Wednesday 24 April 2013

Bitcoin at the 2nd International Conference on Complementary Currency Systems

I've been invited to the 2nd International Conference on Complementary Currency Systems (CCS), taking place between 19th and 23rd June 2013, in The Hague, Netherlands. I'm on a panel on the day 3, the strand of the conference specialised for local government officials and policy-makers. I'm looking forward to meeting the other participants, in particular Jorge Timón from the Freicoin project with whom I exchanged comments on the forum.

Wednesday 17 April 2013

The Mises Institute is clueless about Bitcoin

The Ludwig von Mises Institute (the one behind, located in Auburn, Alabama) posted several articles over the last week or so about Bitcoin:
Prior to that, they posted about it during June 2011, by Justin Ptak:
And one in October 2011, by Jeffrey Tucker:
In the meantime, Jeffrey Tucker became a Bitcoin enthusiast (disclaimer: I've been interviewed by Jeffrey and Laissez-Faire Books is publishing my book about Bitcoin, so I might be biased). Justin Ptak appears now to be friends (on Facebook) with Bitcoin fans. This may or may not imply his change of opinion, but it doesn't look like he published anything more about Bitcoin at least.

Now, there is nothing wrong with criticism. And Bitcoin can be criticised, there are many legitimate objections to it. But to criticise something merely because someone feels a pressure to criticise, and then rushes to hastily print something quickly is not scholarly work. It is symptomatic that the institute didn't publish anything in between. They are under pressure to publish something when there is a media interest in Bitcoin, and hence hastily rush to assemble something. But actual academic research appears to be absent.

The main issue appears to be the conflation of money, unit of account and a medium of exchange. Unit of account is not a necessary function of either money or a medium of exchange. It is merely a possible byproduct of it.

A further issue is the "self-reinforcing monetary spiral" (i.e. which mainstream economists call the network effect), in which one medium of exchange emerges victorious and beats all other media of exchange and that we call money. But this is merely a hypothetical model that is furthermore prone to misinterpretation. First of all, transaction costs can prevent this spiral to escalate to its final stage. Currently, we have hundred something currencies all over the world. Legal restrictions prevent this spiral from playing out. But this is merely an empirical factor, rather than a rule that gives legal restrictions magical powers. Even without legal restrictions, we can't be entirely sure that the transaction costs won't hinder a full monetisation.

The second issue is the neglect of other media of exchange, those that are not money. Mises calls them "secondary media of exchange", and Rothbard calls them "quasi-monies". These goods are liquid, and a part of their demand is due to their liquidity. They are not liquid enough to be money, but nevertheless they serve, not only through their other uses, but also through their liquidity, a valuable purpose. This is what Bitcoin is. This is what gold is too. And this is also the pool for potential candidates for money. Before something can be money, it must be a medium of exchange.

Bitcoin is somewhat liquid, and it has a very important advantage against extant money: it reduces transaction costs. It reduces transaction costs even further than existing payment mechanisms, so that even a fluctuating price is not enough to offset this reduction. We can therefore expect Bitcoin to be used as a payment mechanism in those areas where it can substitute for other payment mechanisms. This is also a type of network effect. Once they use it as a payment mechanism, people may decide that they do not actually need to fully convert it to/from fiat, and use Bitcoin as a store of value as well. Indeed, Tony Gallippi from BitPay reported (can't find the link now) that their customers are increasingly opting to keep a larger proportion of the payment in Bitcoin, whereas at the beginning they just converted the whole sum to fiat money.

Yet again I have to quote White in his brilliant insight, which has not yet been processed by other Austrians:
"Coinage reduces transaction costs compared to simple exchange, because of authentication and weighing. Bank liabilities also reduce transaction costs. But these are empirical factors, and not something inherent in all possible monetary systems."
Rather, other Austrians make empirical (!!!) statements like this (Salerno):
"With the use of clearing systems, money substitutes are virtually costless to transfer."
An adoption as a payment mechanism, and expansion into a store of value are the early stages of monetisation. This is the same mechanism as the Austrians hypothesised occurred during pre-monetary times, only we now already have a different money. But already existing money is not a showstopper for this mechanism to work. Liquidity is just not the only factor influencing the choice of media of exchange. The argument of Hoppe that
"Driven by no more than narrow self-interest, man will always prefer a more general, and if possible, a universal medium of exchange to a less general or non-universal one."
is therefore false. It is only apodictically true in a world without transaction costs. It still may happen in a world with transaction costs, we merely can't be sure about it. And Bitcoin is a hint that empirical factors can't be dismissed entirely. EDIT: If the statement of Hoppe was true, once money exists, it could never have been replaced by a new money, and we clearly know from history that that's not the case.

