Friday, January 25, 2013

Bitcoin is an amoeba, central banks are blowfish


When my mother asked me yesterday if I was still buying the bit points, I took it as a sign that it was time for another bitcoin post.

One of the most popular reasons for owning bit points—sorry, bitcoin—is that the supply of coin is fixed whereas the supply of central bank money can be increased ad infinitum. Like an amoeba colony nearing population saturation, the bitcoin supply is growing at a decreasing rate as it approaches the magic 21 million number, the ceiling specified by designer Satoshi Nakamoto. Bitcoin advocates believe that this controlled supply effectively grounds the price of bitcoin while leaving the value of central bank money to flap in the wind.

But this ignores the mirror image of this argument. Yes, a central bank can rapidly increase the supply of notes and reserves. But blowfish-like, a central bank can just as quickly suck this supply back in—indeed, a central bank can go to the extreme of extinguishing every last liability it has ever issued. Bitcoin, on the other hand, can never be destroyed by its issuer—it has no issuer. The implications of this for the values of bitcoin and central bank money are important. I'm going to use someone else's model to show why.

Mencius Moldbug has a recent post called How Bitcoin Dies. If I'm not mistaken, he's making reference to Adam Ferguson's book When Money Dies (pdf), an account of the Weimar inflation. In any case, I agree with much of what Moldbug (his real name?) writes. Bitcoin's value is highly tenuous, and it wouldn't take much of a shock to send it to 0. I'm going to borrow Moldbug's model of a bitcoin economy and use it to explain a central-bank economy. This will help us to see the core difference between the stabilities of these two exchange media.

Moldbug starts out by imagining that there are 2 types of bitcoin users. First, there are "speculators" who hold bitcoin over time, hoping to earn a return. The second type of user, the "exchangers," only hold bitcoin for brief moments to engage in daily transactions, selling all coins to speculators at the market close. Only speculators, therefore, hold bitcoin overnight, presumably selling it back to exchangers the next morning.

If we abstract a bit from this, what Moldbug is really talking about is the two famous "functions" of money, that of serving as a store-of-value and a medium-of-exchange.

Now if speculators all flee the market at once, then desperate exchangers have no one to sell to come evening time. Bitcoin's overnight price falls to 0. Since bitcoin no longer has any purchasing power, the next day exchangers will find their bitcoin useless as an exchange media. Bitcoin becomes just bits. What might lead to this result? Moldbug hypothesizes that a government closure of the various bitcoin exchanges would spook speculators, causing them to all exit and drive the price down to 0.

I want to show why the same thing can't happen to central bank money. During the day, banks in an economy need large quantities of central bank balances for clearing and payments purposes. A central bank provides these balances to banks in return for collateral. At the end of the day, what do the banks do with these unwanted balances? Well, let's say that they sell them onwards to speculators to hold overnight. Speculators accept the trade because they think they can make a return. The next day the banks repurchase these balances in order to use them for their daily payments. This is very much like the stable bitcoin economy described above.

What happens when the speculators suddenly exit the market? Banks now have no one to sell their clearing balances to at the end of the day. As in our bitcoin case, won't the value of balances fall to 0? No. The issuing central bank will offer to buy all of the balances back.* With what? With the collateral that was originally used to buy them. Unwanted clearing balances will therefore be slurped right back up by the central bank. So long as the central bank holds adequate collateral, it will be able to suck every single clearing liability it has issued, contracting its balance sheet to 0.

This, in short, is why the bitcoin price is highly unstable whereas the price of central bank liabilities is highly stable. All that underpins the value of bitcoin is the presence of a few speculators in the market—whatever random event causes these speculators to depart will be the end of bitcoin. In the case of central bank money, the original issuer is committed to repurchase whatever is unwanted, even if speculators scramble to leave.

