Monday, November 5, 2012

A bit of bother for bitcoin.

Digital currencies were considered a fringe idea unworthy of serious consideration a decade ago. Yet,  with the level of uncertainty about how the Euro crisis is going to play out, questions about currencies like Bitcoin are being contemplated by central bankers and mainstream economists.

The digital currencies are distinct from electronic payment systems like PayPal or Visa in that instead of negotiating the transfer of money from one place to another, electronic cash systems are designed to _store_ value instead of transmit it.

There are a number of criticisms of e-cash systems, which I won't go into here, but there is one I think will be a deal breaker is something that appears to be intrinsic to any system of algorithms.

But first, there is the question of whether any store of value can be sustained without some kind of market enforcement mechanism. Even gold is backed by assayers so that banks don't fill their vaults with fake, tungsten ingots. The main threats to the value (and hence, purchasing power and usefulness) of a currency are not necessarily debasement and counterfeit. Hot capital flows and trade imbalances are more likely to cause currency shocks than efforts to undermine the money supply. The actual consequence of undermining the currency causes greater swings, but these efforts are less frequent. It's what people call "tail risk" these days.

The premise that Bitcoin and other crypto-currencies advocate is that they can develop sufficiently powerful encryption algorithms that, notwithstanding their physical implementation details, can mitigate these counterfeit and debasement risk exposures by ensuring the integrity of the e-currency mathematically for the foreseeable future.

It seems like a better bet to trust mathematicians then politicians, except there is a fallacy in this that e-currency advocates seem to have missed. The short version is that we can trust mathematicians to do their best, but politicians will always win...

The technical problem of breaking codes, while difficult, is not an adequate substitute for the political problems of currency debasement, capital controls and financial repression. Solving one does not necessarily solve the other.

For e-curency to "work," unfortunately it must first carve out an impregnable niche of political legitimacy that will protect it from these very abuses it was designed to resolve.  Since Bitcoin is by its nature antagonistic to political control, every e-cash transaction will have to price in a premium of political risk in the form of reduced purchasing power. Any government could ban it tomorrow without significant political consequence, and people found to be transacting in it could be rounded up and charged with trafficking in contraband, laundering or evasion, or whatever else is expedient. Even if this weren't likely in the immediate term, the political risk makes putting your savings in it seem reckless.

The main argument I have heard from technologists is that with strong cryptography and a peer-to-peer network of untraceable onion-routed nodes, hence secrecy,  the political risks are all but eliminated. I am sure this was a similar argument to what UBS and other Swiss banks have been giving their clients for 60 years. Banking secrecy is apparently written into the Swiss constitution, and therefore the identities of clients, and in turn their personal wealth, would be safe for foreseeable generations. Swiss bankers held off the Nazi's, the Soviets,  the English, the Chinese and every revolutionary government in recent memory. They even kept out of reach of successive US administrations and its vast intelligence apparatus.  And for the last century or so, it worked. That is, until it didn't.

Political risk exposes it to enhanced volatility proportionate to the secrecy it requires. The implicit risk premium on holding Bitcoins for any period of time will create significant distortions in exchange rates and add that risk premium to prices denominated in it. Even if e-cash were used only for brief periods -  the way we use paper cash now by holding it for a day or two before depositing it or trading it for goods - it reduces the unique function of e-cash to that of an anonymizing proxy - and again, one that antagonizes authorities.

Cryptography, like any information system, has tremendous exposure to massive, radical, instantaneous change. When a given crypto implementation gets broken, tens of millions of devices that use that cipher scheme are vulnerable to anyone who knows how to exploit it. This happens all the time, officially once every few years various standards bodies declare certain cipher schemes obsolete, but the cipher implementations used by different businesses are broken every few months.

