the organization, and the accessible nature of the platform they are built upon.

Governance models in DAOs however still present challenges.

Background

Governance is at the core of many DAOs. Unlike solidified protocols like Bitcoin, DeFi projects might need a greater level of granularity in decision making. It’s imperative that the process of decision making is transparent and incentivized accordingly.

Delegation mechanisms can play a positive role in reducing the amount of work required by token holders. Parties with significant delegated tokens need to have enough reasons to not be malicious.

Stimulating participation in governance is a combination of reducing friction from a user perspective, ensuring the distribution of information about governance proposals to the involved token holder, and incentives.

The proposed model below is designed to free the average DAO token holder from day to day work, while ensuring continuity in governance for specific tasks. Average token holders can any time remove decision power from delegated parties.

A staking model for active delegated governance

In the following pages, we introduce a DAO governance model that is designed to create operative departments within the DAO to work on a specific task, with aligned incentives via staking to boost participation and punish misbehavior via slashing.

DAOs with a large number of token holders often suffer a common condition: low participation in day-to-day activities. The vast majority of holders are voting only for high impact/controversial proposals. A few are instead voting often on operational decisions without getting rewarded or explicit delegation.

This model aims to change that.

Stakeholders

DOUGH Token Holders: DOUGH is the most important asset for decision making. Any decision can be taken or revoked using DOUGH. DOUGH is a non-transferable asset that has the sole purpose of facilitating participation in the PieDAO governance mechanism. Holders of DOUGH have the ability to vote on decisions in the DAO and exit the DAO (aka rage-quit) whenever there is not an active vote that they initiated.

FLOUR Token Holders: DOUGH holders can burn DOUGH for FLOUR at any time. FLOUR is free to transfer, trade, and stake. One DOUGH can always be burned for one FLOUR.

Team member: In order to become a Department member a candidate has to stake FLOUR into the election contract. The candidate's stake sets the upper limit on what others can stake towards their candidacy. The Candidate Stake represents 5% of the maximum amount on what others can stake towards their candidacy. Similarly how a candidate

Approved members must vote on questions raised by members of the PieDAO community. They may vote YES, NO, or ABSTAIN. Failure to vote results in the slashing of a portion of that member's stake.

A team member can charge a fee on rewards generated by delegated funds. Such fee is extracted to the Team member at the end of each epoch providing the Team member has not been slashed for lack of participation.

Delegators: DOUGH holders can delegate to a Team member in order to generate more DOUGH.

Team Formation

Once per epoch (t = number of blocks in the epoch), per team, an election is held to pick an amount (n = number of members to elect for the team) of team members. The election runs from startBlock to endBlock, once endBlock is passed the n candidates with the most FLOUR staked on them win the election and serve for the epoch unless removed.

Misconduct and Slashing

The system recognizes 3 security threat levels to the system. Each level of penalty, represented in percentage points, indicates the amount slashed according to level.

Level 1 Misconducts that are likely to happen eventually to most participants. Proposed slashing 10.0% of the stake.

Level 2 Misconducts that can occur in good faith, but show bad practices. Proposed slashing 50.0% of the stake.

Level 3 Misconducts that are unlikely to happen in good faith or by accident, considered malicious. Proposed slashing is 100% of the stake.

Right after an election, every Team member starts at Level 1. If they fail to fulfill their role or generally act in bad faith (ie Team excessive misconduct), DOUGH Token Holders can decide to raise their penalty level.

It's in each Delegator's interest to make sure their Team member performs work required. Otherwise, a Delegator should ask DOUGH Token Holders to increase the penalty level of their Team member.

Levels of Misconduct are multiplied by the Member Participation Percentage, applied retroactively at the end of the quarter.

Delegators are slashed a maximum % value Delegator_Penalty of the amount their Team member is slashed. Proposed: 40%.

Baking DOUGH, inflation module

Through the process of staking FLOUR, new DOUGH is created.

The inflation rate is decided by governance which jointly decides an Epoch Reward amount and Bonus Reward.

Team members with good performance at the end of the year and users delegating to them are entitled to the Bonus Reward.

Good actors are rewarded for bootstrapping the governance mechanism and participating in governance.

The Epoch Reward is distributed every quarter according to the number of votes the Team Member participated in compared to the total number of votes in the quarter.

Rewards / Slashing Function

TM = Team Member.
DL = Delegator to Member.
MPP = Member Participation Percentage.
MPP = TM_Total_Votes/TotalVotesInEpoch.
MIP = Member Informed Percentage
MIP = (Yes_TM + No_TM) / TotalVotesInEpoch.
// MPP and MIP are calculated every quarter.
E_Rewards_TM_% = 1 - (1 - MPP) * Level_of_Misconduct_%;
E_Rewards_TM = ((TM_Stake / 4) * (TM_Stake * EpochReward)) * E_Rewards_CM_%;
E_Rewards_DL = 1 - (1 - MPP) * Level_of_Misconduct_% * Delegator_Penalty ;
MIP_TM = MIP_e1 + MIP_e2 + MIP_e3 + MIP_e4;
TM_Bonus_Reward = TM_Stake * Bonus_Reward * MIP_TM;
Slash_TM = TM_Bonus_Reward - TM_Stake;

DAO Architecture

PieDAO is an Aragon DAO running the Dandelion organization structure.

