Regression theorem & Bitcoin revisited

[hazek]

The cost to make something does not give it value.

I haven't read all the stuff here, but just about that: I would argue, that in the case of gold, bitcoin & co, it boils down to reduction of entropy. for gold, you have to find and combine all the tiny bits (atoms) to bigger chunks. That reduces the scattering, i.e. entropy. this needs energy! bitcoin: the low hash value is one of the purest forms of entropy reduction you can get. also needs energy.

so, only the cost of making something doesn't count, but it counts if its goal is to make something "special" or "unique".

Actually this is a really good point. Gold that is still in the ground isn't worth anything. It's only when it's dug up that it gets it's value because now someone put effort into transforming that lump of rock into something more pure and more beautifully shaped that the person who dug it up and other people now value. And this is as true today as it was back when the only use gold saw was as jewellery  (imagine some guy finding a pretty rock and coming home and giving as a present to his wife saying "look what I found, isn't it pretty..).

The same could be argued about the calculated hashes.. By themselves they are meaningless and worthless but it's because someone did the calculations and found those numbers and gathered them into a blockchain that they were valuable to them and then to others.

[Adrian-x]

But, as I said, bitcoin is not a consumer good, nor does it have any connection (present or historical) to any consumer good.

This quote from the OP's original post is the key, when reviewing the regression thrum in reverse from its origin through time; you see the only real money is a commodity that has a measurable benefit (something like wheat / corn) and all other money is a derivative of that.

The transition from a corn to gold was most likely a market driven transition. Arguably the food commodity money transitions to a token system like gold was a messy process and took generations. That jump is very similar to the transition from Fiat to Bitcoin crypto currency.

A side note: the original lure of gold was most likely it reflected excess recourses in a community, and as a result gave one a symbol for "food prosperity" hence its desirability and people willing to pay in life sustaining recourses to attain it.  (bitcoin's  comparison to digital Jewellery is appropriate in my opinion)

In support of the regression thrum, Adam Smith gives us insight as to how the value of mined metals were derived, during this transition, effectively you needed excess production in corn to invest in feeding the miners  and the metal was valued higher than its production cost, market forces them sculpted human action in relation to how corn was invested.

Bitcoin fits this model to a tee,  people originally exchanged it for near to the production cost, then for pricing determined by the market, all the while the transition is analogues to that of a food commodity being exchanged for a metal like silver of gold, (just at the speed of fibrotic light).

That said, Miesis understood the issue of money inflation more than anyone and his regression thrum addresses this issue but doesn't emphasise it enough, over 98 % of money to day is just money by association, and its only connection to money is it was printed in the same name as the money before it.

Leaving the argument with Bitcoin and regression thrum a little moot, the real problem is Fiat, not lining Bitcoin to the regression chain to prove its validity.

[Longmarch]

This never seems to come up in these regression theorem discussions:

https://en.bitcoin.it/wiki/Proof_of_work

The most widely known and used proof-of-work is the hashcash cost-function which is used by Bitcoin, and also some anti-spam systems and as an anti-DoS mechanism in a number of other protocols. In the context of anti-spam, a proof of work on the recipients address can be attached to the email in an email header. Legitimate senders will be able to do the work to generate the proof easily (not much work is required for a single email), but mass spam emailers will have difficulty generating the required proofs (which would require huge computational resources).

So, the proof-of-work token is originally valuable because it lets you send email, and because it is also divisible fungible and countable it inevitably gets used as money. 

[xxjs]

This never seems to come up in these regression theorem discussions:

https://en.bitcoin.it/wiki/Proof_of_work

The most widely known and used proof-of-work is the hashcash cost-function which is used by Bitcoin, and also some anti-spam systems and as an anti-DoS mechanism in a number of other protocols. In the context of anti-spam, a proof of work on the recipients address can be attached to the email in an email header. Legitimate senders will be able to do the work to generate the proof easily (not much work is required for a single email), but mass spam emailers will have difficulty generating the required proofs (which would require huge computational resources).

So, the proof-of-work token is originally valuable because it lets you send email, and because it is also divisible fungible and countable it inevitably gets used as money. 

That was a long shot. Do you really need there to be intrinsic value in bitcoins?

[Longmarch]

That was a long shot. Do you really need there to be intrinsic value in bitcoins?

If bitcoin is to fit the regression theorem, then yes bitcoin would need to have an intrinsic, or more accurately, a barterable value.

This discussion isn't really about bitcoin, though, is it?  Bitcoin clearly has a value and clearly has utility as a money.  This discussion, I think, is really about the validity of the regression theorem.  Because if bitcoin doesn't fit the theorem, then the theorem is disproven.