Will Bitcoin ever become money? That's an empirical issue and we can't know this in advance. But equally we cannot dismiss it, unless we find a competitor to Bitcoin based on fiat money (or precious metals like gold) that is able to mitigate the transaction cost advantage of Bitcoin. And that would be very difficult to pull off. One of the reasons is the transaction costs associated with the boundary between money in the narrower sense and money substitutes (such as redemption, settlement, and so on), which Bitcoin does not need. Bitcoin is form-invariant and can exist in practically any imaginable (and unimaginable) form. The second one is regulation (management of money and money substitutes is strictly regulated and burdened by many restrictions which have nothing to do with monetary policy, such as anti-money-laundering, capital controls, war on drugs and so on). Even if regulation affects Bitcoin, unless there is an alternative that isn't affected by regulation, there is still no reason for Bitcoin users to switch to something else.

But Bitcoin can do much more than become money. With algorithmic contracts, it can make large parts of the financial sector obsolete. With its ultra low transaction costs, it can make money substitutes redundant (and, obviously, without money substitutes there is no credit expansion and without credit expansion there is no business cycle). Again with its transaction costs and abstract base, anyone can make a payment to anyone, anytime, anyplace. Nothing that has existed so far in the history can do that. And even if we disregard it as a hypothetical, we just need to remember that gold already failed, because it was reduced from money to a secondary medium of exchange. This was only possible because money substitutes emerged. If nothing else, Bitcoin shows that money substitutes are merely an empirical quirk and not an inherent feature of monetary systems, and for that alone is it should change the landscape of the Austrian literature forever, and open a wide spectrum of possibilities for research and our understanding of money.

For more in depth analysis, I recommend my master's thesis, or if you wait a while you can get an updated version in a book format.

Wednesday 10 April 2013

The economic fundamental of money substitutes

Since Patrik Korda still criticises my approach to money substitutes, I thought I'd summarise here their fundamentals. This is not a definition of a money substitute, nor an exhaustive listing of all possible types of substitutes. Rather, it is to explain their essence in a way that is easily understandable and clarify why Bitcoin has nothing to do with them.

The fundamental of a money substitute is a causal link between money in the narrower sense and the substitute. This causal link is the main, necessary, factor explaining the price of the substitute. For argument's sake, I'll leave it open whether other, optional, factors exist. This is the essence of the point Mises was making, he was trying to explain the price of money substitutes. And he found the explanation in the causal link to the money in the narrower sense.

For Bitcoin, there is no such underlying causal link to a money in the narrower sense that explains the price of Bitcoin. The price of Bitcoin is determined by other factors.

Therefore Bitcoin is not a money substitute (or it wouldn't be if Bitcoin was money).

Tuesday 9 April 2013

Another interview with me, and my reaction to the recent media hysteria

The Slovak bitcoin Blog "" published a brief interview with me (I'll translate it when I have a bit more time):

However, there is an important thing that I want to get off my chest, because I don't think I've concisely addressed it yet. As far as I can remember, I have been refusing to give predictions about the future price of Bitcoin, I only indirectly explained how and why it will be influenced. In my master's thesis, I explained the mechanisms  influencing the demand for Bitcoin and what it means for the future of Bitcoin. Here I'll summarise it here very briefly.

