I have been doing a fair bit of reading in the history of money and monetary theory to try and understand money, particularly its origin and use. A thing reveals its nature through history, to understand the history of something is to much better understand it.
I was aware that modern macroeconomics is bedevilled by a lack of a common analytical language in talking about money. But the problem goes deeper than that; money is a multi-dimensional phenomenon.
I am not talking here about debates over monetary base (coins, notes and bank reserves with the central bank) and the various monetary aggregates (M1, M2, M3 etc). These debates are even less interesting than they appear, not least because expectations matter so much.
Nor even that when economists talk of the “demand for money” they mean the demand to hold money, not the willingness to spend money (which, in a monetary exchange economy, is the demand for goods and services; and money actually spent on goods and services is aggregate demand). Though one way of thinking of money held is as a proportion of money passing through, and shifts in the proportion of money held rather than spent can be very important in its macroeconomic effects. Especially if the money supply does not adjust accordingly, because then an increase in the demand to hold money can lead to a fall in transactions and hence (since one person’s spending is another person’s income) a fall in incomes, spiralling down into those transaction crashes/goods-and-services gluts we call recessions or depressions.
The problem in talking about money qua money, is that there are so many institutional possibilities with money. Let us just consider the situation were notes exist, so cost of production is not an effective constraint (and ignoring coins and the history of money before notes).
A “theory of money” has to consider to following dimensions:
(1) Is there a monopoly supplier of local notes or are there competitive suppliers?
(2) Are the notes convertible into gold, silver or some other commodities?
(3) If not convertible, are the notes backed in some other way?
(4) If not, is the supply of notes limited in some other way beyond cost of production?
If (2) is true, then the price level (P) is determined by the ratio of monetised backing (for example, gold if a gold standard is operating) to output (y), since notes can always be “swapped out” for gold, so the level of monetised gold (mG) sets the swap value of notes for output. (So P = mG/y.) So, if monetised gold rises faster than output, prices rise. If monetised gold rises slower than output, prices fall.
Note that if (2) applies, then the swap value of the notes (in terms of gold) is set, but their supply is not.[i] This setting of the swap value by a fixed price in gold leads to considerable price stability, as the gold/output ratio generally changes fairly slowly.[ii] Changes in the rate of turnover of notes (i.e. in the demand to hold money) can easily be accommodated by changes in the supply of notes.
This price stability tends to lead to low levels of fluctuations in the rate of turnover/proportion of notes held–in Sweden, for example, in its gold standard era from 1873 to 1914, there was a steady decline (pdf) in the proportion of notes held (so gentle increase in rate of turnover).
If (2) does not apply, then the above is not how things work. So convertible or not convertible becomes a vital issue for effective analysis.
If you have a monopoly supplier of notes that are not convertible nor backed nor otherwise supply limited–so (2), (3) and (4) do not apply–then you have hyperinflation, for the reasons explained nicely here (pdf) using the example of the German 1922-3 hyperinflation:
throughout the hyperinflation episode the Reichsbank’s president, Rudolf Havenstein, considered it his duty to supply the growing sums of money required to conduct real transactions at skyrocketing prices. Citing the real bills doctrine, he refused to believe that issuing money in favor of businessmen against genuine commercial bills could have an inflationary effect. He simply failed to understand that linking the money supply to a nominal variable that moves in step with prices is tantamount to creating an engine of inflation. That is, he succumbed to the fallacy of using one uncontrolled nominal variable (the money value of economic activity) to regulate another nominal variable (the money stock).
Putting it in terms of the Fisher-Friedman equation of exchange (money stock x rate of turnover of money for goods and services = prices x output: i.e. MV = Py) as, in this situation, notes are neither supply nor demand constrained, M increases exponentially. The swap value of notes for goods and services plummets. So money turnover increases towards maximum practicable as people unload notes as quickly as possible due to notes plummeting as a store of value and people generally cannot avoid being paid in the monopoly notes. If M and V are both increasing (the former at an accelerating rate), prices head towards infinity, as prices are much more upwardly flexible than output.
If all these conditions do not apply, then you do not get hyperinflation. Which is why worrying about the size of the monetary base in the US leading to hyperinflation is so much blather. There is no way the US Federal Reserve is going to issue an accelerating supply of notes if prices start rising. Indeed, it has shown that it is able to keep inflation down no matter how high the monetary base gets; markets are also expecting that inflation will remain low.
To put it another way, we have a monopoly supplier of local notes (check the signatures on your banknotes) which are not convertible (except in foreign exchange markets) but are backed (by the assets of the central bank). The backing sets the supply of money. The notes issued are the intermediary between the asset backing and output. The ratio between output and the supply of money (adjusted for changes in rate of turnover; so what counts is money spent on goods and services) sets the price level.[iii]
Of course, if note issue is competitive, people can insist on the notes which are better stores of value. Indeed, it has been reasonably argued that competitive note issue would drive the issuers towards convertibility, for precisely that reason.
So, by changing which of conditions (1) to (4) apply, we get profoundly different outcomes.
If people do not specify clearly what conditions they are talking about in analysing money, then folk will misunderstand, talk past each other and generally get involved in debates which never get satisfactorily resolved. For example, disagreeing or agreeing with the quantity theory of money without specifying whether, for example, (2) applies is unhelpful. Money matters, but it is also a multi-dimensional phenomenon. There is no such thing as “just money”.
[i] So any analysis which is predicated on the proposition that the gold standard controls the money supply is mistaken. There may be a legally imposed “cover ratio” (a, usually minimum, ratio of gold held to notes issued) which, if followed, affects money supply, but that is not an inherent feature of the gold standard.
[ii] Unless you are the Bank of France and the US Federal Reserve in the 1928-1932 period and you are treating an international monetary regime as a game of “who can suck more gold out of the monetary system”, so causing the Great Depression (pdf) by reducing the supply of monetised gold, so driving down prices (since monetised gold shrank dramatically compared to output) causing a massive transactions crash/goods-and-services glut.
[iii] I am agreeing with Nick Rowe and disagreeing with Mike Sproul (pdf) on the law of reflux (the proposition that notes issued in excess of the demand to hold notes reflux straight back to the issuer). Also, backing strikes me as fundamentally different from convertibility since, with convertibility, the swap value of notes for gold is a set quantity per unit; the price is set as an external (exogenous) anchor. (A currency board uses a set exchange rate as such an anchor.) In the case of backing without such convertibility, the backing assets generally do not have a set external (exogenous) price. But I found Mike Sproul’s papers very helpful.