[anologwintermut]

The fact that Bitcoin is deflationary is purely a product of the software/protocol. Given consensus to change it, one could make it inflationary easily. E.g. by having blocks always pay out 25 btc as a reward or even have successive blocks pay out increasing rewards.

[nhaehnle]

It would still be inferior to modern fiat currency from a macro-economic perspective, because whatever your rule is going to be, it will not adapt to the development of the economy.

Think of it this way: The abundance of money should somehow correlate with the abundance of real resources. When this correlation breaks down, bad things happen to the economy.

In modern fiat currencies, the supply adapts automatically, without government intervention to the development of the economy, via the mechanism of bank loans. When a bank makes a loan, the money supply grows. When a loan is paid back, the money supply shrinks.

Bitcoin cannot work that way by design, which makes it a bad choice for a common-use currency - at least from the perspective of macro-economic behavior.

[aaren]

I hadn't thought of that before. So bitcoin is like gold in this regard, when gold was used as currency.

It occurred to me that bank notes today are the ancestors of a receipt that promised you a certain quantity of gold. Cash is just a promise - an abstraction from the underlying thing of value.

Whilst bitcoins are limited, you can create as many promises as you like - you just need an extra layer on top of bitcoin. Does this solve the loaning problem?

You would still have bitcoin as the base, like the gold standard but with none of the physical limitations of actual gold and with people still being able to easily transact in bitcoins.

I'm not sure whether having two things that you can pay with - bitcoin / promise of bitcoin - would be implementable. Would it require another crypto layer on top of bitcoin, with some parallel blockchain?

Gold standard: http://en.wikipedia.org/wiki/Gold_standard

This is an interesting read about bank notes: http://www.bankofengland.co.uk/banknotes/pages/about/history...

[nhaehnle]

I think it's possible that introducing reserve banking on top of Bitcoin could solve the lending problem for a while, just like it solved the lending problem on top of gold - for a while.

In the end, though, the same problems would crop up again after a while: bank runs, the attempt to deal with bank runs via a central bank, the tendency for downward spirals in countries with a trade deficit, and so on. Eventually, countries would have the same incentives to get out of a "Bitcoin standard" as they had with the gold standard.

[jpadkins]

> Think of it this way: The abundance of money should somehow correlate with the abundance of real resources. When this correlation breaks down, bad things happen to the economy.

citation please. The abundance of real resources in the cpu, memory, and storage markets have been outgrowing the dollar supply for several decades (evidence by the steady drop in prices even with a much larger dollar + credit supply). Nothing bad has happened to those markets. The growth of real resources varies wildly from resource to resource. If your first statement was true, shouldn't we see see resources that are not correlated with the dollar supply have 'bad things happen'?

Also its almost impossible to measure 'abundance of real resources', so any fiat currency pegged to that metric will not be aligned.

[anologwintermut]

To be clear, I don't think bitcoin that good a currency either. I wanted to point out that the deflationary aspect is an arbitrary (and I think stupid) restriction.

You are right, modern fiat currencies can adapt and that is the main thing bitcoin lacks. My understanding was that happened mainly as a result of central banks who can adjust all kinds of parameters. This is antithetical to Bitcoin's entire design principles, let alone the underlying technology (which could me modified to accept some parameters modifications from a board of governors).

Normal banks can't do that. To the extent that they do cause things to happen, can't they do that with Bitcoins. I.e. issue loans and collect payments with them.

[robotcookies]

But you're only looking at bitcoins as a replacement for fiat. What if it worked in conjunction alongside the dollar. You can have both. Just like gold is not used as currency but alongside printed money.

[nhaehnle]

I think that's quite a plausible future for Bitcoin, though I find it hard to give an estimate of likelihood and of the relative relevance that Bitcoin will have.

The next-most plausible (though I think much less likely) future is that the hype will die down, there will be widespread disappointment, and it will turn into a kind of ghost town.

One big problem of Bitcoin is that in terms of efficiency, pretty much any centralized service can beat the shit out of it. If you think of Bitcoin as a ledger, and then look at the amount of resources spent on maintaining that ledger, and compare it to the amount of resources required to run a simple centralized ledger... well, Bitcoin loses big time.

This means that whenever Bitcoin comes close to having a large mindshare for everyday internet transaction, there is a serious window of opportunity for non-Bitcoin startups to take its place.

