EIP 1559: Wherefore Art Thou?
Update: Since publication of this article, EIP 1559 has been confirmed to be deployed in the London hard fork, slated for July 2021!
TL;DR majority of miners opposed, other stakeholders support
With the extreme congestion seen in Ethereum recently, a topic long-discussed has reared its head once again: EIP 1559. While broadly welcomed by most stakeholders in the ecosystem, it has received heavy push back from a majority of miners. In this article we summarize the issue for the average user, going through:
- Why EIP 1559?
- Miners’ resistance
- Potential compromises
Why EIP 1559?
Better fee estimation
EIP 1559 changes the fee market for Ethereum from the current first-price auction mechanism to one with a base fee plus miner tips. It mainly provides a better estimation of fees, not necessarily a reduction in fees.
The fee estimation problem is common for pretty much any Ethereum user over the last year. You might check your wallet or a gas price oracle and see an estimated gas price of 100 gwei quoted to get a transaction confirmed in 2 minutes. Many of us know this often turns out wrong! Other users may come in afterwards with higher fees, cutting in line ahead of us and causing a long, uncertain wait.
It‘s a bad user experience for everyone — both the user with a stuck transaction (having to wait an arbitrarily long time) and the user who did get a transaction through (paid more gas than necessary).
With EIP 1559, a base fee is always made clear by the protocol. The user can be confident that their transaction will go through smoothly in almost all cases.
Reduce variance in times of high demand
Having a predictable base fee also helps reduce the crazy variance of gas fees in times when there is a sudden spike in demand for transactions — imagine a Banksy NFT sale or an Etherscan token airdrop*!
*for illustration purposes only
Deflationary pressure
As base fee gets burned, total supply of ETH is reduced, benefiting every ETH holder. In a scenario where total base fee paid per block exceeds the block reward, total ETH supply may even be deflationary.
While this is often hyped as a driver of ETH value (lower supply = higher price), in reality the meme itself may be more powerful than any actual deflation.
Locking in ETH as the base currency for Ethereum
With the rise of MEV (Maximum Extractable Value), off-chain transactions where users pays miners in a currency other than ETH becomes more likely. EIP 1559 requires that the base fee be paid in ETH, ensuring that ETH as Ethereum’s currency is locked in to the protocol.
Protects against vulnerability of a network with higher transaction fee than block rewards
A well-documented vulnerability of POW blockchains with lower block reward than transaction fees has been known for some time. The research focuses on Bitcoin, but while Bitcoin‘s halving schedule means it may not have to deal with this issue immediately, Ethereum‘s transaction fees has grown tremendously to the point where this vulnerability may soon become a reality for the network.
Miners’ Resistance
An estimated 63% of miner hash power (as at end of Feb) signals official opposition to the proposal, with 10% in support and 20% unknown.
Why the resistance?
Reduces income for miners
The major sticking point for miners is the burning of base fee. As an example, if base fee were to make up 90% of transaction fees post-EIP 1559 and income followed the February 2021 breakdown, miners would be losing 31.5% of their income.
An important factor to consider is by quantifying the effects of MEV. Research by Deribit indicates that the decrease to miner income will be between 20–35%, as only the non-MEV portion of transaction fees (yellow area in chart below) will be burned. The MEV portion — bidding to be ordered at the top of blocks — will be unaffected by the EIP.
Another angle to look at is whether Ethereum is overpaying for security. Comparing Ethereum with Bitcoin, Ethereum pays 2.5x Bitcoin’s block reward amount as well as higher transaction fees.
Introduces vulnerability to the network
With a drop in miner’s income, hash rate may drop drastically as miners avoid incurring losses. The mining hardware that is unused could then be rented out to entities using them to attack the Ethereum network.
To investigate this likelihood, we take a look at historical drops in ETH price and in block rewards.
On DeFi Black Thursday (12 March 2020), ETH price dropped by 43% overnight. At the same time, total estimated hash rate dropped by only 3%. Even zooming out slightly, the largest drop in hash rate in that time period was only 9%. Similar results are seen when checking other time periods with large price drops (17 June — 16 July 2017, 13–20 Jan 2018, 28 Jan — 5 Feb 2018 and 5 March — 5 April 2018)
Ethereum has reduced block rewards twice in its history: with the Byzantium hard fork in block 4,370,000 and Constantinople in block 7,280,000. In neither case did we see a drastic reduction in total hash rate.
Another factor to take note of is the reputation of the largest miners officially against EIP 1559.
Spark Pool and Ethermine make up a combined 44% of the hash rate on Ethereum. Spark Pool came from EthFans, the Ethereum community in China while Ethermine is an active participant in the Ethereum ecosystem — even receiving a grant from the Ethereum Foundation to create an Eth2 explorer. Having invested heavily in the space it is unlikely that either will stoop to attack Ethereum for a one-time profit.
Potential Compromises
Several options have been discussed to reduce impact to miners and the chances of an exploit of any vulnerability. The necessity of deploying them depends on the community’s overall views on whether miners are sufficiently compensated and the likelihood of exploits.
EIP 969 — removing ASIC miners
Coming from the majority GPU mining side, it is argued that ASICs should be blocked from mining in Ethereum. As the latter are high-performance devices focused purely on mining, they may outmuscle hobbyist miners out in a post-EIP 1559 world thereby reducing the decentralization of Ethereum. On the other hand, their specific focus on only mining ETH means that ASICs are necessarily aligned to the long term health of Ethereum — any successful attack on the network directly affects their bottom line.
Increase block reward
The simplest one to reason — increasing the block reward for miners from 2 ETH to 3 ETH per block. However, doing this may affect the narrative on ETH’s monetary policy. While historically it has only been revised downwards, making this change provides ammunition for naysayers to complain about ETH having an ‘unreliable’ inflation.
l-smoothing
As suggested by the economic analysis made by Tim Roughgarden, an alternative to base fee burning is rewarding the same amount proportionally to miners who mined the prior l blocks (l a large, pre-set constant). This method ensures that miners keep the base fee, but removes the meme-able narrative of a deflationary ETH supply.
Implement in Eth2
Since Ethereum will fully move to Proof of Stake (POS) when Eth1 is merged into Eth2, several miners have asked to delay implementing EIP 1559 until then. POS has been on the cards from the start of Ethereum so it won’t be seen as such a rude shock for miners. The argument against this is that the Eth2 merge may take a couple years, in which time unpredictable fees may drive usage away from Ethereum.
Above is a high-level view of the benefits, resistances to and potential compromises for EIP 1559. To keep up with the proposal, bookmark this excellent resource by Tim Beiko.
What do you think about the proposal? Let us know!
