Can you spot the flaw in this argument?
1) Over time, Bitcoin will become increasingly important as a global currency.
2) Therefore, over time, the price of Bitcoin must rise.
The first is a statement of (alleged) fact and hence not (in its raw form) susceptible to logical analysis. This is not a post about whether it is true, probable, possible, improbable or false. I’m just going to accept it for the sake of argument.
The problem occurs at the word “Therefore”.
I suppose that what underlies the “therefore” in some people’s minds is some instinct like this: “Prices have to equilibrate supply and demand. With supply fixed and demand growing, the price must rise”.
Here’s exactly where that argument goes wrong: Bitcoin has two relevant prices: The price of the asset itself and the transaction fee you pay every time you use it. If the transaction fee rises to equilibrate the supply and demand for transactions, there’s no obvious reason for the price of the asset to rise.
In other words: Yes, it is true that fixed supply plus rising demand implies a rising price. But when there are two prices, it’s not immediately obvious which one must rise.
I can easily imagine a world like this:
1) The demand for Bitcoin transactions is extremely high.
2) Therefore the transaction fee is extremely high.
3) Therefore people use Bitcoin only for large transactions. Large transactions tend to be foreseeable. (I don’t buy a car without first thinking about it for a while.) Therefore there’s plenty of time to acquire Bitcoin just in time to make my transaction. Otherwise, there’s no reason to hold it. (And plenty of reason not to, just as there’s plenty of reason not to hold any other non-interest-bearing asset. For example, if you’re hedging against inflation, you can buy interest-bearing inflation-indexed bonds.)
4) Therefore, at any given moment, the demand for Bitcoin is quite low, although the demand for transactions is quite high.
It seems to me that much of what I read about the supply and demand for Bitcoin starts from the assumption that we can model the demand for Bitcoin the same way we model the demand for dollars or euros. But neither dollars nor euros have transaction fees, and that seems to me to make a world of difference.
Am I missing something?
I think you’re completely on the money here.
One of the things I find fascinating about bitcoin and to some extent all cryptocurrencies is that right now the actual transaction fee is hugely if not nearly entirely subsidized by other people. Any time a random American buys or sells BTC, they’re free-riding an extremely expensive infrastructure that is largely being run by… well, who it’s actually being run by is a fascinating question, but it seems pretty clear that the vast majority of bitcoin transaction processing (“mining”) is being done in purpose-built datacenters somewhere in the People’s Republic of China.
For now, whoever is running those datacenters seem to be pretty convinced that it’s worth the money. (One suspects that for a number of reasons their real costs are extremely low compared to doing it anywhere else in the world.) I wonder what will happen when and if that stops being the case?
People do not just buy Bitcoin for transactions. Many, and maybe even most, buy it as a means for stored value. For this, it competes with buying gold.
As you say, there are reasons to avoid non-interest-bearing assets. But there has been a market for gold, above and beyond jewelry and industrial uses, for centuries.
It is hard to estimate how many investments I have missed because I expected them to be logical. One of my friends made a small fortunes with investment strategies that were economically unjustified.
The bitcoin fee seems to be assessed by whomever is handling the transaction. The size of the fee depends on demand, but also on the size of the transaction in bytes.
The size can depend on how many transactions a coin has gone through in the past, since those are checked every time a coin is used in a new transaction.
The transaction, I assume, will be priced at whatever the market will bear, and will have to factor in the fact that blockchains have limited capacity. So we have these odd situations where average transaction cost went above $60 back in April, but then fell to less than $10 a few weeks later.
Since I was worried about the oddities of the economics, I stayed away from bitcoin back when it was at $32 per coin. I lost out on a huge opportunity, simply because I wanted my investments to be rational.
Isn’t there a more straightforward error here? If it’s public information that “Over time, Bitcoin will become increasingly important as a global currency”, and if this were likely to cause the price to rise in the future, then by the efficient market hypothesis that would already be reflected in the price in the present.
The Bitcoin transaction fee is set by the person making the transaction, and its purpose is to make the transaction in question enticing to miners. Basically, Bitcoin transactions compete for the attention of Bitcoin miners, who must assemble blocks of transactions to work on for inclusion in the block chain. Since miners get to keep the transaction fees associated with the transactions in a block that they finish, they prefer high-fee transactions. Currently the average transaction fee for a successful Bitcoin transaction is around USD 5. If the fee on your pending transaction is a lot lower than that, don’t expect the transaction to show up in the official block chain anytime soon.
Right now, Bitcoin miners make money not only via transaction fees but also because they are awarded a bounty (in Bitcoin) whenever they add a block to the official block chain. This bounty, however, is scheduled to dry up eventually, and after that point of time the only way for miners to make money will be the transaction fees. Since miners must pay for up-to-date hardware and ginormous amounts of electricity just to stay competitive, expect the already-steep transaction fees to rise even further.
This, among other things, illustrates why the idea of Bitcoin as an everyday “currency” is laughably absurd.
Anselm (#5): I believe you have the economics wrong.
As long as the supply of transactions per day is fixed by the technology, the transaction fee is determined entirely by demand. To put this another way, the number of miners has no effect on either the supply or demand for transactions, so has no reason to affect the transaction fee. Instead, what will happen when the seignorage disappears is that mining profits will turn negative, leading many miners to exit from the industry, with exit continuing until the only miners left are those efficient enough to get by on transaction fees alone.
It is possible that this will leave so few miners that the whole system collapses, but unless that happens, the system will go on as before, just with fewer miners. Transaction fees might rise or fall for other reasons, but not for this reason.
I don’t think Bitcoin can collapse from lack of miners. The system adjusts the difficulty of the problem miners must solve dynamically according to the amount of mining power available, and there’s nothing wrong in principle, protocol-wise, with having just a single miner (presumably this is how Bitcoin got started in the first place on Satoshi Nakamoto’s machine). It wouldn’t be great for the popular notion that Bitcoin is a decentralised system that no single entity can control, but then again the majority of the currently available Bitcoin hashing power has been concentrated in a small oligopoly of very big mining concerns for some time, so it’s not as if Bitcoin was meaningfully decentralised right now.
The other thing is that transaction fees are pretty high now (at least when you’re contemplating using Bitcoin to buy a cup of coffee), but then again transaction fees only come into play if your Bitcoin transactions, in fact, do happen on the block chain. Since there is way too little room on the block chain for Bitcoin to work as an actual currency (7–10 transactions per second, max), people are looking for ways to do “Bitcoin-but-without-the-block-chain” in order to speed up transactions and make them cheaper (never mind that this undermines one of the main selling points of Bitcoin in the first place). This includes, e.g., the mythical “Lightning Network”, which is impractical and doesn’t seem to work. If – and that is a very big “if” – eventually somebody figures out how to do this right and can convince enough people to go along with it, Bitcoin-mining transaction fees may be much less of a factor in the future. I wouldn’t hold my breath, though.