The deflationary problem

[markm]

I was going to inquire how much gold goes missing each year, but instead I will just quote Meni's nice link from a totally different thread:

See http://en.wikipedia.org/wiki/Crank_(person).

-MarkM-

[Sweft]

I really do not understand how diving satoshi has any meaning to the discussion and flaws that I found in deflationary cryptocurrencies.  

Unfortunate. I'll give it the old college try.

Didn't you suggest an attack that would reduce the amount of coins to 0? Yes, Sweft's Attack I believe, AKA Sweft's Folly. The attack being that the coin would no longer be useful because there weren't enough coins to facilitate trade?

Well, if someone attempts this attack and we start to run low on coins, we can divide the remaining coins in two and you now have twice the supply, which can be used to facilitate trade! The coins which still exist are, of course, more valuable. So the only ones who lose out are the ones destroying their coins during your "attack". The attacker is only hurting himself!

Bitcoin is extremely divisible. I don't have a lot of technical knowledge, but the wiki says "infinitely divisible". The rogue miner better get busy!

I think you're missing the bigger picture here.  As much as you divide bitcoin, it is still deflationary.  The only thing you're doing is moving numbers around.  If you make a 1 a 100000 you still have the same amount.  Value doesn't increase if you move numbers around.

Let me try to explain is to you.

If tomorrow they moved the decimal 2 points right making each coin you have worth 100 coins, the exchange rate will also move in corespodence by decreasing the value of each coin by 100.  Eventually the rogue miner will still take every coin out of existence if he processes transactions and destroys the coins.  Moving the decimal doesn't change anything, the currency is still deflationary.

Why?

In prior scenario say 1 BTC was worth 50$ and miners demanded .01 btc fee which is 1% or 50 cents.

If we move the decimal 2 places to the right.

Now 1 BTC is worth .50$ miners charge .01 BTC per transaction or 1%.
100 BTC is worth 50$ miners charge 1 BTC per transaction or 1%.

The impact to the currency is exactly the same whether .01 BTC fee exists for 1 BTC at 50$ or a 1 BTC fee exists for 100 BTC at .5$.

Regardless of how divisible BTC is a rogue miner will eventually take out every bitcoin in existence unless the currency inflates at a rate greater than transaction fees.

[akspecs]

  To all those people comparing bitcoin to gold - don't kid yourself into thinking they're very similar.  Gold exists in a physical sense.  Of course this has been brought up before.  But it hasn't been taken seriously.

  It is much easier for a bitcoin to disappear forever than it is for an oz of gold to disappear.  Therefore, it is quite possible for there to be a much more rapid rate of deflation than of what there would be with gold / silver / oil or any other supposed to be deflationary assets.  No one can simply claim how the market will react to this deflationary asset because none like it have ever existed.

  This is an experiment.  This is a beta.  Really.  And to be frank, the most recent monetarist experiment of the 20th century failed miserably.

  Bitcoin is meant to be a universal currency.  Not something just for libertarians.

  

[akspecs]

The only think it's lacking is a history of acceptance.

I agree with practically everything you said.  Really.  

I just want to consider everything.  Although I am convinced that inflation is usually very bad for practically everybody (except for *cough*, you know who), I am not convinced that rapid deflation will provide the simple after affect of increased purchasing power.  It can't be that simple.

Looking at the history of gold and silver - there have been plenty of instances (historically speaking) where the volatility in those assets would be unbearable for a conservative investor.  History has shown us that a few entities have the power to play ping pong with those assets.  Manipulation is very prevalent in these markets.  What makes you think bitcoin is immune to this?

In the end, I am very interested in bitcoin's future.  If bitcoin proves to be successful in the long term, I will be the first to defend the statement that the "deflationary spiral" is just a myth.

[MoonShadow]

The only think it's lacking is a history of acceptance.

I agree with practically everything you said.  Really.  

I just want to consider everything.  Although I am convinced that inflation is usually very bad for practically everybody (except for *cough*, you know who), I am not convinced that rapid deflation will provide the simple after affect of increased purchasing power.  It can't be that simple.

It's not simple, and extreme deflation can also be detrimental.  However, just like inflation, the cause of the deflation matters.  In the case of Bitcoin, the vast majority of the deflation occurs because the economy is expanding at that rate.  While deflation can impede that growth, (which is not necessarily bad, btw) it cannot prevent it; for to whatever degree that deflationary effect impeded the growth of the economy also limits further deflation.  It's a self-limiting process.

[Sweft]

The divisibility of coins does not change the deflationary aspect of the currency.  As many time as you divide them, you're still experiencing a deflationary event if someone uses a Sweft attack.  As the miner accumulates more coins, more money is lost. 

The deflationary aspect of bitcoin destroys the vibrancy of the mining ecosystem that is necessary to support a cryptocurrency.

Design a crypto with a minimum 2% inflation and it can survive outside threats.

Let me explain why deflation is bad.

How much transaction fees will support a 2% inflation rate?

