Topic: The deflationary problem - page 8. (Read 32501 times)

I predict that by 2040 there will be so many transactions that people would rather not get a reward for their hashing out a block, just for the transactions. I cant prove anything, but that is what I believe.

That's very dangerous because there is no way to know when it has peaked and there's always a risk that the price will fall before you can sell the oil you held back. Certainly some people will do that, but it's not a universally winning strategy.

The problem is that the predictable aspects of the future price are already baked into the present price. If the price wasn't rising fast enough, people would hold back oil, and the price would just rise faster.

All of these affects are already cancelled out in the present price. The present price must be fair to sellers, or it would rise. The present price must be fair to buyers, or it would drop. For any reasonably-liquid commodity priced in any reasonably-liquid currency, the price will be tend towards the price that is equally fair to the typical buyer and the typical seller.

If I managed an oil company and the price of oil was going up very quickly I would of course try to profit from the rising price by not selling to the market until the price has peaked.
Since oil demands huge volumes of storage It would be much easier to just deliver the oil to the customers, and instead buy one future contract for every barrel deliverd until our analysts would say the bubble is about to crash. (summer of 2008, oil peaked at 145$/barrel and crashed 80%).

Maybe you'll see it more clearly if you look at the same argument in the context of an inflationary currency or good that's appreciating in value. Suppose the price of gas has been rising continuously and fairly predictably, either because the currency is dropping in value or because the demand for gas is increasing. Are gas stations going to close down for a few weeks so that they can sell gas for higher prices in the future?

The answer is clearly no. Because if they were going to do so, that would simply mean the price of gas wasn't high enough *today*. You don't see a gas station closing because they'd prefer to sell their gas later when the price is higher. The price is already high enough -- supply and demand ensures it.

The same is true of a deflationary currency. If it's actually encouraging hoarding, that would simply mean the price of goods wasn't low enough *today* to encourage people to buy them and the price would fall until it is low enough. But of course, that will have already happened -- having a deflationary currency means the price of goods has dropped.

And it is not right. For someone to be rich, someone has to be poor.

I guess that even bitcoin has a natural point of equilibrium after which it will only deflate by the yearly value that its economy will produce.
If for example at equilibrium all bitcoin users have a total of 10 houses and someone builds one more, the total value of the 21 mil bitcoins will become equal to eleven houses and ALL bitcoin users will be richer by a house divided by their percentage of the total bitcoins.

Ok, so let's assume everything is constant except for Moore's Law. By Moore's Law, the cost to produce a fixed level of hashing power will decrease over time. With all other things being equal, mining becomes more profitable. Now other potential miners will take note of this, so more miners will join, incentivised by the increased profit margin. Thus - all other things being equal - the hashing power will increase with Moore's Law.

I think your point about the price of Bitcoin is right - but that should be obvious. Bitcoin being less valuable means exactly that - it's less valuable, therefore it's only right that less hashing power is devoted to it. In this case, less hashing power is the effect of Bitcoin being less valuable, not the cause.

There is a limit on bitcoin because the person who made it never intended it to be used the way this community has decided it should be used. Kind of ridiculous to ignore that fact and try to rationalize the nature of BTC away and make it fit nicely into your ridiculous internet libertarian world view.

That's not true. If it were, no one would hold inflationary cash, as they would be spending it today, rather than lose value tomorrow. It would also mean that no one would buy any computers.

Just as contracts can account for inflation, they can also account for deflation. The only real difference is psychological - more stuff has to be better than less stuff, right?

There is nothing to stop people taking out inflation rate trackers, which can go negative. That is, the repayments get reduced during a deflation. The real value remains positive regardless.

Deflation will kill bitcoin businesses - why spend your bitcoins today if they are worth more tommorow, or next week, or next month...
Bitcoin simply becomes an asset to speculate in or hedge against inflation instead of being used as a trading currency.
The best thing that can happen to the bitcoin community is a crash followed by a stable price for a long time so people get used to spend their coins instead of just hoard the coins.

One solution to the deflation problem is that if everybody votes for a change in the bitcoin protocol so a 2%  inflation can go on for ever even after 21 million coins have been mined.

Yes, understood. However, this is just asserting that if the hash rate falls, the difficulty will adjust. The problem isn't the difficult level, but the falling hash rate (reducing the barrier against a network attack).