For the foreseeable future, I expect the price of Bitcoin to be subject to fluctuations, just like it has been behaving until now, because the market isn't sufficiently liquid. It may currently be in a bubble phase, or not. I don't know. But it is also irrelevant for the future of Bitcoin, because a medium of exchange on the market gains market share through its reduction of transaction costs against competitors, not through its price stability. Indeed, in A case for a genuine gold dollar, Rothbard criticised Hayek's Denationalisation of Money for precisely the same point: that price stability is not a valid reason for people to choose one medium of exchange in preference for another. Rothbard invoked, from the point of view of the network effect, the argument of critical mass: the competitor wouldn't be able to compete with something people are accustomed to. But other factors are unmentioned by Rothbard, in particular the reduction of transaction costs. This is because until Bitcoin, Austrians (with the exception of Lawrence White, whom I quoted several times) assume that money substitutes have lower transaction costs than money in the narrower sense. Bitcoin exposes this as an implicit assumption which is actually an empirical factor, and leaves most of the Austrians dead in the water, not understanding what's happening.

Will Bitcoin ever become money? I don't know, and while that would be terrific, it's also not relevant for the broad question of the future of Bitcoin, because of the "money or nothing fallacy" (which is a false dichotomy that many critics, Austrian or not, of Bitcoin, invoke, arguing that if we can refute that Bitcoin is or will be money, it follows that it is nothing). But that's bogus. As long as it provides an significant advantage in transaction costs, it will be here and compete, even if it never becomes money (whether that's due to too much leverage the states have over fiat money, or due to inertia, or even if we admit all the arguments of Gertchevs and Kordas about impracticality of electronic money, even though 31% of Kenya's GDP is now being spent through mobile phones, i.e. if we accept whatever assumptions they make up on the fly). Non-Austrians have called such a medium of exchange "metacurrency", for example, Krugman (who is critical of Bitcoin, but missing that it confirm his own 30 year old papers) calls it "vehicle currency". Austrians also have a name suitable for such class of media of exchange, for example "secondary media of exchange" (Mises) or "quasi monies" (Rothbard). But whatever our classification of Bitcoin may be, from economic point of view it is an immaterial good with ultra low transaction costs and an inelastic supply, a perfect fit for a reform of payment mechanisms and the financial sector, irrespective of whether it also sucks up all the "moneyness" from fiat, whether it "monetises" or not, and irrespective of what people argue about it. In order to pull this off, Bitcoin does not have to be perfect, it just needs to provide a significant comparative advantage over competitors. This it has, and I expect it to stay that way for the foreseeable future.

Bitcoin may fail for one reason or another, but it won't be because its price is susceptible to bubbles, because we argue how to classify it, because of low barriers to entry, because economists or pundits don't understand it and have to hastily make things up to patch the holes in their arguments. It will be because it will stop having a comparative advantage in transaction costs, i.e. it will be replaced by something which provides better utility for the users.

Response to Patrik Korda #4

This is a followup to the previous replies to Korda:

Korda made it to the daily, and he presented a couple of comments at the bottom of my post

