Money Mischief_ Episodes in Monetary History - Milton Friedman [73]
If the quantity of goods and services available for purchase—output, for short—were to increase as rapidly as the quantity of money, prices would tend to be stable. Prices might even fall gradually as higher incomes led people to want to hold a larger fraction of their wealth in the form of money. Inflation occurs when the quantity of money rises appreciably more rapidly than output, and the more rapid the rise in the quantity of money per unit of output, the greater the rate of inflation. There is probably no other proposition in economics that is as well established as this one.
Output is limited by the physical and human resources available and by the degree of knowledge and capacity for using those resources. At best, output can grow only somewhat slowly. Over the past century, output in the United States grew at the average rate of about 3 percent a year. Even at the height of Japan's most rapid growth after World War II, its output never grew at much above 10 percent a year. The quantity of commodity money is subject to similar physical limits, though it has at times grown more rapidly than output in general, as the examples of the flood of precious metals from the New World in the sixteenth and seventeenth centuries and of gold in the nineteenth century illustrate. The modern forms of money—paper and bookkeeping entries—are subject to no such physical limits.
During the German hyperinflation after World War I, hand-to-hand money increased at the average rate of more than 300 percent a month for more than a year, and so did prices. During the Hungarian hyperinflation after World War II, hand-to-hand money increased at the average rate of more than 12,000 percent a month for a year, and prices at the even higher rate of nearly 20,000 percent a month (see Cagan 1956, p. 26).
During the moderate inflation in the United States from 1969 to 1979, the quantity of money increased at the average rate of 9 percent a year and prices at the average rate of 7 percent a year. The difference of 2 percentage points reflects the 2.8 percent average rate of growth of output over the same decade.
As these examples show, what happens to the quantity of money tends to dwarf what happens to output; hence our reference to inflation as a monetary phenomenon, without adding any qualification about output. These examples also show that the rate of monetary growth does not have a precise one-to-one correspondence to the rate of inflation. However, I know no example in history of a substantial inflation lasting for more than a brief time that was not accompanied by a roughly corresponding rapid increase in the quantity of money; and no example of a rapid increase in the quantity of money that was not accompanied by a roughly corresponding substantial inflation.
A few charts (Figures 1–5) illustrate the universality of this relation. The solid line on each chart is the quantity of money per unit of output for the country in question, year by year for the various periods. The other line is a price index—either a deflator or a consumer price index.* To make the two series comparable, both have been expressed as percentages of their average values for the period as a whole. Moreover, the vertical scale is logarithmic; that is, equal distances record equal percentage changes. The reason is straightforward. What matters for prices is the percentage change, not the absolute change. A $1.00 increase in price is a far more drastic change for an item initially priced at $1.00 than it is for an item initially priced at $100; the price of the $100 item would have to double to $200 to undergo a comparable change. The scales on the various charts are not identical, but they are comparable, in the sense that the same slope corresponds to the same rate of inflation.*
The two lines in each chart necessarily have the same average level,