In real life, speculators typically don't hold central bank clearing balances overnight. Central banks usually repurchase all clearing balances back at the end of the day, returning the collateral to the banks so they can use it for the next market day. Central banks are like blowfish.** They blow themselves huge during the day in order to accommodate the needs of banks for clearing balances, then suck themselves tight at night when they aren't needed. Bitcoin is like a slowly growing amoeba. It can't contract itself when it needs to.
______________________________________________________________________________

PS: What about gold? If speculators all leave the gold market, demand will still be anchored overnight by those who desire gold for ornamentation, dental, and manufacturing purposes. Thus, when the market reopens next day, exchangers will find their gold still has a positive value, although probably far less than the night before.

Disclaimer: I am long bitcoin. Why? I'm curious about bitcoin and the best way to learn is by doing. Secondly, I'm speculating that before it hits $0, it could hit $50. There are a lot of people out there who don't yet realize that they'll be buying over the next months. Keynes's beauty contest and all.

*Central banks are required to ensure that the value of clearing balances stay moored to basket of consumer goods, or a CPI target. They typically do this by setting the overnight bank rate. If no one wants to contract to hold reserves overnight, the interest rate will collapse. The central bank will have to conduct open market sales - basically repurchasing clearing balances with collateral - in order to reduce the supply of clearing balances until the overnight interest rate has returned to its prior level. In any case, that's why a central bank needs to "suck" the money supply back in.

** I'm borrowing the imagery from Alex Tabarrok's 2008 post, although he uses a bullfrog.

27 comments:

  1. Why wouldn't the exchangers hold the bitcoin overnight? I hold notes and coins overnight. The exchangers would become speculators, if nobody else would pay them to borrow their bitcoin.

    ReplyDelete
    Replies
    1. It's just an assumption. You're right that we are all capable of acting simultaneously as exchangers and speculators. It seems useful to me to imagine a world in which we split these roles apart so that one sort of person plays only the first role, another sort only the second role, and then work through what happens when one or the other exits.

      Delete
  2. Speculators who go long in bitcoins will drive bitcoin up, but you have to remember that speculators who go short in bitcoins will drive it down. So in an efficiently functioning market, speculation might have no effect on bitcoin's value. (I suppose that restrictions on short selling might weaken its effect. Lots of complications crop up when we start down that road.)

    ReplyDelete
    Replies
    1. Mike,

      I agree with you that an efficiently functioning market would offset the longs with the shorts. However, it is currently difficult to hold a short position on Bitcoin, there is no sufficiently efficient market that has this functionality. There used to be one, and according to my empirical analysis, while it was operating, the value stabilised. Liquidity was higher and price volatility lower.

      I think that while JP is onto something, his approach is skewed by his experience with the central banking mechanism. In our fiat system, practically all money is debt. It is created to earn interest and this is the reason driving the behaviour of banks. This desire for interest payments drives liquidity. But Bitcoin is not debt. Its liquidity is a consequence of its low transaction costs. While its liquidity may be subject to higher shocks than that of a typical fiat money, as long as it has a comparative advantage in transaction costs, there will be demand for it.

      I've heard several people saying that if there's another big dip (i.e. if the Bitcoin price falls significantly), they will buy up all the Bitcoins they can. Bitcoin is thus not like an amoeba. It is like a black hole, a singularity, it sucks the whole world into itself while only gaining weight but not size. And while a blowfish might think it made the escape velocity, its perception might be affected by the doppler effect.

      Delete
    2. Mike, that may be. But given the setup of the above imaginary world, the only way for bitcoin to maintain its value overnight is the willingness of speculators to hold it. So speculators (those not holding bitcoin for transactional purposes) surely have an effect on bitcoin's value. Of course, our definitions may vary.

      Peter, intraday clearing balances don't pay interest. They are held purely for transactional purposes. Mind you, I do like the imagery we're amassing here - blowfish, amoeba, black holes, doppler effects.

      Delete
    3. Hmm, I'm not an expert in the operational details of the internal banking market, so I'll take your word for it. One more thing though to remember is that Bitcoin trades 24/7 worldwide, so the concepts of "end of the day" and "overnight" make no sense. If I had to hazard a guess, I would say it shows that Bitcoin is more mature than fiat, a truly global medium of exchange.