This is an attempt to address the outstanding points, in a more direct, quote by quote, manner, so it's not really a full blown article with a beginning and a conclusion. I just want to have the points sorted out.
"I believe that bitcoins are a lose-lose proposition for libertarian minded people going forward."
So what will people use to reduce their transaction costs of trade instead? Will people switch from the internet to the post office and library?
"Jeff Garzik stated that (1) bitcoins were either a bubble that will collapse or (2) bitcoins were coming up to the level at which they would trade as a worldwide digital currency, this was his response to the dramatic price increase at the time. In retrospect, bitcoin did in fact turn out to be a bubble."
Korda seems to ascribe the argument of Garzik too much power. He has yet to explain why the bubble is relevant. The relevance of the bubble is only for people who use Bitcoin for speculative purposes. People who use it for a reduction of transaction costs are unaffected, to the extent that the remaining prices still allow trade. Similarly, the tulip bubble was relevant for people who speculated on the tulips. But last time I checked, tulips still exist and are regularly bought and sold. In addition to that, there has been a lot of investment into Bitcoin projects, and these have nothing to do with the changes in price. I don't think there have been comparable investment into tulip infrastructure, and tulips do not decrease transaction costs, so there is no need to hold them for transactional purposes. In fact, we already did have a spike and a following crash in the prices of Bitcoin, in 2011. Yet, the market share grew irrespective of this development, and so did investment, the variety of goods and services related to Bitcoin. So the relevance of the argument is directly contradicted by empirical evidence.
"How could this be bad for libertarians? The answer to this question was also addressed by Jeff Garzik in the interview I posted in my article. Most consumers are not libertarians."
But many, if not all, consumers have liquid assets, such as money, and can recognise the benefit in decreasing transaction costs. Similarly as you don't need to be a geek to want to use the internet.
"With mainstream acceptance comes mainstream policy, which includes registration of the exchanges as money service businesses, regulation, fraud insurance, and higher fees to pay for all of this."
The exchanges already have long been subject to MSB regulations. But Bitcoin does not need exchanges to work. Even during communism with high capital controls, illegal forex brokers were quite present. Currently, we see the situation in countries Argentina and Iran.
"I simply do not attribute as much weight to the network effect nor as much rationality to the market process as you do."
Wait a minute. Now this is a turn in 180 degrees. If there is no network effect, then there is no demand for a new competitor in the first place.
"Nor does it always follow that a product which overtakes another has to be superior in some fashion or another. Standard economic theory would lead one to believe a competitor can win out by either offering a superior product or a lower price. However, there are times when offering what is virtually the same product but at a higher price is the right strategy, as Charlie Munger would point out."
In network effect, the "lower price" manifests itself as transaction costs, because the object is a tool, not a consumption good. It is used indirectly. What is the price of a language? What is the price of the internet? There is none. There are costs associated with using them, and benefits that are accessible through their use. In network effect, the transaction costs in the narrower sense can be compensated against by the network size, and if the transaction cost difference is small, the relative market share size can be permanent.
"The classic example of this is red bull, which used to come in large cans and sell for the same price it does today. Another example would be facebook, which is generally more clean-cut and does not allow as much personalization as does myspace."
Korda now himself admitted that there can be differences, and these might cause a switch, whereas before he argued that the switch can happen without an apparent reason. Of course, something like this could happen with Bitcoin. A new competitor can provide something different, which the users of Bitcoin consider relevant, and which Bitcoin can't mimick. But the chances of this happening are not very high. Bitcoin is open source, available for anyone without restriction, and its feature set can be enhanced long before the competitor starts gaining market share. If we do a bit of research on Xanga, we find out for example that they were found to be violating  "Children's Online Privacy Protection Act". In addition to paying a penalty, they were then subjected to monitoring by the FTC. If we do a bit of research on myspace, we find out for example that their system didn't scale, wasn't able to grow outside of US, made problematic decisions with respect to marketing and so on. I don't know which of these factors is relevant, but clearly there were a lot of differences and many of them could have been relevant.
"People typically do not have an incentive to make up their own language out of nowhere. On the other hand, there has since time immemorial been a great incentive for alchemy."
I already commented on this one in the comment section. This argument is a variant of the labour theory of value, i.e. it misses the demand part. Merely because people invest a lot of money into creating something, it does not follow that this creates demand for it. Indeed, the more people start their own competing projects in the area of network effect, the less likely it will be that any of them will reach significant market penetration, and the costs would be more distributed as well. Now, if someone started a single, big project, with the idea to overcome Bitcoin, and invested billions of dollars into its development, integration, promotion, and so on, that might work. But that would then raise the bar even higher for a subsequent challenger, requiring him to spend even more money. While there is no barrier to entry, there is a barrier to be able to reasonably compete. The network size is also a relevant factor, that gives preference if no relevant distinguishing (technical) factor is present.
"This is the primary reason why people, such as myself, have piled into litecoins earlier this month. I banked on my parable."
However, this exactly proves my point. The relevant factor for market share of Bitcoin is liquidity, not price. Investment into LTC infrastructure, good and services is still much smaller than that of Bitcoin, even if Mt. Gox opens a LTC market. Korda probably spent too much time on stock exchanges and too little in the payments industry.
"Moreover there are always, and I mean always, unforeseen complications. The most recent one as far as bitcoin is concerned are the so-called forks."
This is probably not apparent for the majority of observers, but the ability to fork are actually good, because they provide a method for resolving problems that arise.
"I believe that digital token money may very well have a big future, but that doesn’t mean bitcoins cannot go the way of xanga. Contrary to the network effect, the fact that bitcoins were first may very well be a detriment. If something major goes wrong, rationally or irrationally, people will switch over to a competitor, even if that competitor may be inferior or nearly identical with little to no improvements."
This is now again an entirely opposite argument, indeed this is the same thing that I said. There are valid reasons for people to switch, even if we do not understand them. But that presented by Korda (low barriers to entry) is not one of them.
"Actually, initially you argued that it [price of a money substitute, ed] must be fixed. It was only subsequently that you budged and stated that it can in fact fluctuate, but only downwards according to you."
The reason why I did this was that I am unhappy about the Austrian definition of a money substitute. If I insisted on it, then Korda's argument about classification of Bitcoin would be refuted outright and my job would be done. Korda uses Mises' Theory of Money and Credit to classify Bitcoin, but simultaneously rejects his whole reasoning and definitions when doing that.
"I don’t think this is correct. You can have credit money on a metallic standard. But as regards my article and our exchange, this point is irrelevant. All that needs to be said is that bitcoins are not a credit money."
Korda entirely missed my point here. My point was that a money substitute can evolve into something that is not a money substitute, and the primary cause for this switch is the breaking of the coupling of the prices between the money substitute and the underlying good. I was trying to explain the essence of Mises' classification system.
"This is more akin to the labor theory of value in my opinion. It takes time, energy, and in essence money to produce a desired outcome in a videogame. However, it does not follow that therefore the various gamers out there should be getting paid for accomplishing those objectives. The essence of the regression theorem is that the object at hand has use-value prior to attaining exchange-value."
I can see how one can mistake my argument for labour theory of value. However, what I still don't understand why Korda is missing the point of my argument. What I was doing here was analysing empirical data, not presenting a theoretical argument. I was referring to this:
"During 2009 my exchange rate was calculated by dividing $1.00 by the average amount of electricity required to run a computer with high CPU for a year, 1331.5 kWh, multiplied by the the average residential cost of electricity in the United States for the previous year, $0.1136, divided by 12 months divided by the number of bitcoins generated by my computer over the past 30 days."
and followed by a table of exchange rates between USD and BTC, rangining from 1/737 to 1/1622. In other words, the producer was using his costs to establish his ask position. If Bitcoin was a money substitute, he would not mention the costs at all, and the table would contain percentages (discounts/premiums or agios/disagios), possibly would not exist at all (because BTC would trade at par with USD). This is as clear as it gets that Bitcoin is not a money substitute. I don't know how else I can make it more apparent. The production costs of money substitute do not play a role in deciding the ask position of the producer, rather the market price of the underlying good. Which, in case of Bitcoin, doesn't exist.

