Effects of MACI voting on DAO governance delegation

[JohnGuilding]

@ctrlc03 spoke with Shutter (https://docs.shutter.network/) the other month and they brought up an interesting point. In governance it is very common to delegate voting power to others, and if a vote is run using MACI, the votes are kept private. Users are then not able to prove beyond any reasonable doubt how they voted. In this scenario, how can users that delegate, ensure that whom they delegated to is voting with their best interest in mind?

[JohnGuilding]

This is a major contradiction of how MACI is supposed to work. The point of receipt freeness is that you cannot prove how you voted, so delegates would not be able to prove to delegators how they voted. So this is basically not possible with MACI. For delegation, the delegators would just have to trust the entities they were delegating to if we wanted to keep receipt freeness. Or just vote themselves.

People might start delegating less, and perhaps vote themselves, increasing participation?

What Shutter said is they actually noticed the opposite, private voting reduces participation because people need to actually think by themselves how to vote, and are not influenced by intermediate results. At the same time, the quality of the actual votes is higher as people take more time to think, and make up their own mind etc. Shutter was experimenting with private dao voting - votes are encrypted and decrypted publicly after the proposal ends. So these findings are based on those initial experiments.

[JohnGuilding]

One of my questions was that if delegation is a hard requirement and you didn't want to trust the delegate, could receipt freeness be optional if delegating to some entity?

Ale's response was that removing receipt freeness would downgrade the value proposition of MACI.