In any voting-based system, there are cases for possible collusion or unavailability of stakeholders. For this reason, PieDAO implements rage-quit, a functionality popularized by MolochDAO which guarantees that participants can exit if they disagree with the decisions other members are making.

Governance parameters

Time

Action

MIN Approval

Support

72H

Voting

10%

60%

Apps installed

  • Redemptions: Allows users to manage a list of eligible assets held within an organization's Vault and allows members of the organization to redeem (burn) organization tokens in exchange for a proportional amount of the eligible assets.

  • Token Request: Allows users to propose minting tokens in exchange for a payment to the organization, subject to the approval of existing members.

  • Time Lock: Allows an organization to require users to lock a configure amount of tokens for a configurable amount of time in order to forward an intent.

  • Dandelion Voting: An enhanced version of Aragon One's voting app which implements an ACL Oracle which allows an organization to configure permissions that restrict actions based on whether an address has recently voted Yes.

Permissions

App

Permission

Grantee

Manager

ACL Oracle

Kernel

APP_MANAGER

Dandelion Voting

Dandelion Voting

None

ACL

CREATE_PERMISSIONS

Dandelion Voting

Dandelion Voting

None

EVMScriptRegistry

REGISTRY_MANAGER

Dandelion Voting

Dandelion Voting

None

EVMScriptRegistry

REGISTRY_ADD_EXECUTOR

Dandelion Voting

Dandelion Voting

None

Dandelion Voting

CREATE_VOTES

Time Lock

Dandelion Voting

None

Dandelion Voting

MODIFY_QUORUM

Dandelion Voting

Dandelion Voting

None

Dandelion Voting

MODIFY_SUPPORT

Dandelion Voting

Dandelion Voting

None

Dandelion Voting

MODIFY_BUFFER

Dandelion Voting

Dandelion Voting

None

Dandelion Voting

MODIFY_EXECUTION_DELAY

Dandelion Voting

Dandelion Voting

None

Agent or Vault

TRANSFER

Finance and Redemptions

Dandelion Voting

None

Finance

CREATE_PAYMENTS

Dandelion Voting

Dandelion Voting

None

Finance

EXECUTE_PAYMENTS

Dandelion Voting

Dandelion Voting

None

Finance

MANAGE_PAYMENTS

Dandelion Voting

Dandelion Voting

None

Token Manager

MINT

Token Request

Dandelion Voting

None

Token Manager

BURN

Redemptions

Dandelion Voting

None

Redemptions

ADD_TOKEN

Dandelion Voting

Dandelion Voting

None

Redemptions

REMOVE_TOKEN

Dandelion Voting

Dandelion Voting

None

Redemptions

REDEEM

ANY ENTITY

Dandelion Voting

Dandelion Voting

Token Request

SET_TOKEN_MANAGER

Dandelion Voting

Dandelion Voting

None

Token Request

SET_VAULT

Dandelion Voting

Dandelion Voting

None

Token Request

MODIFY_TOKENS

Dandelion Voting

Dandelion Voting

None

Token Request

FINALISE_TOKEN_REQUEST

Dandelion Voting

Dandelion Voting

None

Time Lock

CHANGE_DURATION

Dandelion Voting

Dandelion Voting

None

Time Lock

CHANGE_AMOUNT

Dandelion Voting

Dandelion Voting

None

Time Lock

CHANGE_SPAM_PENALTY

Dandelion Voting

Dandelion Voting

None

Time Lock

LOCK_TOKENS_ROLE

ANY ENTITY

Dandelion Voting

Token Oracle

Token Oracle

SET_TOKEN

Dandelion Voting

Dandelion Voting

None

Token Oracle

SET_MIN_BALANCE

Dandelion Voting

Dandelion Voting

None

Agent

RUN_SCRIPT

Dandelion Voting

Dandelion Voting

Agent

EXECUTE

Dandelion Voting

Dandelion Voting

Staking Aragon App Architecture

TBD

Election Aragon App Architecture

TBD

Appendix

One year staking journey

New elections for Tech Team Alpha is about to open with parameters:

Election Values

Val

Epoch Reward

2.5%

Bonus Reward

10%

Delegator Penalty

40%

Summary Q1

Alice registers as a Team member candidate by staking 10,000 FLOUR tokens. Since Alice staked 10,000 FLOUR, up to 200,000 FLOUR can be delegated towards her candidacy.

Bob delegated 200,000 FLOUR to Alice.

