Change inflation to not have more go to the proposer

[ValarDragon]

Inflation going through the normal fee process where it gets divied based on number of validators who signed the block, with a slight advantage to the proposer is sub-ideal. This is necessary for fees, but not for inflation.

Instead I think inflation should be divied evenly to all of the staked tokens. This will remove the incentive for proposers to publish blocks with no tx fees, but that is a good thing. If there are no tx fees, then that means there only txs that are from the proposer. Thus it is fine not incentivize this blocks creation, as that would just create larger state bloat / increase sync times.

_Originally posted by @ValarDragon and @cwgoes in #2990 (comment)

[rigelrozanski]

I think this is in the intended behaviour as per slack discussion. Obviously should be addressed with F1 implementation

[sunnya97]

@rigelrozanski I'm pretty sure the agreed upon PoS Economic Design was that block rewards get distributed evenly, and tx fees have a slight proposer extra reward?

[rigelrozanski]

Correct, ultimately, distribution ought to have regard for separation of block rewards (no proposer reward) and fees (with proposer reward) - it just makes the code more messy so currently everything simply goes through the fee distribution mechanism (which was also agreed upon in other discussions in Berkley).

We could do this prelaunch if we so desire, I don't think it's going to make a huge difference though and may just be wasted effort as we should really be doing this at the same time as F1 updates.

[jackzampolin]

This will require a governance proposal but will be a relatively small change otherwise.

[zmanian]

We should have bonuses for including signatures on the commits but no proposer bonus. Social liveness incentives appear to be enough

[cwgoes]

Hmm, I would really like to see a more comprehensive analysis here a la #3529.

(though others may come to different conclusions than I did!)

[gavinly]

I'd like to see math evidence of the proposer bonus accelerating the centralization of stake.

[cwgoes]

I'd like to see math evidence of the proposer bonus accelerating the centralization of stake.

It doesn't, at least not in the first-order distribution, since all rewards are stake-proportional (linearly) and proposer election is as well. @ValarDragon's point is that we probably don't want to pay a proposer reward for proposing blocks which have zero transaction fees (and thus just cost full nodes / validators storage & compute).

[github-actions]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[cwgoes]

This is still applicable, albeit low-priority.

[ValarDragon]

I now agree with Zaki, all we really should be incentivizing is precommit inclusion. I think proposer bonus for everything else is unneeded.

In the abstract, I think its right for fees to have proposer bonuses (and nothing else).

Concretely though, I've now generally changed my mind, that we should be aiming for tx fees to in totality be negligibile worth, and constantly be pushing our scalability to enable this. This counteracts with there ever being sufficient incentive for inspiring validator action in improving mempools (or dis-incentive to not run a mempool at all).

(Granted, I potentially have a hot take here, since unlike ~every other big blockchain ecosystem, I want Cosmos to eliminate MEV, rather than democratize its extraction for better tx-fee-reward-equivalents)

[tac0turtle]

closing in favour of #13139