The only market where this logic doesn't apply is the Silkroad-type market and the stuff that caters to gold-bug/libertarian ideologists.

[icebraining]

I agree with you that Bitcoin is an expensive ledger to maintain. That said, it's nonsense to say that low transaction costs (that's what efficiency means to the end user) is the sole factor for anyone besides Silkroad users and ideologists.

People pay a lot for security, and the problem with centralized systems is that they grant the organization running them both the means and incentives all the rents it can get away with, not only directly with added fees, but by using abusive tactics to pass the costs of fraud onto the user (see PayPal's infamous freezing policy) and by skimping security as much as it can.

Of course, you can reduce the likelihood of such events by adding external auditing, insurance, regulation, etc, but all of that adds friction which will inevitability result in increased transaction costs (TANSTAAFL).

If risk was irrelevant, nobody would invest in treasury bonds.

[Guvante]

On a related note: It boggles my mind when people bring up Bitcoin as a way of doing distributed work, it is nearly the exact opposite of how you would want to do distributed work, increasing the duplicated workload tremendously.

[icebraining]

I'm probably missing something, but why can't Bitcoin work that way? Say I deposit 10 BTC in Coinbase, which records that as a liability in their books, and then they lend out 8 BTC to someone else. Isn't that the definition of fractional reserve banking?

[Jtsummers]

Fractional reserve banking effectively creates more money. Assume you have to keep 20% available, you can lend out 80% of your deposits. You lend that out, but your depositors still have access to "spend" their money, so you've increased the money supply by 80%. Now suppose the borrowors repeat this process at a second bank, and that bank loans out 80% of their deposit, we now have 164% extra money in the system, all available for spending.

If you try the same with BTC, it's not possible. You give me 10BTC to hold, I loan out 8 BTC, you can't spend from your 10 anymore (well, 8 of it). The total currency in the system is unchanged, just who can use it has changed.

[icebraining]

You lend that out, but your depositors still have access to "spend" their money

They have access to their money because the bank keeps a reserve that is enough for covering regular activity, but if the depositors as a whole tried to withdraw more than that, the bank can't simply invent USD with which to pay its depositors. That's why banks default when there's a run.

And of course, Bitcoin banks could also keep a reserve of BTC, so I really don't see why you can't have fractional reserve banking with Bitcoin.

[Jtsummers]

Only partial for that reason. If everyone demanded all their money, yes, it would break the bank. But if 30% of the money were requested, and only 20% were held in reserve, the bank can still act as if it has the money. There's no similar mechanism for bitcoins.

[icebraining]

Sure, but that doesn't mean that new money can't be created with a Bitcoin-based fractional reserve banking system, it just means they wouldn't have a central bank acting as a lender of last resort - though they could still get loans from other banks.

[trts]

Even though it's deflationary, I don't see that as necessarily a problem since it is infinitely divisible, which fiat currency is not. It may just take time for people to start thinking in hundredths and thousandths instead of hundreds and thousands. If Bitcoin becomes less volatile in the future and deflates at a predictable rate, then I could see it being used more as currency.

Interesting to think of the reverse of "when I was a kid, a Snickers bar only cost 5 cents," e.g. "what I just paid for my new Macbook only bought me a pair of shoes ten years ago."

[nhaehnle]

Deflation, meet downward rigidity of wages (and other prices).

Basically, the problem with deflation is that everybody who has only one (or very few) income stream, but at the same time many expenditure streams (especially fixed contract expenditure streams) is going to fight very hard against reduction in that income stream.

And empirical data suggests that that fight tends to be fairly successful, too. This is why deflation is poison for economic activity.

[tbrownaw]

And empirical data suggests that that fight tends to be fairly successful, too. This is why deflation is poison for economic activity.

I suspect that demanding constantly increasing income would be rather less successful if the system weren't rigged to make expenses never decrease.

[nhaehnle]

The empirical data is not about constantly increasing income. It's about a number of things, but the most striking is that in a deflationary environment, wage (and other price) changes have an abnormally large peak at a change of zero; this peak does not fit into any simple assumption about how price changes should be distributed (in particular, it is clearly not a normally distributed effect).

[herge]

It's a lot easier to give someone a raise than to cut someone's salary, even after explaining them all about how 1 bitcoin is worth more now than when he or she started working for you.