As for whether my point is a long shot, it isn't.  It's exactly dead on.  That's why I'm surprised that the point never comes up.  It's impressive that Satoshi saw the barterability of proof-of-work tokens and jumped right from there to turning them into money.  To understand both cryptography and economics enough to see this possibility and actually implement it is a stroke of genius.

[xxjs]

That was a long shot. Do you really need there to be intrinsic value in bitcoins?

If bitcoin is to fit the regression theorem, then yes bitcoin would need to have an intrinsic, or more accurately, a barterable value.

This discussion isn't really about bitcoin, though, is it?  Bitcoin clearly has a value and clearly has utility as a money.  This discussion, I think, is really about the validity of the regression theorem.  Because if bitcoin doesn't fit the theorem, then the theorem is disproven.

As for whether my point is a long shot, it isn't.  It's exactly dead on.  That's why I'm surprised that the point never comes up.  It's impressive that Satoshi saw the barterability of proof-of-work tokens and jumped right from there to turning them into money.  To understand both cryptography and economics enough to see this possibility and actually implement it is a stroke of genius.

I don't see the need to satisfy the regression theorem. With fiat being around as an example for fifty years, people have no problem seeing that a currency without intrinsic value could work. The problem is only to disperse some bitcoins and trigger one exchange. Maybe it was the historic pizza exchange. Mises developed his therories over time, it did not come to him from the gods. Had he lived now, we might have another version.

If you really need it to have intrinsic value, you can anyway not use the cost of production. It has to be a utility value. Bragging power comes to mind. Satisfaction of inventing the thing. Satisfaction of having optimized the mining program. Satisfaction of dreaming about bitcoins as a world currency while sitting in a dark room at night, looking at the hash per second tachometer. Showing a wallet display to your little sister, feeling smarter. Something like that. If anything at all, the utility value of each bitcoin must have been miniscule in the very beginning.

[n8rwJeTt8TrrLKPa55eU]

I don't see the need to satisfy the regression theorem. With fiat being around as an example for fifty years, people have no problem seeing that a currency without intrinsic value could work. The problem is only to disperse some bitcoins and trigger one exchange. Maybe it was the historic pizza exchange. Mises developed his therories over time, it did not come to him from the gods. Had he lived now, we might have another version.

If you really need it to have intrinsic value, you can anyway not use the cost of production. It has to be a utility value. Bragging power comes to mind. Satisfaction of inventing the thing. Satisfaction of having optimized the mining program. Satisfaction of dreaming about bitcoins as a world currency while sitting in a dark room at night, looking at the hash per second tachometer. Showing a wallet display to your little sister, feeling smarter. Something like that. If anything at all, the utility value of each bitcoin must have been miniscule in the very beginning.

I like your two lines of thought.

To put it in experimental terms, the widespread existence and acceptance of fiat money today is a different initial condition versus the starting point for metallic money.  And his theorem only applies under the single initial condition where humans have not yet experienced unbacked units of account.

Non-money primitive humans and post-money fiat humans could be thought of as two conceptually different species in terms of their capacity for abstract thought, with previous humans only being able to attach value to physical objects versus current humans who have been conditioned over decades (for better or worse) to accept intangible fiat tokens as valid representations of value.

So let's all thank Nixon for closing the gold window and making Bitcoin possible

[swappermall]

"So let's all thank Nixon for closing the gold window and making Bitcoin possible"

Thank you, Nixon.  I never thought I would ever say that.

[xxjs]

I don't see the need to satisfy the regression theorem. With fiat being around as an example for fifty years, people have no problem seeing that a currency without intrinsic value could work. The problem is only to disperse some bitcoins and trigger one exchange. Maybe it was the historic pizza exchange. Mises developed his therories over time, it did not come to him from the gods. Had he lived now, we might have another version.

If you really need it to have intrinsic value, you can anyway not use the cost of production. It has to be a utility value. Bragging power comes to mind. Satisfaction of inventing the thing. Satisfaction of having optimized the mining program. Satisfaction of dreaming about bitcoins as a world currency while sitting in a dark room at night, looking at the hash per second tachometer. Showing a wallet display to your little sister, feeling smarter. Something like that. If anything at all, the utility value of each bitcoin must have been miniscule in the very beginning.

I like your two lines of thought.

To put it in experimental terms, the widespread existence and acceptance of fiat money today is a different initial condition versus the starting point for metallic money.  And his theorem only applies under the single initial condition where humans have not yet experienced unbacked units of account.

Non-money primitive humans and post-money fiat humans could be thought of as two conceptually different species in terms of their capacity for abstract thought, with previous humans only being able to attach value to physical objects versus current humans who have been conditioned over decades (for better or worse) to accept intangible fiat tokens as valid representations of value.