End of Q1

Alice Q1

Votes

Q1 Yes Votes

0

Q1 Abstain Votes

1

Q1 No Votes

0

Q1 Total Votes

2

Q1 MPP

50.00%

Q1 MIP

0.00%

Q1 Penalty Level

1

Alice has only voted in one of the 2 votes, her Member Participation Percentage (MPP) is 50%. Alice closed the quarter with Level_of_Misconduct of 1.

MPP = AliceVotes / TotalVotesInQuarter

To calculate Alice Q1 rewards, we then use the following formula:

Q1_Rewards% = 1 - (1 - MPP) * Level_of_Misconduct%;
Q1_Rewards% = 1 - (1 - 50%) * 10%;
Q1_Rewards% = 95%;
Q1_Rewards_Alice = ((AliceStake / 4) * (AliceStake * QuarterlyReward)) * Q1_Rewards_%;
Q1_Rewards_Alice = ((10000 / 4) + (10000 * 2.5%)) * 95%;
Q1_Rewards_Alice = 2612.5;

Alice Scenario at the end of Q1

Max Possible

Realized

% of max

2,750

2,613

95.00%

Bob is penalized by Alice's bad behavior. To calculate Bob Q1 rewards, we then use the following formula:

Q1_Rewards% = 1 - (1 - MPP) * Level_of_Misconduct_% * Delegator_Penalty;
Q1_Rewards% = 1 - (1 - 50%) * 10% * 40%;
Q1_Rewards% = 98%;
Q1_Rewards_Bob = ((DelegatorStake / 4) * (DelegatorStake * QuarterlyReward)) * Q1_Rewards%;
Q1_Rewards_Bob = 53,900;

Bob Scenario at the end of Q1

Max Possible

Realized

% of max

55,000

53,900

98.00%

Summary Q2

Alice Q2

Votes

Q2 Yes Votes

0

Q2 Abstain Votes

1

Q2 No Votes

0

Q2 Total Votes

10

Q2 MPP

10.00%

Q2 MIP

0.00%

Q2 Penalty Level

3

Alice is performing poorly, participating in only 10% of the votes.

At Bob's request, DOUGH holders have voted to raise Alice's Level_of_Misconduct to 3.

Alice Scenario at the end of Q2

Max Possible

Realized

% of max

2,750

275

10.00%

Bob Scenario at the end of Q2

Max Possible

Realized

% of max

55,000

35,200

64.00%

Summary Q3

Alice Q3

Votes

Q3 Yes Votes

2

Q3 Abstain Votes

21

Q3 No Votes

2

Q3 Total Votes

25

Q3 MPP

100.00%

Q3 MIP

16.00%

Q3 Penalty Level

3

Alice is performing better, participating in 100% of the votes.

Alice closed the quarter with her Level_of_Misconduct still set at 3.

Alice Scenario at the end of Q3

Max Possible

Realized

% of max

2,750

2,750

100.00%

Bob Scenario at the end of Q3

Max Possible

Realized

% of max

55,000

55,000

100.00%

Summary Q4

Alice Q4

Votes

Q4 Yes Votes

1

Q4 Abstain Votes

0

Q4 No Votes

0

Q4 Total Votes

1

Q4 MPP

100.00%

Q4 MIP

100.00%

Q4 Penalty Level

2

Alice is back on track with 100% MPP.

DOUGH holders have voted to lower her Level_of_Misconduct to 2 based on her performance in Q3.

Alice Scenario at the end of Q4

Max Possible

Realized

% of max

2,750

2,750

100.00%

Bob Scenario at the end of Q4

Max Possible

Realized

% of max

55,000

55,000

100.00%

Bonus Rewards calculation

The bonus is calculated by dividing the number of votes in which the member took an informed position (YES or NO) by the total votes brought to get the Member Informed Percentage (MIP). The final bonus is equal to ten percent of the initial stake, multiplied by the Member Informed Percentage.

MIP_Alice = MIP_q1 + MIP_q2 + MIP_q3 + MIP_q4;
MIP_Alice = 29.00%;
Alice_Bonus_Reward = AliceStake * Bonus_Reward * MIP_Alice;
Alice_Bonus_Reward = 8,678;
Alice_Slash = Alice_Bonus_Reward - AliceStake;
Alice_Net = -1,323

Alice Bonus reward

Staked

Realized

% of Slash

10,000

-1,323

13.22%

Bob Bonus reward

Staked

Realized

% of Slash

200,000

4,900

0%

Conclusion

  • Alice didn't do great on Q1 and Q2.

  • Alice did better on Q3 and Q4 and managed to save most of her stake.

  • Bob was impacted by Alice's behavior but ends net positive.

  • Bob had to make sure Alice was voting enough.

  • DOUGH Holders kept the possibility of exiting the DAO at any time.

  • FLOUR Holders have renounced the possibility of exiting the DAO by delegating in exchange for staking rewards.