Sure, I'm fine with this too. I asserted that 'if the transaction rate becomes stable', then so will the competition for block space. In turn, this implies a stable hash rate (fluctuations considered, ofc).


I asserted that 'mining ceases to be profitable' as a boundary condition. I'm not convinced that miners drive the price of Bitcoin; It's the people who value Bitcoin that dictate the price. If it becomes too expensive to mine at the market rate, mining will not be done. Therefore, I'm not convinced that miners will necessarily step up to increase the hash rate.

To conclude, I am still unconvinced that this thread is moot. The hash rate needs to increase proportionately to Moore's Law and I'm not certain this will happen automatically (without some relatively minor alternations).

I'm happy to be proved wrong though, if you think I'm still missing something.

The problem is in this statement:


The hash rate is technically independent of transaction volume. If the hash rate were to drop to 1Mhash/s right now, then the target difficulty would soon follow. Bitcoin is set up to produce a new block every 10 minutes, regardless of hash rate. Thus block space remains constant.

Now, given that block space is a finite resource, there will be competition. More competition (more transactions) will drive up fees, which will make mining more profitable, thus increasing the hash rate as more miners join in to get that profit. Conversely, less transactions means less competition, less fees, less mining, reduced hash rate.

Now, when hardware becomes cheaper, mining becomes more profitable. Thus more miners join in to get that profit. Any guesses what that results in?

Bottom line: as long as there is competition for block space, this entire thread is moot.

Assuming that:

- Transaction rate becomes stable.
- Mining ceases to be profitable.

Why would there be increasing competition for block space (in turn, an increasing hash rate)?

Logically, the hash rate only needs to be sufficiently high to process all of the transactions requested. If hardware to provide said hash rate continues to fall, at some point the cost of hardware to do this will fall to a price where it is economical to attack the network.

Apologies if I am misunderstanding something technical which is critical to this debate.

No - block space is independent of hash rate. Competition for block space => increased fees => increased hash rate.

I don't think that is the problem (although I haven't read that thread yet - will do later).

The problem isn't that there would ever be insufficient block space; the problem is that providing sufficient block space may become too easy. That is, the hashing power needed to confirm the transactions may get so cheap that it would be economical to attempt to take over the network.

If a single, cheap, machine had enough hashing power to process all the transactions in the world, then the fees would be tiny. Equally, anyone with a couple of servers could take over the network. Ofc, this is an extreme example, but I think it is what the OP is concerned about. I would assert that it is a valid concern too, but one that can be rectified.

Do you mean a floor on the price? If not, I'm not sure what you mean.

If you have a minimum fee as a percentage, then there should be no hard floor on the price.

Tragedy of the commons doesn't apply as long as there is sufficient demand for block space (transactions), and some form of limit on block space to facilitate competition.

See this ongoing thread here:
https://bitcointalksearch.org/topic/funding-of-network-security-with-infinite-block-sizes-157141

The problem with a minimum fee, is it puts a floor on the exchange rate.

I've read the initial assertion and it seems that the primary problem is of Moore's Law resulting in hashing power getting cheaper, which could threaten a static/shrinking Bitcoin network.

Surely this has a simple solution - agree a minimum fee* (a percentage, perhaps - TBC) and apply it to the software. This doesn't need to be decided now, as it would only become useful when the total hashing power starts to flat line. Adding a minimum fee wouldn't fork the block chain either.

Why a minimum fee? It would mean that as hardware halved in price, the amount of hashing power could be doubled to soak up the return on the minimum fee. Therefore, hashing power would scale up, at a similar rate to the drop in hardware costs.

Additionally, the fact that you can't charge a fee over 100% is irrelevant. You don't need an ever increasing hash rate [EDIT: relative to Moore's Law], you just need a total hash rate which is large enough to make network attacks uneconomical (which means tracking Moore's Law - see above).



EDIT: * To add, the reason why a minimum fee would be preferable to a market solution is due to the 'tragedy of the commons'. If same miners start dropping their rates to catch more transactions, then it will price out those with high rates. Setting same sort of minimum fee, for the good of the network, would seem appropriate to combat this. There may be alternatives to this though.

I found a cool comic that is kinda related to the 'mining all the coins via transaction fees' thing i said earlier.
+ It's kinda funny

What he said!

I'm not at all worried that "Sweft's Law" will ever come into play, nor that it is even correct in the first place.