Now to the second part, the regression theorem. I don't know to what extent Korda was paying attention when reading my thesis, but I clearly explained how the regression theorem fits into Bitcoin. But even if I was wrong, empirical evidence clearly shows that Bitcoin is a medium of exchange. So whatever objections Korda has regarding Bitcoin with respect to the theorem, they are irrelevant. It's like attempting to refute the existence of GPS by pointing out that the earth is flat. It's a methodological blunder, attempting to mix theoretical reasoning with empirical data.
"I have known of families and individuals who hoard things which are extremely durable, such as gold. However, I have not heard of anyone hoarding binary digits."
Surely though, Korda must have heard about the internet. While people are not hoarding the internet, they are using it to reduce their transaction costs.
"The primary reason that I classified bitcoins as token money was that there was no other place to put them."
But shouldn't then the proper approach be to attempt to understand the reasoning for the classification system and amend it based on that, instead of making up new stuff? That's what I tried to do. I established that the reason for the classification system is the different mechanisms by which the price of those goods emerges. What is Korda's underlying reasoning for his system? Is it based on catallactics or not?
"If we presume that bitcoins will eventually become money, which is presumably what most bitcoiners would like, then we are back to square one in having to classify them. In your thesis, you seem to claim that bitcoins are a commodity money, which is presumably why you have reformulated the regression theorem."
Now again we have several issues. First of all, there is no necessity for Bitcoin to become money, nor is there a necessity that we derive the evaluation of Bitcoin from it becoming or not becoming money. I dub this fallacy "money or nothing", it's a false dichotomy. It can remain a medium of exchange, and be primarily used as a payment mechanism, in a world dominated by fiat. How is that supposed to be "bad" for the future of users of Bitcoin?

