In his comment, Ralph Benko, editor of the website The Gold Standard Now, defends the gold standard, or at least alludes to defenses. One argument is that, historically, use of the gold standard correlates to economic growth. Benko and his unnamed advocates "suspect, with good reason, that there is more than a coincidental correlation." Benko also asserts that my criticism of the gold standard as limiting economic growth "conflates quantity and quality theory of money." Benko compares the gold standard to abstract units of measure, and implies that holding money to a gold standard does not affect economic growth any more than holding the definition of a yard constant affects the production of yardsticks. These defenses, however, are facially thin.
A suspicion of causation is not very impressive. Correlation is evidence of causation, but it is at best only preliminary evidence, and tells us nothing about the direction of causation. For example, if changing economic conditions caused the gold standard to begin to limit economic growth, or if some underlying change in economic conditions caused both slower economic growth and the inefficiency of the gold standard, then we would expect to see the same correlation: declining economic growth at the same time the gold standard was being eroded. Benko suggests that the correlation might have something to do with Total Factor Productivity, but I can't seen any plausible connection. Benko admits this defense is just a suspicion, so I think it's justified to not take it too seriously.
The more substantive objection is that I'm conflating quality with quality. But this objection at least needs more elaboration. Money is not an abstract unit of measure, it's a reified social construction that we pretend has actual existence; we actually pretend to store and move money in time and space like we move physical objects in space. There's nothing wrong with that pretense; pretending to store and move money around is just a way of modeling and representing the storage and movement of real goods and services.
In one sense, the quantity of money doesn't matter to money as a representation of transactions of real goods and services, only its quality. If we could move money arbitrarily quickly, then one unit of money would suffice, so long as that unit had a consistent representation. "Wealthy" or "poor" people would just be those through whose hands the unit of money passed relatively more or less often. But money cannot move arbitrarily quickly, for two reasons. First is just physics: it takes time, energy, and effort to move money, even by computer; doubling the actual units of money in use halves that effort, because moving two units in one transaction has about the same cost as moving one unit.
The second reason is because of money's other use: as a liquid store of value. We don't just move money around to represent the movement of goods and services, we store money to virtually transfer consumption in time (by actually transferring it in space between different people: my saving must be your excess spending, and vice versa).
Because money must be stored, the quantity of money is directly related to its quality. Indeed, as we know, Y * P = M * V: real GDP (Y) times the price level (P) is, by definition, equal to the quantity of money (M) times the average velocity of money (V). Everything else being equal*, increasing the quantity of money increases the price level in the short term, as more M is chasing the same amount of Y. But the price level is the "quality" of money. Thus (under ordinary circumstances) there is a direct relationship between the quantity and quality of money.
*Which is sometimes not the case: when interest rates are near the zero lower bound, for example, the velocity of money just decreases when the money supply increases, without changing GDP or the price level.
There is, however, a third issue: socially constructed agreements directly denominated in money, i.e. purely financial agreements. When I borrow money, I both borrow and agree to repay only money, not any actual goods and services that money represents. Financial agreements must involve positive nominal interest rates: no one would ever loan money at a zero or negative rate; they would just hold the cash. This means that if the price level decreases, people who owe money would owe more goods and services than they originally borrowed. Since interest rates cannot be negative, there is no easy way to reflect expectations of future deflation in interest rates. This is why real economists fear deflation more than (modest) inflation: deflation — or even just the expectation of deflation — causes money to stop moving, and the way our system is presently set up, money in motion is economic activity. Deflation causes a decline in real GDP. Because inflation can be easily reflected in interest rates, modest inflation does not affect real GDP in as dramatic a way as does even the smallest deflation.
What does all this have to do with the gold standard? Under the gold standard, M is the physical quantity of processed gold in the hands of the government and private individuals, and its increase (or decrease) cannot be socially constructed. (If Benko means something else by a gold standard, he will have to explicitly describe it.) Assuming the velocity (V) doesn't change (under ordinary circumstances it doesn't), the quantity of gold (M) must represent all the real goods and services we're capable of producing (divided by a constant V) plus all the real goods and services people want to store. Given that there's no particular reason to believe that real GDP will increase at the same rate that the quantity of gold increases, then either P or Y must change. But P, the price level, is the "quality" of money. It is how much goods and services a unit of currency will produce. So, if we let price levels float, the quality of gold will vary just as does the quality of a fiat currency; it's not fixed objectively. Alternatively, if we demand price stability, then we must match real GDP to increases in the money supply, which is just having the tail wag the dog. Either gold doesn't actually do anything macroeconomically significant, or it does exactly the wrong thing.
Not only does a gold standard have zero or negative macroeconomic effects, it also cannot limit government spending. The government can effectively create a monopoly on gold, simply by demanding whatever taxes it wants in gold, and paying for goods and services in gold. By changing the ratio of the two, the government can set the price level to whatever it wants: an ounce of gold is "worth" whatever the government will pay an ounce of gold for, which people will provide because they need the gold to pay their taxes.
So, demanding a gold standard seems as weird and arbitrary as demanding that all legal documents should change to using the Courier font for all legal documents or declaring that a court does not have jurisdiction because the flag in the room has a fringe. It really doesn't matter if our price levels are arbitrarily denominated in ounces of gold or simply fiat dollars. The demand for a gold standard, therefore, is disingenuous. It has nothing to do with "sound" money, since gold is just a slightly less efficient way of doing what fiat currency already does. It is, instead, aimed at delegitimizing the democratic republic, in favor of placing political power in the hands of the people who happen to own a lot of gold right now.