Saturday, December 2, 2017

Is Bitcoin A Speculative Bubble?

There are at least two definitions of a speculative bubble.  The first, and most widely accepted, is that it involves a price of an asset that rises substantially above its fundamental and then falls back towards that fundamental.  The other, not necessarily all that clearly distinguishable from the former, involves an asset price that rises due to people buying due to an expectation that they will get a capital gain from its expected future price rise, with this then happening due to a self-fulfilling prophecy, with eventually the price falling sharply, with this not necessarily involving a fall towards a fundamental because the asset may have no fundamental at all.

I note before proceeding futher that there is an enormous debate and literature on identifying fundamentals at all, even when they might theoretically exist.  There is a serious body of opinion dating from Flood and Garber a quarter of a century ago that one cannot identify them econometrically ("Tulipmania").  This has been shown to be false by me and many others in various papers, including particularly on closed-end funds where the net asset value of the fund minus taxes and fees is a fundamental, and when those soar to twice the value of the underlying net asset value, well, we are looking at a bubble. The lit is there and decisive.  I asked Garber to comment on a paper presented a long time ago at a conference on this point, but the chicken shit did not show up to admit that he was just plain wrong.  He had no legitimate excuse for his absence.

Of course Bob Shiller has made pretty reliable estimates regarding housing prices based on price to rent and price to income ratios.  His studies of these by 2005 were pretty decisive to anybody who was remotely paying attention (including at the Fed, Janet Yellen, the only one there to take this seriously at the time), that the US housing market was in a serious bubble that was going to crash big time.  The matter of forecasting how bad it would get with the Great Recession became a matter of who had figured out how deeply the financing for all that had gotten involved in world financial markets at a fundamental level, and very few did figure that out.

But then there are assets that have no fundamental at all, even theoretically, quite aside from all the horrendous econometric identification problems pointed out by such serious people as Jim Hamilton, for whom I have the deepest respect. The question for cryptocurrencies in general is whether they have a fundamental, and it may well be that they do not. If that is the case, then only the second definition of a speculative bubble may be relevant, and that is much harder to determine than the former, already admittedly a difficult matter.

How might bitcoin have a fundamental?  One reason might be for its use as a medium of exchange.  However, while it is certainly being used as that, for regular commodity transactions as of now it remains as a sometimes difficult alternative to cash dollars.  As near as I know there are no regular commodity transactions in the real economy that require it.  So, it may have no fundamental, and from what I have heard, this is widely accepted among the more sophisticated bitcoin traders.  This would make it like art, such as the recent sale for $450 million of the possibly faked "Salvator Mundi" supposedly by Leonardo da Vinci.  And, of course, there is good old gold that has a long had a value as a store of wealth about an order of magnitude above its likely strictly commercial fundamental value.

However, it may be that bitcoin has a fundamental value above zero, if wildly fluctuating and hard to estimate, far beyond the econometric difficulties identified by people such as Hamilton, much less Flood and Garber.  It is that to buy most of the other cryptocurrencies one must use bitcoins to do so.  Oh, this is really a gas.  The fundamental for one asset is based on its ability to purchase even more lacking in fundamentals and totally speculative assets one must own it.  Wow.

Thus we have that any real fundamental for bitcoin is a derived demand for any among the vast universe of other cryptocurrencies, and we should keep in mind that bitcoin itself is horrendously inefficient compared to others because of its accelererating and already very large "mining" costs.  As near as I can tell there are only two other cryptocurrencies that have any relation to the real world out of the multitude of crytpos. They are ethereum and ripple.  The first has been a matter of much speculation itself, given its potential for writing contracts, giving it a serious possibility of much longer use and actual usefulness in the real world.  This may yet occur, although for its future it would probably be better for it if it could be bought directly with cash/dollars from the real world rather than having to use the horribly socially inefficient bitcoin, a bad example of first mover advantage, or perhaps the ultimate proof that the first mover should be sent to last.

Which gets us to the quiet and most obscure member of the crypto world, ripple.  This cryptocurrency, which is the hardest to buy of them all, is probably the one with the most serious actual real world use, and thus possibly providing a real foundation for bitcoin to have a fundamental, although I confess at this point that I am not certain to what degree purchasing it does rely on using bitcoins.  I know that as of fairly recently one had to use bitcoins to buy ethereum, but I am not sure about ripple, which moves with bitcoin, but probably more weakly than any other of all the cryptocurrencies.

The source of its value is that has been adopted by a significant number of banks for their interbank transactions.  This now appears to be firmly established, not to go away whatever happens to bitcoin or any of the other cryptocurrencies.  Indeed, the underlying idea of blockchains is clearly a brilliant and useful forward movement in managing financial and economic transactions, assuming that is managed in a reasonable and efficient manner, without me remotely dealing with issues of transparency or legality.  But it seems that at least some banks have decided to use ripple, which I understand uses a more efficient mining technology.

So, there may well be a fundamental for bitcoin, despite what I understand to be the current consensus among smartass bitcoin traders.  But, of course, that fundie is probably way below the current price, as if that matters at all.

Barkley Rosser

Addendum, 12/2:  This has been picked up by Naked Capitalism where some commentators have pointed out that apparently one no longer needs bitcoin to buy other cryptocurrencies, especially ripple and etherium.  If anything it is etherium that is being used to buy other cryptos.  In any case, that reduces the case for bitcoin having a fundamental greater than zero.