The reason why I classified Bitcoin as commodity money I explained in my thesis (indeed included two references to support my decision), and elaborated on it in three (now fourth) blog posts. The reason was catallactics, the analysis of supply and demand for Bitcoin. And last but not least, the reason why I reformulated the theorem is that I have not seen it formulated in a coherent manner anywhere, and people were ascribing it features which it doesn't have. One of those people is Korda, who ascribes to it quantitative aspects ("handful of people consume them"), whereas it is a qualitative argument (emergence of price and liquidity). Even if Bitcoin didn't exist, my reformulation would be relevant for fiat money, for example, because many Austrians interpret it very weirdly, for example that the power to tax gives creates demand for money (which still does not explain the price and liquidity of fiat money).
"In your thesis, you claim that “the price of bitcoin correlates with the public interest in bitcoin”".
If this is an attempt to fully explain the price of Bitcoin, it is what what Mises calls an "acatallactic monetary doctrine". It is missing the essence. I explicitly say with respect to Chapter 4 that for Austrians, empirical data are not arguments.
"Although this is true, a more precise way of putting it would be that interest in bitcoin follows the price of bitcoin."
There is a research paper which comes to the opposite conclusion, i.e. that the price follows the interest. But that still does not answer the actual question.
"This is exactly why I think the price of bitcoin is relevant towards its potential success and that a bubble will be very detrimental."
 "I’ve been in the markets long enough to see how this story ends."
Korda has been around the wrong markets. Or more accurately, those that are not relevant for Bitcoin. The relevant factor for the future of Bitcoin isn't its price, but liquidity. In other words, the payment systems industry, and to a certain extent, the forex market.
"Your first article practically claims that all of my points are irrelevant."
Yes and that is still correct. I have yet to see why any of the arguments are relevant.
"You went on to say that the primary advantage of bitcoins were the lower transaction costs. While it is certainly true that exchanging binary digits is much cheaper than having to move fiat, or worse yet, specie around the globe, there is another edge to this sword. Let us not forget that the most common arguments for fractional-reserve banking or fiat money are in fact lower transaction costs."
Exactly. That is why on a free market with a metallic monetary standard, I expect fractional reserve banking to out-compete full reserve banking, due to transaction costs. And that is again why I expect Bitcoin to outcompete fiat money and gold as long as it can keep its comparative advantage in transaction costs.
"I think transaction costs can be greatly eliminated via companies such as BullionVault or GoldMoney."
I present theoretical arguments in my thesis why this is probably inadequate to compete with Bitcoin. Furthermore, empirical evidence shows that Bitcoin has already outgrown GoldMoney (2.1 vs 1.9 billion USD). And, you can't use GoldMoney anymore to pay, unless you live in Jersey. We may argue that this is due to regulation, but that is precisely my point, regulation is just yet another factor influencing the suitability of individual goods to reduce transaction costs.
"As for bitcoin, I think they are at present a lose-lose proposition. To reiterate, if it pops this is not good news for the holders of bitcoins. If it does not pop and is instead reasserting itself as a worldwide currency, then it is bound to lose its libertarianism as Garzik pointed out."
But now this is yet another argument. If it collapses, it means that it was replaced by something better suited for the users. If it doesn't, it means that it provides a better service than the competitors. Whether this has to do with libertarianism is irrelevant.

Last but not least, I want to address the name of the article, "Bitcoin: Money of the Future or Old-Fashioned Bubble?". As I have been hopelessly trying to explain, these are not the only options, neither are they mutually exclusive. It's a "money or nothing" fallacy I've been complaining about for a long time. Even if the current price was a bubble, it would have negligible effect on the ability of Bitcoin to decrease transaction costs, and therefore its prospects as a medium of exchange.