      Delete
  3. "But blowfish-like, a central bank can just as quickly suck this supply back in—indeed, a central bank can go to the extreme of extinguishing every last liability it has ever issued."

    So what if they buy a bunch of garbage debt that isn't worth nearly what they paid for it? Can they still buy back all of their liabilities afterward?

    ReplyDelete
  4. "PS: What about gold? If speculators all leave the gold market, demand will still be anchored overnight by those who desire gold for ornamentation, dental, and manufacturing purposes. Thus, when the market reopens next day, exchangers will find their gold still has a positive value, although probably far less than the night before."

    What if gold's non-monetary demand is so insignificant relative to its monetary demand that your argument applies equally well to gold, practically speaking, as it does to bitcoin? Would that make it a valid reductio ad absurdum?

    ReplyDelete
    Replies
    1. Yes, I'd agree with that.

      Mind you, all it takes is one manufacturer willing to buy a tonne of gold for $1 (and therefore world's entire stock of gold for $150) to anchor the price overnight so that the next morning exchangers can use it as a medium of exchange.

      Delete
    2. I have been wondering about the same thing myself.

      Delete
  5. Peter's empirical results imply that the introduction of short selling does not reduce bitcoin's value. That's the same result we see in all securities markets. Two weird implications:
    1. Bitcoin's value is not preserved by restricting short sales.
    2. But short-selling of money always creates a claim to that money (The claim would be properly called a hypothecated bitcoin.), and since that claim can be used to buy stuff, the introduction of short selling should reduce monetary demand for bitcoins and thus reduce its value. But given Peter's results, this doesn't happen. Conclusion: Bitcoin does not get its value from its moneyness, but from something else. What else? The two things that come to mind are (a) curio value (same reason we value old baseball cards) and (b) the speculators' belief that a greater fool will be along shortly.

    ReplyDelete
    Replies
    1. Mike,

      the price of Bitcoin is influenced by liquidity, but it's not the only factor. I think (and I would say that JP agrees) that liquidity is a more important factor than price. I'm not getting excited about changes in the Bitcoin price anymore, unlike when I started researching Bitcoin. I also researched correlation between liquidity and price, and the result was that liquidity drops as price rises. So my interpretation is that the high price spikes are bubbles. Unfortunately I don't have data yet for the latest developments, but my impression when I was looking at the charts during the spikes was that liquidity was lower, and that at the moment it appears to be slowly stabilising, but please consider this impression as anecdotal evidence.

      Also, as I said before, merely because something is a claim, it does not follow that market accepts it as a nearly perfect substitute to the underlying. This requires that the claim reduces transaction costs. Which is difficult to pull of if the underlying already has very low transaction costs, like with Bitcoin. So claims on Bitcoin, while they might trade at the same price as their nominal value, are unlikely to be widely accepted as a substitute medium of exchange.

      The absence of short selling facilities for Bitcoin is an empirical issue. It requires a trustworthy and skilled entrepreneur, or on the other hand a working decentralised system. The last big shorting market (Bitcoinica) unfortunately was plagued with operational mistakes and is currently in liquidation. But eventually I think a working replacement will be available.

      Actually, my research hints that the availability of the short selling facilities probably DID reduce Bitcoin's price. It kept it down and prevented a bubble.

      Delete
    2. Actually, I pretty much agree with Mike's conclusion --- although I'm not too keen on his way of getting there, via his short-sale theory, for the reasons same reasons as Peter --- it's very difficult for a BTC IOU to replicate what Bitcoin does in terms of transactionability. [Mike, why don't you buy some bitcoins? We'll walk you through it... it'll give you a better feel for the technical side of bitcoin.]

      Anyways, back to Mike's conclusion. If you look at the disclaimer I added to the end of my post, my motives for holding bitcoin are precisely Mike's two reasons. I'm trying to learn about this odd thing called bitcoin -- that's my demand for a curio. And I'm speculating -- it's gonna hit $50 before $0.