So let's all thank Nixon for closing the gold window and making Bitcoin possible

Well put.

[n8rwJeTt8TrrLKPa55eU]

[painlord2k]

People here didn't understand the Regression Theorem abstractly. The point is to generalize it more, not less.
Try to use praxeology, not other tools, to understand economics.
The praxeological analysis never concern itself with the reasons of actions, only with the effects of actions.
BTW praxeology is the tool used by Austrian economists to analyze the market behavior.

There is a simple explanation of why bitcoins have value and Mises's Regression Theorem is right.

http://wiki.mises.org/wiki/Regression_theorem

"People today expect money to have a certain purchasing power tomorrow, because of their memory of its purchasing power yesterday. We then push the problem back one step. People yesterday anticipated today's purchasing power, because they remembered that money could be exchanged for other goods and services two days ago. And so on."

The "two pizza for 10K bitcoins" is the first recorded exchange of something with a hard value for bitcoin.

The reason the guy gave two pizzas for 10k bitcoins is unimportant. What matter is the act of buying "two pizzas for 10k bitcoins". 

The day after, people (the pizza guys and the bitcoin guy and everyone that knew of the transaction) have a price point to remember.
It doesn't matter if it a good point, a bad point, a neutral point or a random point.
The day after people would remember a price point and they would decide if they want pay less or more to obtain bitcoins.

The same was true for any other good or service exchange for bitcoin. The reason people do the exchange is immaterial. But for every exchange there is a price point to be remembered.
When they started to be enough, these price points started to converge on a smaller band (like the theory would expect) because people try to maximize their expected profits.

[xxjs]

People here didn't understand the Regression Theorem abstractly. The point is to generalize it more, not less.
Try to use praxeology, not other tools, to understand economics.
The praxeological analysis never concern itself with the reasons of actions, only with the effects of actions.
BTW praxeology is the tool used by Austrian economists to analyze the market behavior.

There is a simple explanation of why bitcoins have value and Mises's Regression Theorem is right.

http://wiki.mises.org/wiki/Regression_theorem

"People today expect money to have a certain purchasing power tomorrow, because of their memory of its purchasing power yesterday. We then push the problem back one step. People yesterday anticipated today's purchasing power, because they remembered that money could be exchanged for other goods and services two days ago. And so on."

The "two pizza for 10K bitcoins" is the first recorded exchange of something with a hard value for bitcoin.

The reason the guy gave two pizzas for 10k bitcoins is unimportant. What matter is the act of buying "two pizzas for 10k bitcoins". 

The day after, people (the pizza guys and the bitcoin guy and everyone that knew of the transaction) have a price point to remember.
It doesn't matter if it a good point, a bad point, a neutral point or a random point.
The day after people would remember a price point and they would decide if they want pay less or more to obtain bitcoins.

The same was true for any other good or service exchange for bitcoin. The reason people do the exchange is immaterial. But for every exchange there is a price point to be remembered.
When they started to be enough, these price points started to converge on a smaller band (like the theory would expect) because people try to maximize their expected profits.

But the regression theorem requires that the money have value each prior day, even when you get all the way back to the day before the moneystuff became money. Hence the need for an intrinsic value. According to the regression theorem, the pizza maker must have some idea of bitcoin worth, and with no prior exchanges it must have been the intrinsic value. Thats why it seems the regression theorem does not work for bitcoins. You either have to use fantasy and invent some intrinsic value, or you have to modernize the regression theorem.

[justusranvier]

Thats why it seems the regression theorem does not work for bitcoins. You either have to use fantasy and invent some intrinsic value, or you have to modernize the regression theorem.

Or you don't define Bitcoin as money, so the regression theory doesn't apply.

Bitcoin : money :: Internet : newspaper

[Walter Rothbard]

But the regression theorem requires that the money have value each prior day, even when you get all the way back to the day before the moneystuff became money. Hence the need for an intrinsic value.

I thought the regression theorem came from economists who do not believe in intrinsic value and believe that all value is subjective.  I think you may be mixing and matching incompatible economic concepts.

[xxjs]

Thats why it seems the regression theorem does not work for bitcoins. You either have to use fantasy and invent some intrinsic value, or you have to modernize the regression theorem.

Or you don't define Bitcoin as money, so the regression theory doesn't apply.

Bitcoin : money :: Internet : newspaper

You definitely have to define bitcoin as money. It has all the required features, and it is used in exchanges.

[xxjs]

But the regression theorem requires that the money have value each prior day, even when you get all the way back to the day before the moneystuff became money. Hence the need for an intrinsic value.

I thought the regression theorem came from economists who do not believe in intrinsic value and believe that all value is subjective.  I think you may be mixing and matching incompatible economic concepts.