Robert Flood said...

Actually I gave comments on a paper published recently in Econometrica in which the authors had data on very very long lived lease holds that I thought identified that no bubble was present in british housing prices.

Barkley Rosser said...

That may be. I have said nothing about UK housing prices, although I fully agree with Shiller that US ones were a speculative bubble, if subject to some question marks about misspecified fundamentals. Anyway, in the US, the price to rent and price to income ratios after 2000 rose to levels never seen before in US history, pretty suggestive of a bubble, and, of course, they fell long and hard later.

I apologize through you to Peter Garber for snarking too much at his failure to show up to discuss a paper he said he would discuss without ever giving any excuse. My remarks overdid it, although the facts are as I stated them.

That paper, eventually published in JEBO prior to when I edited it (with Ahmed, Koppl, and White) was about closed-end country funds, especially the price movements that occurred in late 1989 and early 1990. Without getting into too many gory details, a lot of them, including ones where people could freely buy the underlying assets (not always the case for some countries, such as South Korea at the time), with Spain and Germany prime examples, had their closed-end fund prices soar to about twice the level of their net asset values, before crashing pretty hard in February, 1990. A lot of the action at the time seems to have been driven by Nomura, with much of the money in this particular market fleeing the about to crash and then crashing Japanese stock market.

We did not come up with the idea that closed-end funds are a case where, assuming ability to buy the underlying assets, one can pretty reasonably accept that the fundamental is probably slightly below the net asset value due to tax and fees issues, as I noted in my post. Others have noted this, including Thaler as well as Brad DeLong, who in one paper noted that due to the misspecified fundamental problem we cannot say for sure that the US 1929 stock market was a speculative bubble, but that we can be pretty sure that the closed-end fund market was, where at that time it behaved like closed-end country fund market in 1898-90, with fund prices roughly doubling the underlying net asset values before finally crashing hard.

Robert Flood said...

The thing that puzzled me about your interaction with PG involves my recollection of Garber's and my discussions about identification. As I remember this (from 1980, which is 37 years btw) is our saying we could not tell the difference between a greater fool bubble that burst (which we got from Brock/Blanchard)) and a payoff expectation that did not materialize. Have a look at the chart in the back of Kindleberger - his bubbles (suspiciously) occur around new markets and new inventions - some of which pay off and some don't. Also Tulip mania was NOT a greater fool type bubble since the price that 'bubbled' was not a spot price but a futures price so time does not move in the pricing equation. The deal was for delivery of tulip bulbs at a fixed future date. Oh and the point of the UK think was to tout the very cool data the authors used - spot prices for freehold and for leasehold of very similar properties - and they had lots and lots of data. said...


Well, my (non) interaction with PG happened in the 90s, the paper being discussed having been based on the closed-end country funds bubble that blew in 1990.

I have not looked at your Tulipmania recently, but my memory is that in the end the bottom line was that the most likely actual bubble was only in the final month for oridnary tulip futures, not the high priced ones that got all the publicity in MacKay and other such sources. There were special enough details and circumstances regarding them that your misspecified fundamentals problem.

My reading of Kindleberger in his various editions involves more cases that are not new commodities or markets, although many of them were.

Robert Flood said...

BTW, I have never written anything about Tulipmania. That's Peter's thing. I gave some lectures on bubbles at ND - before retiring - and read Peter's work. Peter's puzzling about the common tulip varieties in the final month or 6 weeks always seemed odd to me. As you know, the tulip contracts were voided by the king - or someone - so actual settlement was just "cents on the dollar." My point is that people might well have anticipated the probabilistic voiding of the contracts. So when one sold the bulb future there were two possible payments - so expected price was: (prob no king intervention)* nominal price agreed + (prob of king intervention)* some unknown low price. So if I were a drunk dutchman in a bar, I'd put the nominal agreed price up pretty high to compensate (in expectation) to the chance that the king would void. The king did void, of course, and I can see the prob of voiding rising as the voiding approached, which would account for the so-called bubble in common bulb prices.

Barkley Rosser said...

Fair enough, Robert, although you have written on the misspecified fundamental problem with him, I believe, JPE 1980 if I am not mistaken, and I have always taken the critique seriously, hence the interest in such cases as closed-end funds where it looks like one might actually be able to get pretty close to pinning down an actual fundamental. While I poked at him for not showing up at that conference, Garber's "Tulipmania," both the article and book versions, are interesting and informative.

If I have misrepresented your position at all (aside from appearing to include you as a coauthor on Tulipmania), I apologize.

Anonymous said...

better described as a pyramid scheme than a bubble said...

That is not obvious, Anonymous. A blockchain is not a pyramid scheme, per se. Transactions do not depend on each other in the way those in a pyramid do.

Robert Flood said...

Pyramid schemes, Ponzi schemes ect are what economists call "greater fool" schemes/bubbles. The thing they have in common is no final condition or price. Technically this whole thing came out of the "necessity of transversality" discussion, which is technical talk for "yea but if there is a bubbles people can't be maximizing - so if people are really maximizing there can't be bubbles." BTW, that is why Garber and I looked at the German hyperinflation since price level is the only case where the econ maximizing assumption does not imply no bubbles. said...


Transversality is only relevant for an infinitely lived agent. Not relevant in either an OLG model or the real world.

Hyperinflations are not bubbles, they are collapses of value, not escalations based on greater fools or self-fulfilling sunspots or whatever.