The deal breaker for crypto-currency is that in the eventuality that the Bitcoin cipher or any of its peers are compromised, after an instant wave of algorithmically driven counterfeit, there will be a complete liquidity collapse. To zero. That is, nobody will accept the e-currency until trust is restored in it. The problem is there is no obvious way to restore the trust. Both the beauty and irony of an e-currency system is it does not require trust in anyone because trust was based on the integrity of a proof, and liquidity: the two things that evaporate after a crack.

How do you breathe trust and liquidity back into a currency scheme? Peg it to the dollar or the euro? Create a central e-currency bank to loan it out and sell bonds denominated in Bitcoin? The answers are interesting, and it's possible that bitcoin could be a useful model for thinking about what "real" currencies are going to have to do. The difference is that Bitcoin lacks viable institutions whose business is denominated in an e-currency.

In a decentralized e-currency scheme, there is nobody to step in and provide either trust or liquidity, precisely because the system is designed to function without the political legitimacy necessary to bootstrap it.

Even if someone will likely appear, e.g. in the form of some geek who creates a new proof, it is likely the damage may be done.  The human element here is that even if the e-currency system can be fixed after it has collapsed, given the consequences of failure and the alternatives available now, why should anyone expose themselves to this kind of tail risk through a significant investment? Something to gamble on, but not to put real capital in. It's a bit of a bootstrapping problem in that there is probably a natural limit to the amount of liquidity in such a sound-crypto-money e-currency system -  while there is nothing offsetting this risk of collapse.

The volatility of an e-cash equivalent could be similar to that of common stock. It's fine under specific and relatively controlled circumstances, like a merger where both parties can influence its value and price, but stocks are notoriously hard to value in a way that would get you the same loaf of bread from one day to to the next.  Stocks also require a "greater fool" to redeem value from them, and in the event of a bitcoin crash, useful fools and market makers may be in limited supply. This is precisely because the currency is designed to function without institutional backing. Relative price stability is a goal of most currency systems, and for now, e-currencies just seem to be arbitraging tail risk - or in simpler terms, getting in while the ponzi scheme is winding up. 

If stories about the Silk Road online black market are true, it is possible that drugs transactions are the main source of liquidity for Bitcoin, in that the currency may derive much of its value from being readily exchanged for pot. If you can always buy dope with Bitcoin, then you can buy other things, because worst case, anyone can always redeem their Bitcoin for pot. Maybe that's the problem it solves. It's a black market currency backed by liquidity in contraband. If that's the case, it could sustain itself as long as there are pot heads.

Our physical fiat currencies have a similar elliptical logic that faith begets liquidity, which begets faith, but these have the slight advantage that a bunch of governments and institutions derive their legitimacy and power from their promises to make payments denominated in a given currency. The perhaps unintended consequence of the multi-trillion dollar derivatives market and shadow banking system has been that it created a Mexican standoff between banks and governments that prevents any of them from acting unilaterally to debase the currency. It has been surprisingly stable.

E-currencies are useful for modelling technical money problems, and they may even someday serve as a kind of structured investment vehicle that tokenizes and proxies anonymous transactions, like a magic cookie that facilitates a transaction, but as a store of value e-currencies are a simulacrum, a copy of a copy. Computational work is not physical work; e-currencies may represent a proof of work, but a symbol is not a substitute.

The only institutions likely to back such a decentralized currency would operate in black markets. It is not intentional that this presumes criminality in bitcoin transactions, but as soon as you operate in the realm of cash and untaxed anonymous transactions in grey areas, you become vulnerable to prosecution on a variety of fronts.

What Bitcoin and other e-currency projects ironically demonstrate is that people are in fact not free to trade between themselves without political accountability. Regulations aren't just a list of verboten contraband to protect people from themselves, the e-currencies demonstrate that the only legitimate economic activity is sanctioned and supervised by the state. The rest is in black markets. The limits of the crypto-currency schemes show that there is in fact, no escape from state supervision.

It's good to have this settled, but it suggests there could be some very pissed-off libertarians when they figure this out.

No comments:

Post a Comment