IDEA: Supply cap & dynamic block reward based on average base fees

There is a lot of excitement around the “ultra sound money” meme, the idea that Ethereum might become deflationary after the issuance reduction that will coincide with the merge (aka the “triple halvening”). Based on my research, Ethereum will be deflationary if gas prices stay at or below 7 gwei.

While I understand the excitement, I believe that Ethereum would be a better form of money if it had a predictable (net) issuance like Bitcoin. Personally I’d prefer holding a currency with an absolute supply cap over a currency that will probably have a net negative issuance but that does not have an explicit supply cap. In other words, I’d prefer holding an asset that is definitely not inflationary over an asset that will probably be deflationary. The fact that only 21 million Bitcoin can ever exist is an incredibly strong meme. In my opinion, the most sound money would have an issuance of zero, neither negative, nor positive. Moreover, it is difficult to denominate debt in a currency with negative expected issuance.

I propose the following:

  • Reduce issuance to 0
  • Compensate miners using the current EIP-1559 burns (I’ll refer to them as base fees), in addition to priority fees
  • Stabilize the block reward by allocating base fees to a base fee reserve, and assign 1/N of the reserve to the validators of the block

N = 1000 (this could also be much larger to smoothen out base fee rewards even more)

Assume base fee reserve is 2000 ETH at block T

Block T+1:
Total base fees in that block: 2
Validator base fee rewards: (2000 + 2)/1000 = 2.002
New base fee reserve: 2000 + 2 - 2.002 = 1999.998

Block T+2:
Total base fees in that block: 3
Validator base fee rewards: (1999.998 + 3)/1000 = 2.003
New base fee reserve: 1999.998 + 3 - 2.003 = 2000.995

Block T+3:
Total base fees in that block: 0
Validator base fee rewards: 2000.995/1000 = 2.001
New base fee reserve: 2000.995 - 2.001 = 1998.994

I would probably also allocate slashing penalties to honest validators, but I left those out for now to keep the proposal simple.

Would love to hear your thoughts on the general idea of a supply cap and this mechanism.