A congregation of sorts, or a 21st century town hall meeting, is to take place tomorrow at 2PM UTC where a livestreamed discussion will hopefully lead to an answer on whether ethereum’s issuance is to be reduced and if so by how much.
The matter arises due to the necessity to delay the difficulty bomb following a slight change in the implementation of Proof of Stake (PoS), which was expected this summer but now is estimated for summer 2019.
That delay changes, for the second time, the foreseeable level of ethereum’s supply, with Vitalik Buterin’s eth supply estimate last year missing by some 5 million eth, or $1.5 billion at current prices.
If it continues at current levels, then ethereum’s supply will increase by more than ◊7 million within a year, or about $2 billion.
That’s nearly 10% of its current market cap and does not include transaction fees which go to miners.
Such fees are at ◊600 for the past 24 hours. If we extrapolate, then miners would receive an additional ◊215,000, or $60 million, within a year.
Those are all very big numbers paid by ethereum holders with a proposal on the table to reduce them by 35% or 70%.
One reason for this proposed reduction is because ethereum plans to remove miners completely and replace them with staker which only require eth to engage in validation rather than GPUs and Asics.
To make that transition smoother, thus, whereby block rewards will fall to just ◊0.82 once PoS Casper, a reduction to 2eth per block has been proposed or 1eth, with the latter being a favorite of token holders according to a vote.
Token holders, or someone to represent them, have not been invited to this meeting, presumably because some devs are themselves holders and presumably because it is difficult to see who exactly can represent them.
Miners, on the other hand, have been invited and we naturally expect them to be against the proposal, but there is an opportunity here to seek clarification from miners on certain aspects.
One reason why supply is so high is because uncle rates are very high. Why? Miners would probably say because the network is being used a lot, but eth’s network, at full capacity, was running at about 1MB per ten minutes.
When bitcoin was running at 1MB and even above, orphan (uncle) rates did not likewise rise. That’s because miners in bitcoin do not get paid for orphans. Their endeavor to increase profits, thus, has led to a very efficient propagation network in bitcoin.
In ethereum, miners do get paid for uncles. Less than an actual block, but its still some 2 eth per uncle and 3eth per proper block.
The block propagation method in ethereum, thus, is very inefficient with miners not taking any steps to reduce uncles, in contrast to bitcoin where they have compressed a 1MB block to just 2kb, better than dial-up.
If eth miners did the same, capacity would probably be a lot higher. Their incentives are however misaligned, at least to some extent, with those of network users and eth holders. First, miners want higher fees, so they limit capacity. Second, they want higher uncles, so they increase supply.
Yet miners are needed for now, so just how much should they be rewarded for keeping those blocks moving?
The answer to that, some say, is a political question, but in many ways it is a technical question of how much security is sufficient.
A question which doesn’t have an easy answer because miners in many way think only in fiat and price of course can change up or down.
Yet one somewhat objective parameter that can be looked at is the amount of eth moving a day. Currently that stands at 1.6 million eth or just 1.6% of the market cap.
For bitcoin, it is at 809,000 btc, more than their yearly supply of 700,000 btc. While for ethereum, their yearly supply is 4x more at above ◊7 million.
If we look at that parameter only, then ethereum’s supply can be reduced by 75% and still be just about as secure as bitcoin.
Another parameter we can look at is the amount sent per hour. For bitcoin, that was about 34,000, with miners there receiving 75 btc as a reward, or 0.002% of the amount moved.
For eth, it stood at about ◊68,000, with miners receiving ◊816 in total, including uncle rewards, or about 0.012%.
That suggests eth is paying more than bitcoin for its security by about 6x. So a reduction to 2eth would probably be quite safe as far as security is concerned, as may be a reduction to 1eth which would give ethereum about the same level of security as bitcoin.
The only political question here thus might be whether miners should be paid more than is necessary to secure the network, and the answer to that, accounting for some leeway as these are estimates, is to ask why should they at this stage?
Previously, a high inflation rate was necessary to distribute the coins, but now with supply overshooting what was foreseen by quite a bit, distribution is probably not part of the equation.
Yet what might also be up for discussion at the meeting is whether GPU miners should be given priority over Asics miners. Something which is a political question because you are picking winners and losers.
One matter to take into account there is that anyone can buy asics as Bitmain, and other asics manufacturers, sell them to anyone who wants to pay for it. Moreover, new GPUs now are getting closer and closer to asics capabilities.
So that would probably not achieve much and it wouldn’t be an easy endeavor to fork off asics because in ethereum they are fairly similar to GPUs. The code therefore would certainly not be ready for the Metropolis fork which is planned for just before Devcon in October.
For issuance reduction, devs said it is easy to implement it, so they can do it quickly, but now with just weeks to go, a decision does need to be made so that they can start testing because no one wants a repeat of the Devcon Shanghai opening in 2016 when a debugging night preceded the conference keynotes.
They haven’t said whether a decision will be made in this meeting, but there isn’t much left now to reach a conclusion.
Holders have made their views very clear. Miners would probably oppose or might try and delay by focusing on asics, but general consensus does appear to favor a reduction to at least 2eth and maybe even 1 eth.