      Taking the entire universe of bitcoin holders, I'd assume that the split between curio/speculator is massively biased to the latter. I'm going to assume that curio holders only need to hold 1 bitcoin to be satisfied. If speculators leave the market and the BTC price plunges, there's probably no reason for curio holders to buy more curio value.

      [A note on definitions so as to avoid confusion: Peter notes that liquidity is very important to bitcoin's value, and I agree. Mike says that moneyness is not key to bitcoin's value, but that curio and speculative motives are, and I agree. Except for one caveat. I haven't really gotten into this, but I subsume the speculative motive under the category of moneyness/liquidity. Both the demand for a stock of liquidity and the demand for a stock of speculative assets are the same... in both instances, a stock of assets is being held for purposes of future sale, not permanent ownership. The difference between speculation and transactional demand is only in degree, since the transactor expects to spend it tomorrow, the speculator next week. Both are betting on a bigger fool, so to say, and both add to an asset's liquidity premium.]

      Delete
    3. That's interesting JP,

      I need to think about it more. But just for the time being, I agree that the purely speculative motive (i.e. hoping that the price will rise compared to, say, USD) is present, and that it will probably be present with some market participants any level of liquidity. But, as liquidity of a good increases, I would say that the proportion of people would tend to hold it for transactional (as you call it) purposes relative to speculative purposes is increasing. Anecdotal evidence (apart from my research, I'm also involved in the Bitcoin community directly and I've been in personal contact with many of the Bitcoin users) suggests that indeed the transactional purposes are gaining traction. But you don't need to take my word for it, check out the two recent news items:

      Bitcoin Casinos Release 2012 Earnings: http://www.forbes.com/sites/jonmatonis/2013/01/22/bitcoin-casinos-release-2012-earnings

      BitPay Surpasses 10,000 Bitcoin Merchant Transactions, Zero Cases of Payment Fraud: http://blog.bitpay.com/2013/01/bitpay-surpasses-10000-bitcoin-merchant.html

      This is consistent with the argument I made in my master's thesis, that we can expect Bitcoin to expand in those markets that can gain a significant improvement in transaction costs, or that are sensitive to transaction costs.

      I've been thinking lately, and I think that there are two reasons why money(ness) emerges: uncertainty and transaction costs. I think both are a necessary precondition for money(ness). But I haven't formed a fully coherent argument yet so I'll just leave it at that for the time being.

      PS. Both JP and Mike, you should read my master's thesis, I think it does a good job in explaining Bitcoin to an economist.

      Delete
    4. Just one more comment, I just remembered that Menger, in On the Origins of Money I believe, mentions speculative reasons as one of the components for a demand for a medium of exchange. So JP, you're definitely onto something when you categorise the speculative demand as a subset of liquidity.

      Delete
    5. I agree that uncertainty and transaction costs contribute to moneyness. Not only actual conditions, but anticipated conditions, drive the premium.

      Peter, your thesis is on my hard drive, have been meaning to read it fore a while.

      Delete
  6. "Mike, why don't you buy some bitcoins?"

    Tell you what JP. I'll borrow 1 bitcoin from you and immediately sell it to Peter. I will pledge to repay either 1 bitcoin + interest, or else the dollar equivalent. No cash will change hands for now, and I won't set an expiration date. We can settle up on our death beds. The swap will just add 1 short to the market (me), plus one long (Peter), while leaving your long/short position unaffected. Just in case bitcoin rises to $1 million, I will include an "up and out" clause that relieves me from having to deliver any more than $200 worth of assets in settlement.

    If bitcoin gets its value from its moneyness, then our deal will push bitcoin down be creating a tradable claim to bitcoin. If it gets its value from investors waiting for a greater fool, then this one short position should partly cancel that effect, since there is now 1 investor (me) who is waiting for a greater genius. If bitcoin gets its value as a curio, then this transaction will not affect it.

