Money Mischief_ Episodes in Monetary History - Milton Friedman [11]
What matters ultimately is the real, not the nominal, magnitudes. As a result, the nominal yield on assets such as bonds has tended to adjust over long periods in order to keep the real yield roughly the same. However, that has been very far from the case over short periods, because of the time it takes for people to adapt to changed circumstances.
For cash balances, one cost—the cost that has been stressed in monetary literature—is the interest return that is sacrificed by holding cash rather than "safe" interest-earning assets, for example, the interest received per dollar of a U.S. Treasury bill as against the interest, if any, received per dollar of cash balances (zero for currency).
Another cost or return—and one less stressed, though often far more important—is the change in the real value of a dollar. If prices are rising at a rate of 6 percent a year, say, $1.00 will be able to buy only as much at the end of the year as 94 cents would have bought at the beginning. To keep the real value of your cash balances constant, you would have to hold balances of 6 percent more at the end of the year than at the beginning. On the average, it would cost you 6 cents for every dollar that you held during the year. Conversely, if prices were falling at the rate of 6 percent a year, you would in effect receive a return of 6 cents for every dollar you held during the year. Clearly, cash is a less attractive asset when prices are rising than when they are falling.
For a nominal asset, the nominal interest sacrificed and the change in purchasing power cannot be added for the reason already noted—that the nominal interest rate is affected by the rate of price change, and so already includes an allowance for it. For a real asset, the cost of holding a dollar has two parts: the loss (or gain) in purchasing power because of rising (or falling) prices, plus the real return sacrificed on the alternative asset.
Over long periods of time, real returns on various classes of assets do tend toward equality. But at any one point in time, real returns may vary widely for different classes of assets. Moreover, what people who acquire assets expect them to yield—the ex ante yield—may differ widely from what the holders do in fact receive—the ex post yield.
Figure 1 is a striking example. It plots over more than a hundred years the observed nominal yield each year on a collection of long-term securities, the actual ex post real yield year by year, and the average real yield for the century 1875–1975. The nominal yield is fairly stable for most of the period, while the ex post real yield fluctuates all over the graph. Clearly, the assets were not purchased in anticipation of such widely variable yields. Such assets are typically held for long periods, so that the ex post yield for a holder of these assets was actually much less variable than the year-by-year calculations show. Indeed, the relatively stable nominal yield implies that the holders of the securities expected, on the average, zero inflation. And, until World War II, they were right: the price level in the United States in 1939 was roughly the same as in 1839.
Figure 1
Nominal and Real Bond Interest Rate, Annually, 1875–1989
The lines on the graph gradually change character after World War II, as the public came to recognize that inflation was more than a passing phenomenon. The nominal yield rose to incorporate that recognition, and the real yield became more stable, started rising toward the long-term average, and then overshot it in the 1980s.
Given the wide range of assets that are alternatives to holding money, it is a great simplification to speak of the cost of holding money. There is in fact a vector of costs, depending on the particular alternative considered. And that is still an oversimplification. Even for a single asset, there is a range of possible yields, both nominal and real. Uncertainty about the nominal yield of alternative assets is one reason for holding money—there is little