Thursday, May 16, 2024

Fee-Only Bitcoin

Mining a Bitcoin block needs to be costly to ensure that the gains from an attack on the blockchain are less than the cost of mounting it. Miners have two sources of income to defray their costs, the block rewards and the fees for the transactions in the block.

On April 19th the block reward was halved from 6.25BTC to 3.125BTC. This process is repeated every 210,000 blocks (about every 4 years). It limits the issuance of BTC to 21M because around 2140 the reward will be zero; a halving will make it less than a satoshi.

Long before 2140 the block rewards will have shrunk to become insignificant compared to the fees. Below the fold I look at the significance of the change to a fee-only Bitcoin

Users wishing to transact bid the fee for their transaction in an auction. When demand for transactions is high, fees are high, at other times lower. The graph shows that around the halving there was heavy demand for transactions and the average fee per transaction rose to $127. This is an average, it is likely that the distribution of fees is highly skewed.

The lower the fee, the less likely a miner will choose to include it in the block they are trying to mine, especially at times of high demand for transactions. Low fee transactions can wait in the mempool for a long time. The average delay on 30th September 2023 was 25,810 minutes (nearly 18 days) while the median delay was 10 minutes. Clearly, there was a huge flood of very low-fee transactions.


As I write the average fee per transaction is $3.21 while the average cost (reward plus fee) is $65.72, so transactions are 95% subsidized by inflating the currency. Over time, miners reap about 1.5% of the transaction volume. The miners' daily income is around $30M, below average. This is about 2.5E-5 of BTC's "market cap".

Lets assume, optimistically, that this below average daily fraction of the "market cap" is sufficient to deter attacks and examine what might happen in 2036 after 3 more halvings. The block reward will be 0.39BTC. Lets work in 2024 dollars and assume that the BTC "price" exceeds inflation by 3.5%, so in 12 years BTC will be around $98.2K.

To maintain deterrence miners' daily income will need to be about $50M, Each day there will be about 144 blocks generating 56.16BTC or about $5.5M, which is 11% of the required miners' income. Instead of 5% of the income, fees will need to cover 89% of it. The daily fees will need to be $44.5M. Bitcoin's blockchain averages around 500K transactions/day, so the average transaction fee will need to be around $90, or around 30 times the current fee.

One might think that, were BTC to proceed properly moonwards, this problem would go away. Lets repeat the calculation assuming BTC = $1M in 12 years. Miners' daily income would need to be around $500M. The daily rewards would be about $55M, so the fees would need to be $445M, the same 11%. Thus the average fee would need to be around $900. The problem scales with the "price".

It seems probable that withdrawal of the 95% subsidy on transactions will cause some problems. Indeed, there is considerable economic research making this point, including:
  • In 2016 Arvind Narayanan's group at Princeton published a related instability in Carlsten et al's On the instability of bitcoin without the block reward. Narayanan summarized the paper in a blog post:
    Our key insight is that with only transaction fees, the variance of the miner reward is very high due to the randomness of the block arrival time, and it becomes attractive to fork a “wealthy” block to “steal” the rewards therein.
  • More generally, the analysis of 2018's The Economic Limits Of Bitcoin And The Blockchain by Eric Budish essentially concludes that, for safety, the value of transactions in a block must be less than the sum of the mining reward and the fees it contains.
  • In 2019 Raphael Auer of the Bank for International Settlements published Beyond the doomsday economics of “proof-of-work” in cryptocurrencies:
    The key takeaway of this paper concerns the interaction of these two limitations: proof-of-work can only achieve payment security if mining income is high, but the transaction market cannot generate an adequate level of income. ... the economic design of the transaction market fails to generate high enough fees. A simple model suggests that ultimately, it could take nearly a year, or 50,000 blocks, before a payment could be considered “final”."
The last time I wrote about this issue was in 2021's Taleb On Cryptocurrency Economics.

1 comment:

Tardigrade said...

So basically at some point the public blockchain has to become at least a semi-private blockchain. Presumably controlled mostly by financial corporations (such as banks) which cover the cost of running the network with sundry fees charged to clients for a variety of services.

Or transition to a proof-of-stake system, which is effectively the same thing as the above but substituting billionaires and private equity firms for the banks.

Either way the anarcho-capitalist philosophy behind the origination of bitcoin is thoroughly dead, for bitcoin at least.