Intrinsic value is value outside of the moneyness value. Gold has it. Subjective? It doesn't matter, as it is up to the actors to define it and express it as actions on the market.

[painlord2k]

But the regression theorem requires that the money have value each prior day, even when you get all the way back to the day before the moneystuff became money. Hence the need for an intrinsic value. According to the regression theorem, the pizza maker must have some idea of bitcoin worth, and with no prior exchanges it must have been the intrinsic value. Thats why it seems the regression theorem does not work for bitcoins. You either have to use fantasy and invent some intrinsic value, or you have to modernize the regression theorem.

"People today expect money to have a certain purchasing power tomorrow, because of their memory of its purchasing power yesterday."
There is no intrinsic value about money. There is no intrinsic value about any and all goods and services in Austrian Economics.
It is "expected purchasing power" depending on the "past purchasing power".

The reason the exchange happened is immaterial after it happened. It doesn't matter What matter is the exchange happened.
The pizza maker could have accepted the transaction for a lot of reasons or just wrote "Yes" in the wrong chat.

The effect is the day after the seller of bitcoin will remember he sold 10K bitcoins for 2 pizzas (10 US$ retail price, 5 US$ cost) (so he valued the 10K bitcoins he sold less than two pizzas) and the pizza seller would remember he bough 10k bitcoins for two pizzas, so he valued them more than the two pizzas he sold. And others would remember the same.
In the mean time, two pizzas would be consumed and 10K bitcoins would be added on the balance sheet of the pizza maker and subtracted from the balance sheet of the coin seller.

Now people know pizzas was bough for bitcoins in the past and they know a price point.
If enough of these exchanges happened in a  short time then market take care of the rest.

Because bitcoin behave as a currency better than other types of currency (because it was designed to do so) as it start to be used (for whatever reason even just random accident) it become a preferred currency to hold than the other currencies. So people tend to exchange their paper currencies for bitcoin and spend their paper currencies more than bitcoin. This raise the price of bitcoin in the other paper currencies and reinforce the feedback.

[justusranvier]

You definitely have to define bitcoin as money. It has all the required features, and it is used in exchanges.

That's not a good reason to classify bitcoin as money. Bitcoin has additional features that money lacks. We don't have a word for it yet but what ever Bitcoin is, money is a subset of that.

[xxjs]

But the regression theorem requires that the money have value each prior day, even when you get all the way back to the day before the moneystuff became money. Hence the need for an intrinsic value. According to the regression theorem, the pizza maker must have some idea of bitcoin worth, and with no prior exchanges it must have been the intrinsic value. Thats why it seems the regression theorem does not work for bitcoins. You either have to use fantasy and invent some intrinsic value, or you have to modernize the regression theorem.

"People today expect money to have a certain purchasing power tomorrow, because of their memory of its purchasing power yesterday."
There is no intrinsic value about money. There is no intrinsic value about any and all goods and services in Austrian Economics.
It is "expected purchasing power" depending on the "past purchasing power".

The reason the exchange happened is immaterial after it happened. It doesn't matter What matter is the exchange happened.
The pizza maker could have accepted the transaction for a lot of reasons or just wrote "Yes" in the wrong chat.

The effect is the day after the seller of bitcoin will remember he sold 10K bitcoins for 2 pizzas (10 US$ retail price, 5 US$ cost) (so he valued the 10K bitcoins he sold less than two pizzas) and the pizza seller would remember he bough 10k bitcoins for two pizzas, so he valued them more than the two pizzas he sold. And others would remember the same.
In the mean time, two pizzas would be consumed and 10K bitcoins would be added on the balance sheet of the pizza maker and subtracted from the balance sheet of the coin seller.

Now people know pizzas was bough for bitcoins in the past and they know a price point.
If enough of these exchanges happened in a  short time then market take care of the rest.

Because bitcoin behave as a currency better than other types of currency (because it was designed to do so) as it start to be used (for whatever reason even just random accident) it become a preferred currency to hold than the other currencies. So people tend to exchange their paper currencies for bitcoin and spend their paper currencies more than bitcoin. This raise the price of bitcoin in the other paper currencies and reinforce the feedback.

You just refuted the regression theorem here. I don't disagree.

[xxjs]

You definitely have to define bitcoin as money. It has all the required features, and it is used in exchanges.

That's not a good reason to classify bitcoin as money. Bitcoin has additional features that money lacks. We don't have a word for it yet but what ever Bitcoin is, money is a subset of that.

We have words.

Compared to paper fiat:

Possible to transfer over the net.
Can be backed up.
You need a computer

What kind of properties does bitcoin have that makes it not money?