If tx fee is .5%, the amount of currency needed to move to get the equivalent of 2% inflation would be 4x the amount of btc in circulation.

Assume tx fee is .1%, the amount of currency you need to exchange to get the equivalent  of 2% inflation is 20x the amount of btc in circulation.

So we assume that when block reward is negligent, that transaction fees will support the economy.

Let's assume the attacker has 50% of network hash.
Let's assume .1% tx fee and 20x velocity of btc.
The miner would in this scenario destroy 1% of BTC a year.

Thus the problem is that without inflation the more BTC that are spent, the more prone to a prolonged deflationary attack.

Thus with a cap on btc we are achieving the opposite of what we want which is spending, not hording.

Why do we not favor hording?

Because miners must profit to secure the blockchain from attack.

Otherwise Sweft's law will make a 51% attack and fork much easier.

[Sweft]

Let's assume we assign a 2% inflation with .1% tx fee and 10x velocity.

The effective mining profit is 3% of BTC a year (2% inflation, 1% fees)


Now an attacker processes 50% of network hash.

Thus he removes 50% of coins generated and 50% of tx fees.

The inflation rate is now 2%-(1%+.5%) or .5%

With a pre-programmed 2% inflation, even with 50% of mining share, the attacker cannot create a deflationary environment as described in the previous post.

Thus the hording and destruction of BTC by rogue agents would be futile because the inflation would ensure more decentralization, more mining profit and a more secure network.

[bitlancr]

So your argument is that Bitcoin is insecure because someone can get all the coins?

Rather than spending all that money on running mining hardware and waiting years to accumulate all the coins, the attacker should just buy the coins directly. Boom, bitcoin network destroyed.

[Bytas]

It's just ridiculous how many times you imply that anyone who does not agree with you is a retard Sweft.

If you are going to make a point and want to be taken seriously, please refrain from offending everyone who you might want to pull into the discussion.
It's a sickening attitude really. There is something called an 'Argument', which is a statement based on facts that sounds completely neutral to anyone who hears/reads it and can be used to convince people without forcing your way into their brain and telling them what to think.

Rule number one of discussion: Never EVER explicitly say "I'm right and you are wrong!" It does not work that way.


Now, as to the content of this topic:

Once Bitcoin matures, the transaction fee will settle in a certain state so that it will still be profitable for miners to mine. Just like any other economic system it will be based on offer and demand. If not enough miners are present, people will increase their transaction fee to get priority over other transactions -> mining will become more profitable -> other people will start mining to get their share of the prize.
An equilibrium will be met.
There will never be a shortage of hashingpower. If there is little hashingpower, people all over the world will start mining because they only need to invest a small amount to get a reasonable percent of the transactionfees in their pocket.

On top of that, its Practically (and mathematically!) impossible to collect ALL the coins by mining transaction fees. Good luck trying completely absorb something by taking away 1% of what is left every time.
100->99->98,01->97,0299->...  
Please go on and tell me when you reach 0 :p

While this "evil person/government" is mining away, he is taking away bitcoin from the market, and while he does that, the worth of bitcoin goes up due to less supply/same demand. In the meantime people still want to use the currency so they divide this infinitely dividable currency(currently 8 digits, more can be added if needed) into smaller parts to trade -> Transaction fees become smaller -> evil dude is having an even harder time to mine them all.
And no, shops will not stop accepting bitcoin because they get worth more.  
If anything, this evil guy is helping the system because he is making transactions cheaper by providing so much hashingpower. (more suplied hashingpower -> cheaper transactions), And as long as he holds all the mined bitcoin, the price will go up, everyone else in the system will get richer. As said before he will never be able to mine all the coins.  

If someone wants to get all the coins, a faster and cheaper way to do that would be to buy them all, and right now that will cost more than one billion dollars. Hopefully by the time bitcoin matures it will be a number somewhere in the trillions. If anyone on the world got that kind of money just laying around and is willing to completely throw it down the drain, go ahead! He will have spend some trillion dollars to buy 21million bitcoins that are now worthless to him and everyone that once owned bitcoins will have their value in Dollars/Euros/... to spend and live from. Everybody wins except for the guy that aims to "destroy us all!"


As to Moore's law. How is it a problem that computers evolve? The better they get, the more powerful the mining-rigs will get, don't forget that mining rigs are computers too.
Why would only bad people take the effort to buy a supercomputer for mining? Moore's law makes the Total hashingpower of the network grow day to day. Its happening as we speak!
Saying that Moore's law will catch up with our hashingpower is like saying that Moore's law will catch up with Moore's law, which is nonsense.



There are some serious flaws in your reasoning that you should evaluate

Kind regards.
Bytas

[Sweft]

Once Bitcoin matures, the transaction fee will settle in a certain state so that it will still be profitable for miners to mine. Just like any other economic system it will be based on offer and demand. If not enough miners are present, people will increase their transaction fee to get priority over other transactions -> mining will become more profitable -> other people will start mining to get their share of the prize.
An equilibrium will be met.
There will never be a shortage of hashingpower. If there is little hashingpower, people all over the world will start mining because they only need to invest a small amount to get a reasonable percent of the transactionfees in their pocket.

