Saturday, March 15, 2008

Money 2

My first post on money.

There is an enormous difference between having a strong argument that a statement is true and proving incontestably that the statement is true. Mencius wrote an interesting post arguing that the world is likely to converge on a single monetary standard, that standard will be the gold standard, and that therefore hoarding gold is likely to be a good investment. Mencius is a bright guy with a lot of interesting ideas, but I think he is wrong here, but wrong in an instructive way.

To begin with, I don't think he really gets the idea central to my first post that money as measurement of value and money as medium of exchange are two fundamentally different things. That is, it is one thing to have prices expressed in units of gold ounces, and something different to generally pay for goods with gold coins. If payment usually is in the form of gold coins, it makes sense that prices will be expressed as a quantity of gold, but the reverse is not true; even if all the world expressed prices in terms of gold, people would likely frequently use some sort of certificates rather than actual gold coins in commerce. Certainly nobody would send payment in gold coins by mail!

Second, I think the strength of the argument for a single money standard is exaggerated. Nick has a very good post where he talks about the advantages of having a currency standard in terms of mental transaction costs. No doubt most people will want to usually use one (or very few) currency standards most of the time, but that's not at all the same as everyone using one single standard all the time.

Finally there is the time element. Prices in a store for immediate purchase can be expressed in any unit, but for contracts that extend over a long time it is important to have currency units with a relatively stable purchasing power, or at least one that will vary slowly and in a predictable way, and I don't think gold is suitable for this task. Gold is a commodity like any other, and is subject to fluctuations in value not merely because of changes in production and industrial demand for gold, but because of changes in peoples' desire to hold "money".

I think the description as to how the gold standard is to come about is also fundamentally wrong. In the story, Sven the fisherman exchanges his fish for gold. Why exchange them at all? because his customer (unnamed, we'll call him Olaf) doesn't have the goods Sven wants, and if Sven just holds on to his fish, they'll rot. Why gold rather than silver? Because it is generally believed that gold will appreciate in purchasing power faster.

There are two fatal flaws with this theory. First, given that gold and silver are both held for their exchange value, it makes no sense that gold and silver are exchanging for a certain rate today but everybody knows that one will be able to get more silver for the same quantity of gold tomorrow. The anticipated future exchange rate largely determines the current exchange rate. Second, if it somehow could be true that everybody knows gold appreciates faster, it is true that Sven would rather be paid in gold, but Olaf would rather pay in silver for the same reason. Why the assumption that Sven "wins"?

Mencius and "John Law" seem to be anticipating a future in which gold continually appreciates faster than other investments. I very much doubt that such a sustainable state of affairs could exist, but if it could, gold would not be used as money because gold holders would not surrender it except under exceptional circumstances. Buying gold because of the self-fulfilling prophecy that others will expect gold to appreciate rapidly and therefore will buy gold causing it to appreciate rapidly is a very dangerous game. The argument that one should sell gold because others will expect gold to fall and hence will sell their gold causing it to fall makes every bit as much sense.

So there are at least two reasons why the user might want to use two or more kinds of money. First, people are risk averse, so given the uncertainty in future purchasing power, one might prefer to hold both silver and gold (or assets denominated in silver and gold) rather than all one or the other. Second, one might accept the type of money being offered even if one would prefer some other type if the alternative is to forgo a profitable trade.

From the point of view of the issuer, how to denominate currency may depend on what one has, or what one can reliably get. For example, if I own a silver mine, it makes sense that I might issue silver coins, or certificates redeemable for a quantity of silver. If one actually has to come to my mine in Nevada to get the silver, it make sense that people might continue to use the certificates as money rather than "cash them in" unless they actually have a use for silver.

I believe that pure electronic money will become important in the future. Not only will the money be transferred and stored electronically, but it will be "backed" by a guarantee to perform services over some network in some predefined way. Why would somebody give up actual physical goods for such "money"? Because they either desire the "backing" services themselves, or because they can exchange the money for good and services they want. All that is necessary for the money to have value is for someone somewhere along the line to actually desire the services, and even that is only necessary to get the system started.