Transaction Fee Mechanism Design - Optimality is Provably Possible

There are a growing number of papers in ETH research circles that claim it is impossible to build a socially optimal transaction fee mechanism. The working paper linked at the address below proves this is wrong - it is achievable.

https://saito.tech/socially-optimal-transaction-fee-mechanism-design/

Sharing here because although the paper is somewhat academic the proof itself is less than 2 pages and should be accessible to anyone with undergraduate economics. It uses the equation for the "utility possibility frontier" developed by Paul Samuelson and Leonid Hurwicz in their Nobel-winning work on optimality of public goods provision.

There are two reasons this matters for Ethereum. The first is that it shows that unless a fee mechanism is pareto optimal it can never be incentive compatible, as otherwise there will always be a coalition of participants who can improve their utility by adopting the "byzantine" strategy of paying a different fee. This is worth noticing because the papers that claim this is an impossible problem to solve are proposing modifications to the fee design that make it impossible to achieve pareto optimality, which means they don't understand the underlying problem. Having more people understand the underlying problem is the best way to provide pushback against bad ideas getting traction.

The second reason is that the paper proves there is a very specific technical property that needs to exist for optimality to happen -- block producers must be incentivized to share unconfirmed and fee-bearing transactions with their peers.

There is a lot of research being done on improving the ability for nodes to increase their profits from hoarding transactions, such as optimizing MEV. A lot of the time this is being done under the guise of "funding infrastructure" but it is harmful for the network.