The Structure of Moneys

(Posted: Friday 22nd January, 2021)

(back to article index)

I seem to find myself explaining on a regular basis the fundamental difference between gold and Bitcoin, and why I've bothered to invest in the former when the latter exists. Plus, what is money anyhow, and why are pounds or dollars worth anything? So, given that the only place I've seen the theory behind this explained is in some particularly dry and long-winded passages scattered through the Wealth of Nations, I shall attempt to present it in a more accessible manner.

What is money for? Conventionally, economists list three (sometimes four) functions of money: store of value, medium of exchange, and unit of account. Of these, 'unit of account' is not a reason to hold money, so does not generate demand, but the other two do. As a store of value, one clearly needs to hold the money in order to store value in it; to use it as a medium of exchange, one generally needs to maintain a cash balance to buffer between buying and selling at slightly different (uncertain) times. Since fulfilling these functions creates value for the holder, this is a source of demand, and in equilibrium will drive up the price of the money by an amount (the 'agio') equal (on the margin) to the value created. Typically, the wider the acceptability of the money, the greater the agio, since a money which all one's trading partners will accept is more useful than one which some will refuse to take.

So what of the various possible moneys? Let's start with the case of gold coinage. Gold has what is sometimes misleadingly referred to as 'intrinsic value'. It's pretty to look at, so there is demand for it in ornamental uses; there is also some demand from industrial uses (dentistry, electronics, etc.). None of this is really 'intrinsic' per se; these uses only have value because people value them, and if suddenly everyone decided that shiny things were ugly, the 'intrinsic' value of gold would fall. But in addition to this, there is also demand for gold as money — it's a lot more convenient than barter — and this additional demand creates an agio. Gold coins are more convenient for this purpose than bullion, so they have a further agio over the bullion price. We can take this further by using claims on stored gold, rather than physical pieces of gold, as the exchange token, whether that's in the form of gold-backed notes or digital gold; these may carry an agio from being still more convenient than gold coins (e.g. it may be harder to forge the note than to disguise some alloyed coin as fine gold) but may also carry a discount from counterparty risk (e.g. the possibility that the issuer may fail to redeem the notes for gold on demand). But whether we consider gold bullion, gold coins or gold notes, in each case the money has a value consisting partly of agio and partly of use value.¹

Contrast this with the fiat moneys issued by nation states, such as US Dollars or Pounds Sterling. While many of these currencies were historically backed by precious metals, today they are not — a £5 note can only be redeemed for another £5 note, not for (say) a defined weight of sterling silver. The value of a fiat money comes entirely from agio: dollars are useful as a currency because lots of people will accept them in payment, and people will accept dollars in payment because they are useful as a currency. A fiat money is a Schelling point, whereby we all agree that this thing has value, so that we can collectively obtain the advantages of having a medium of exchange and a store of value, without having to expend effort digging shiny stuff out of the ground. As long as that Schelling point remains stable, this works; but there is nothing to stop a central bank from issuing more money than is demanded at the current price, thereby producing inflation (reducing the value of the currency until the total value in circulation again matches the demand). If a bank tried the same thing with gold notes, the excess notes would be redeemed for gold, while if a mint produced too many gold coins, the excess would be melted down; but the lack of any way to redeem fiat money means there is no way for the bank's excess issue to return to it. Of course, nation states nowadays have a habit of being in debt, often to eye-watering sums, and this gives them an incentive to produce this kind of inflation, which they sometimes resist in the interests of long-term economic stability.

Let us now turn to cryptocurrencies. There are some that are backed by other assets (e.g. backed stablecoins), but the most visible ones (such as Bitcoin) are generally free-floating and, like national fiat currencies, derive their whole value from agio. An interesting case is Ethereum, where (as I understand it) the currency is effectively² redeemable for a slice of (distributed) computation on which to execute smart contracts; this could be seen either as 'a currency backed by cycles' or as a more flexible form of agio (the money is valuable to the extent that people wish to execute contracts denominated in it, rather than merely transact simple exchanges in it). But generally speaking, cryptocurrencies are usually 100% agio, making them fiat currencies.