    But back to the modern paper dollar, which is the whole reason we care about bitcoin. It's clear that claims to dollars trade just as easily as green paper dollars, and that there is an active market shorting dollars. This tells me that the value of the dollar must be equal to its backing, and that it is therefore not fiat money.

    ReplyDelete
    Replies
    1. Mike,

      the short you create won't be treated as a liquid instrument by the market. It will be almost entirely ignored. This is due to a variety of transaction costs (cost and speed of settlement systems, reputation/homogeneity of the market price of the short, reversibility/repudiation, etc). I'm not saying it's impossible (indeed, there are historical examples where it did work), but you're neglecting the infrastructure. And this is precisely the point of Bitcoin: the historical infrastructure of central and commercial banks is technologically heavily obsolete and probably won't be able to stand up to competition (due to a high amount of fixed capital invested in the old infrastructure).

      Delete
    2. Peter:

      But surely if the folks at bitcoin were able to set up an efficient system for trading bitcoins, some other computer whiz could accomplish the even easier job of setting up an efficient network for trading claims to bitcoins. These claims might be issued in forms that are more convenient than bitcoin, at least for certain kinds of trades (e.g., the claims might be in the form of paper notes spendable at walmart).

      The issuer of each of these claims is short one bitcoin, and therefore profits as bitcoin falls. So if bitcoin does get its value from its moneyness, there's a huge arbitrage opportunity for some potential issuer of bitcoin claims. (Or for anyone who can issue any kind of rival to bitcoin.)

      Also, if bitcoin gets its value from long speculators expecting a greater fool to come along, then that effect will be offset by short speculators expecting a greater genius to come along.

      Delete
    3. Mike, are you saying that in the real world, speculative forces can't keep prices above fundamental for periods of time? How do you explain chain letter ponzi schemes, Bernie Madoff, or the NASDAQ internet bubble? While the EMH is a good baseline by which to start analysis, we need to be careful when we apply it to the real world. There are large gaps in time between the entrance of fools and geniuses.

      Delete
    4. Mike,

      the transaction system is a component of Bitcoin, just like, say, the words are a part of the language. The transaction system does really transfer bitcoins, it is Bitcoin. Bitcoin is a pseudo-commodity with teleport functionality. When trading Bitcoins against fiat, it is the fiat that is the weak link and causing problems. Trading Bitcoin against fiat is far from perfect. I don't think that even mechanisms like OpenTransctions or Ripple can fix this.

      What you suggest is only possible if you use another cryptocurrency. Then you could create transaction messages that offset each other and conduct the transfer upon a predefined trigger, and shorting will be possible at a level where transaction costs won't be relevant. But since other cryptocurrencies have lower liquidity than Bitcoin, it makes no practical sense.

      Delete
  7. His real name is Curtis Yarvin. He revealed as much when he linked another blog of his on programming:
    http://moronlab.blogspot.com/

    ReplyDelete
  8. The Rise Of Bitcoin What Is It And Is It Worth It? | Bitter Bananas

    Bitcoin is starting to get very popular these days, but still the majority of people know nothing about it. It is considered as one of the best investments you could have made, because of its’ massive price jump over the last year.

    So what is Bitcoin?

    This article best describes the overivew of Bitcoin, which is largest online used digital currency in online world.
    http://bitterbananas.com/the-rise-of-bitcoin-what-is-it-and-is-it-worth-it/

    ReplyDelete
  9. This comment has been removed by the author.

    ReplyDelete
    Replies
    1. I think I remember the gist of your question.

      Central bank liabilities are able to reflux back the issuing central bank because as Mike Sproul says, central banks hold assets.

      When the population wants to hold less paper bills, they bring these bills to their bank. The banks, which now hold too many bills in their vaults, send them in trucks to the local Reserve bank. In return they get electronic reserves.

      Most likely the banks will be unwilling to hold these excess reserves, so they'll all try to spend them away. If they do so, the federal funds rate will spike. In order to keep the ff rate on target, the Fed sells assets and buys reserves, thereby reducing their quantity and ensuring.

      Delete
  10. Look where we are now!

    ReplyDelete