I already addressed this numerous times.  If you're having trouble grasping the consequences of a deflationary currency I will explain it once again for those that that yet do not understand.

Mining profit = Block Reward + Tx Fee

In 10 years or less block reward will be negligible.  Thus the sole profits of mining will be Tx Fees.

Your insistence that mining will decrease until mining becomes profitable is only applicable to the ecosystem of miners.  Miners will decrease hash if it is unprofitable to generate hash.  Thus, there will reach an equilibrium between hash and mining profit where miners do profit.

My claim is that this is detrimental to the security of the network.  Any action that requires miners to decrease hash makes the network less secure.  The network is only as secure as the inability of someone to generate a large percentage of total hash.  Thus, to profit, if miners decrease hash, they also decrease security.  It's pretty simple to understand that what we don't want is a scenario where miners have to decrease hash because there is deflation.

The duty of a miner is not only to profit but to continuously increase hash in accordance with Sweft's law.

[mpfrank]

OP's premise makes no sense.  If Moore's Law is continuing, then it is getting cheaper for miners (along with everyone else) to buy computing power.  The continuation of Moore's Law does not weaken Bitcoin's security.  And transaction fees will always be enough to make mining profitable, by the law of supply and demand - once the block size limit is reached, miners will reject transactions that are not large enough to make mining profitable.  If there are not enough such transactions, mining activity (and difficulty) will fall until it is cheap enough to mine that the transaction fees are enough to make mining profitable again.  It's simple economics.  Next, there is no reason to think that transaction fees will come anywhere near 100% in this scenario.  Next, even if coins are destroyed by an attacker, it is not fatal to the economy - Bitcoin is already effectively deflationary (from a price perspective) since its user base is increasing - so additional inflation simply means that the remaining coins become more valuable.  Finally, it is unrealistic to expect Moore's Law to continue indefinitely anyway, since fundamental thermodynamic limits on bit energies (as a multiple of kT) will be reached within the next few decades if not earlier.

[Impaler]

Sweft:  We at Freicoin have long understood the scenario you describe and are in agreement with you on the unsustainable nature of Transaction Fees to support BTC, velocity is far far below 1 so you would need terribly high fees to generate a revenue stream that would replace present mining rates.  So far the response to lower block rewards has been a rise in valuation, which appears to have blow through the mere doubling that would have kept miners at parity and is now an out and out speculative bubble.

The Freicoin solution is to reduce all wallet funds by 5% per year (in tiny amounts per block), all these funds are then payed to Miners.  Total monetary base dose not increase so their is no inflation, after a short initial distribution of 3 years miner rewards remain constant and the network never needs to depend on Transaction fees.  All users essentially pay to secure the network based on their wealth rather then by transaction fees, this is both better economically as well as intrinsically more fair.

We also believe their are a myriad of other detrimental macro-economic effects (Usury) cause by Deflationary currency (putting us radically at odds with the majority of the community that feels it deflation a good thing), and these were our primary motivations, but we've been aware of the death-by-transaction-fee scenario for a long time.  Our website is http://www.freicoin.org/

[Bytas]

OP's premise makes no sense.  If Moore's Law is continuing, then it is getting cheaper for miners (along with everyone else) to buy computing power.  The continuation of Moore's Law does not weaken Bitcoin's security.  And transaction fees will always be enough to make mining profitable, by the law of supply and demand - once the block size limit is reached, miners will reject transactions that are not large enough to make mining profitable.  If there are not enough such transactions, mining activity (and difficulty) will fall until it is cheap enough to mine that the transaction fees are enough to make mining profitable again.  It's simple economics.  Next, there is no reason to think that transaction fees will come anywhere near 100% in this scenario.  Next, even if coins are destroyed by an attacker, it is not fatal to the economy - Bitcoin is already effectively deflationary (from a price perspective) since its user base is increasing - so additional inflation simply means that the remaining coins become more valuable.  Finally, it is unrealistic to expect Moore's Law to continue indefinitely anyway, since fundamental thermodynamic limits on bit energies (as a multiple of kT) will be reached within the next few decades if not earlier.

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.

[Bytas]

On top of that, its Practically (and mathematically!) impossible to collect ALL the coins by mining transaction fees. Good luck trying completely absorb something by taking away 1% of what is left every time.
100->99->98,01->97,0299->...  
Please go on and tell me when you reach 0 :p

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

[Traktion]

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.

[Nesetalis]

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

[bitlancr]

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.

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://bitcointalk.org/index.php?topic=157141.0

[Traktion]

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

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.

[Traktion]

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.

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://bitcointalk.org/index.php?topic=157141.0

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.

[bitlancr]

If a single, cheap, machine had enough hashing power to process all the transactions in the world, then the fees would be tiny.

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