Why does any of this matter? One reason is the bootstrap problem: suppose you wish to create a new currency; how can you entice merchants to accept your currency, when it does not already have an agio from being accepted by other merchants? National fiat currencies generally solve this by making taxes payable in the national money, thus creating an initial agio 'by fiat', or by un-pegging a previously metal-backed money (relying on their sovereign immunity to effectively default on their previous promise to redeem the money for precious metals). Precious metals can bootstrap a currency by virtue of having a use value even in the absence of agio; so can other commodities (iron, say) but the greater convenience of value-dense commodities like gold and silver makes it easier for them to build an agio on top of this. Bitcoin bootstrapped by speculation: enough people came to believe that it would be a widely-acceptable money, that they bought in as a speculative investment, thereby giving it a high enough price to become an acceptable money.

But pure agio is fragile, since it depends on mindshare. Like a network, it's only useful because other people are using it, and if enough people get spooked and turn to alternatives, the agio drops. Fiat currencies can weather this by increasing taxes and thus boosting the "you need this money to pay your taxes with" agio.³ Commodity-backed currencies have a less volatile mindshare: if (say) the price of gold is 80% agio, then a halving of mindshare should only reduce the price by 40%, and that in turn means holders need be less fearful of a drop in mindshare. But a speculation-bootstrapped fiat currency like Bitcoin is very susceptible to a snowball effect: if mindshare drops, agio drops with it, which further drops the mindshare as speculators lose confidence; momentum (whether upward or downward) is self-sustaining. There is no 'bedrock' of non-agio value to reassure holders or to rebuild from after a collapse; the knowledge that re-bootstrapping will be difficult increases the likelihood that it will be needed.

Of course, this may all prove irrelevant. Bitcoin may keep on going up and up until it replaces the US dollar as the 'world currency' (just as the dollar previously ousted sterling); the transaction volume may grow large enough that speculators, with their volatile 'animal emotions', cease to have major effects on the price. But one only has to look at a graph to see that it's not there yet.

And how could cryptocurrency replace national currencies, anyway? The pull factors are not strong enough alone; there would need to be something pushing individuals away from dollars, and the most obvious candidate is a collapse in the latter's value. But that is precisely the situation in which people would become wary of trusting another fiat currency! I think that, if the nature of fiat money and agio is widely understood, the most likely beneficiary of a collapsing dollar would be, not Bitcoin, but gold.

And that's why — full disclosure time! — I'm long 13 ozt of gold in a vault in Zurich.

Update Friday 29th

There's momentum building for a possible short squeeze on silver. Obviously all of the above applies to silver just as well as gold; given that gold makes a better currency and thus has a bigger agio, the only explanation I can see for the greater volatility of the silver price is if there's a lot of short interest. So I'm now long 4kg of silver too, just in case this one goes parabolic.

  1. In equilibrium, the agio of each money should equal the marginal cost of producing it from the next money down — e.g. if coining is unrestricted, the agio of gold coins over bullion is capped by the marginal cost of minting gold coins, since if it rose higher arbitrageurs could profit by minting additional coins. Historically, many nations legally prohibited private coining, allowing their mints to earn a seigniorage by restricting supply and thereby raising the agio, at least domestically, above the cost of coinage. As with other cases of monopoly pricing, this reduces overall economic efficiency.
  2. There's a lot of indirection here. But I think this is how it bottoms out.
  3. Adherents of MMT would describe this as taxation controlling inflation by removing money from circulation; but the theory I present here suggests that merely increasing the uncertainty in the amount people will be taxed, thus requiring them to hold more of the money to ensure liquidity, would have the same effect. Moreover, increasing both taxation and spending together should have no net effect according to MMT, but by increasing the volume of transactions necessarily in the currency, on my theory it should raise the agio.