Progress pill
Bitcoin Central Values

Finite supply

Bitcoin Development Philosophy

Finite supply

  • Block subsidy and transaction fees
  • Conclusion about Finite Supply
This chapter looks into the bitcoin supply limit of 21 million BTC, or how much is it actually? We talk about how this limit is enforced and what one can do to verify that it's being respected. Moreover, we take a peek into the crystal ball and discuss the dynamics that will come into play when the block reward shifts from subsidy-based to fee-based.
The well-known finite supply of 21 million BTC is regarded as a fundamental property of Bitcoin. But is it really set in stone?
Let's start by looking at what the current consensus rules say about the supply of bitcoin, and how much of it will actually be usable. Pieter Wuille wrote a piece about this on Stack Exchange, in which he counted how many bitcoins there would be once all coins are mined:
If you sum all these numbers together, you get 20999999.9769 BTC.
But due to a number of reasons -- such as early problems with coinbase transactions, miners who unintentionally claim less than allowed, and loss of private keys -- that upper limit will never be reached. Wuille concludes:
This leaves us with 20999817.31308491 BTC (taking everything up to block 528333 into account)
However, various wallets have been lost or stolen, transactions have been sent to the wrong address, people forgot they owned bitcoin. The totals of this may well be millions. People have tried to tally known losses up here.
This leaves us with: ??? BTC.
We can thus be sure that the bitcoin supply will be 20999817.31308491 BTC at most. Any lost or unverifiably burnt coins will make this number lower, but we don't know by how much. The interesting thing is that it doesn't really matter, or better yet it does matter in a positive way for bitcoin holders, as explained by Satoshi Nakamoto:
Lost coins only make everyone else's coins worth slightly more. Think of it as a donation to everyone.
The finite supply will shrink and this should, at least in theory, cause price deflation.
More important than the exact number of coins in circulation is the way the supply limit is enforced without any central authority. Alias chytrik puts it well on Stack Exchange:
So the answer is that you don't have to trust someone to not increase the supply. You just have to run some code that will verify that they haven't.
Even if some full nodes turn to the dark side and decide to accept blocks with higher-value coinbase transactions, all the remaining full nodes will simply neglect them and continue doing business as usual. Some full nodes may, intentionally or unintentionally, run evil softwares, yet the collective will robustly secure the blockchain. In conclusion, you can choose to trust the system without having to trust anyone.

Block subsidy and transaction fees

A block reward is composed of the block subsidy plus transaction fees. The block reward needs to cover Bitcoin's security costs. We can say for sure that under today's conditions with regard to block subsidy, transaction fees, bitcoin price, mempool size, hash power, degree of decentralization etc., the incentives for every player to play by the rules are high enough to preserve a secure monetary system.
What happens when the block subsidy approaches zero? To keep things simple, let's assume it actually equals zero. At this point, the system's security cost is covered through transaction fees only. What the future holds for us when this happens, we cannot know. The uncertainty factors are numerous and we are left to speculations. For example, Paul Sztorc's contribution to the subject in his Truthcoin blog is mostly speculations, but he has at least one solid point (please note that M2, as referred to by Sztorc, is a measurement of a fiat money supply):
While the two are mixed into the same "security budget", the block subsidy and txn-fees are utterly and completely different. They are as different from each other, as "VISA’s total profits in 2017" are from the "total increase in M2 in 2017".
Today, it is holders who pay for security (via monetary inflation). Tomorrow it will be the spenders' turn to somehow shoulder this burden, as illustrated below.
As time goes by, the bearing of security costs will shift from holders to spenders
When transaction fees are the main motivation for mining, the incentives shift. Most notably, if the mempool of a miner doesn't contain enough transaction fees, it might become more profitable for that miner to rewrite Bitcoin's history rather than extending it. Bitcoin Optech has a specific section on this behavior, called fee sniping, written by David Harding:
Fee sniping is a problem that may occur as Bitcoin’s subsidy continues to diminish and transaction fees begin to dominate Bitcoin’s block rewards. If transaction fees are all that matter, then a miner with x percent of the hash rate has a x percent chance of mining the next block, so the expected value to them of honestly mining is x percent of the best feerate set of transactions in their mempool. Alternatively, a miner could dishonestly attempt to re-mine the previous block plus a wholly new block to extend the chain. This behavior is referred to as fee sniping, and the dishonest miner’s chance of succeeding at it if every other miner is honest is (x/(1-x))^2. Even though fee sniping has an overall lower probability of success than honest mining, attempting dishonest mining could be the more profitable choice if transactions in the previous block paid significantly higher feerates than the transactions currently in the mempool—a small chance at a large amount can be worth more than a large chance at a small amount.
Throwing a wet blanket over our hopes for the future is the fact that if miners start conducting fee sniping, this will incentivize others to do the same, leaving even fewer honest miners. This could severely impair the overall security of Bitcoin. Harding goes on to list a few countermeasures that can be taken, such as relying on transaction time locks to restrict where in the blockchain the transaction may appear.
So, given that the consensus on finite supply remains, the block subsidy will - thanks to BIP42 which fixed a very-long-term inflation bug - get to zero around year 2140. Will the transaction fees thereafter be enough to secure the network?
It's impossible to say, but we do know a few things:
  • A century is a long time from the Bitcoin perspective. If it is still around, it will have probably evolved enormously.
  • If an overwhelming economic majority finds it necessary to change the rules and introduce for example a perpetual annual 0.1% or 1% monetary inflation, the supply of bitcoin will no longer be finite.
  • With zero block subsidy and an empty or nearly empty mempool, things can become shaky due to fee sniping.
Since the transition to a fee-only block reward is so far in the future, it might be wise not to jump to conclusions and try to fix the potential issues while we can. For example, Peter Todd thinks there's an actual risk that Bitcoin's security budget won't be enough in the future, and consequently argues for a small perpetual inflation in Bitcoin. However, he also thinks it's not a good idea to discuss such an issue at this time, as he said on the What Bitcoin Did podcast:
But, that’s a risk like 10, 20 years in the future. That is a very long time. And, by then, who the hell knows what the risks are?
Perhaps we could think of Bitcoin as something organic. Imagine a small, slowly-growing oak plant. Imagine also that you have never seen a fully grown tree in your life. Wouldn't it be wise then to restrain your control issues instead of setting in advance all the rules on how this plant should be allowed to evolve and grow?

Conclusion about Finite Supply

Whether the bitcoin supply will grow past 21 million we cannot say today, and that is probably not so bad. Ensuring that the security budget remains high enough is crucial but not urgent. Let's have this discussion in 10-50 years, when we know more. If it's still relevant.
Quiz
Quiz1/5
What mechanism ensures Bitcoin's supply limit is enforced without requiring trust in any central authority?