For pushing five decades now, the United States has not defined its currency in gold. Prior to 1971, the nation nearly always had. From the 1790s to the 1930s, the U.S. defined the dollar as 1/20th an ounce of gold. From 1934 to 1971, that definition was 1/35th an ounce. In 1971, President Richard M. Nixon declared that the United States would no longer pay official respect to gold. This new dispensation has prevailed since.
The United States defined its dollar in gold all those years, until recently. Not backed, defined. There is a difference. In the difference lies the reason that the United States no longer has a gold currency.
To define a currency in gold, the issuer names a price in gold, such as 1/20th per ounce, and the deed is done. There is no need for the issuer to have any gold—none. In defining a currency in gold, the issuer proposes that the currency can be redeemed at the issuer for gold at the par value. When redemption claims come in, the issuer enters the market to buy the necessary amount of gold. Given a credible currency, the market trades in gold for the currency at the par price, and the transaction clears.
The idea that a currency issuer should have a large stock of gold on hand to cover redemption requests is a dodgy one. A large stock of gold held by the issuer implies that the issuer does not believe that the market will sell gold to it at the par price. A small vault reserve for day-to-day requests, perhaps—a large one suggests that the currency was overprinted, that the issuer knows that the float exceeds real demand, and that the market will not honor the par price.
Traditionally, the best currency issuers did find themselves accumulating gold. This was because gold holders wanted the best currency for their gold when it came time to use their money, their gold, for investments or to buy goods and services, markets which require the convenience of currency for transactions. Hence the British, and in time the United States, soaked in gold from all sorts of people who wanted pounds and dollars. The issuers often used this revenue as a pretext for limiting taxation.
Things changed with World War I, when the neutral nations, including the United States, piled up enormous gold stocks, because the belligerents were only buying, not selling, goods and services. The neutrals soaked up so much foreign exchange, with no prospect of using it for purchases, they made gold redemption requests from the belligerents that made their official gold stocks soar.
This strange development gave rise to the notion, a false one, that under a gold standard, a currency must be “backed” in gold. That is, the issuer has to have a large fraction in gold of the currency float it has issued. A standard view is that the gold “cover” should be 25%. If the United States has four trillion dollars outstanding, it should have a trillion dollars in gold at the par price in reserve.
This notion is illogical. It implies that the demand for currency should track, quite exactly, the production of gold. There is no scientific reason that this should be the case. World gold production has always crept along at the same rate, perhaps 1% per year. During the industrial revolution, global growth (a proxy for money demand) was regularly 5% per year. Therefore, the amount of currency cannot be a function of how much gold the issuer has.