It is true, as we have said, that the only use for money is in exchange. From this, however, it must not be inferred, as some writers have done, that this exchange must be immediate. Indeed, the reason that a reservation demand for money exists and cash balances are kept is that the individual is keeping his money in reserve for future exchanges. That is the function of a cash balance—to wait for a propitious time to make an exchange.

Suppose the ERE has been established. In such a world of certainty, there would be no risk of loss in investment and no need to keep cash balances on hand in case an emergency for consumer spending should arise. Everyone would therefore allocate his money stock fully, to the purchase of either present goods or future goods, in accordance with his time preferences. No one would keep his money idle in a cash balance. Knowing that he will want to spend a certain amount of money on consumption in six months’ time, a man will lend his money out for that period to be returned at precisely the time it is to be spent. But if no one is willing to keep a cash balance longer than instantaneously, there will be no money held and no use for a money stock. Money, in short, would either be useless or very nearly so in the world of certainty.

In the real world of uncertainty, as contrasted to the ERE, even “idle” money kept in a cash balance performs a use for its owner. Indeed, if it did not perform such a use, it would not be kept in his cash balance. Its uses are based precisely on the fact that the individual is not certain on what he will spend his money or of the precise time that he will spend it in the future.

Economists have attempted mechanically to reduce the demand for money to various sources.7 There is no such mechanical determination, however. Each individual decides for himself by his own standards his whole demand for cash balances, and we can only trace various influences which different catallactic events may have had on demand.

• 7J.M. Keynes’ Treatise on Money (New York: Harcourt, Brace, 1930) is a classic example of this type of analysis.

One of the most obvious influences on the demand for money is expectation of future changes in the exchange-value of money. Thus, suppose that, at a certain point in the future, the PPM of money is expected to drop rapidly. How the demand-for-money schedule now reacts depends on the number of people who hold this expectation and the strength with which they hold it. It also depends on the distance in the future at which the change is expected to take place. The further away in time any economic event, the more its impact will be discounted in the present by the interest rate. Whatever the degree of impact, however, an expected future fall in the PPM will tend to lower the PPM now. For an expected fall in the PPM means that present units of money are worth more than they will be in the future, in which case there will be a fall in the demand-for-money schedule as people tend to spend more money now than at the future date. A general expectation of an imminent fall in the PPM will lower the demand schedule for money now and thus tend to bring about the fall at the present moment.

Conversely, an expectation of a rise in the PPM in the near future will tend to raise the demand-for-money schedule as people decide to “hoard” (add money to their cash balance) in expectation of a future rise in the exchange-value of a unit of their money. The result will be a present rise in the PPM.

An expected fall in the PPM in the future will therefore lower the PPM now, and an expected rise will lead to a rise now. The speculative demand for money functions in the same manner as the speculative demand for any good. An anticipation of a future point speeds the adjustment of the economy toward that future point. Just as the speculative demand for a good speeded adjustment to an equilibrium position, so the anticipation of a change in the PPM speeds the market adjustment toward that position. Just as in the case of any good, furthermore, errors in this speculative anticipation are “self-correcting.” Many writers believe that in the case of money there is no such self-correction. They assert that while there may be a “real” or underlying demand for goods, money is not consumed and therefore has no such underlying demand. The PPM and the demand for money, they declare, can be explained only as a perpetual and rather meaningless cat-and-mouse race in which everyone is simply trying to anticipate everyone else’s anticipations.

There is, however, a “real” or underlying demand for money. Money may not be physically consumed, but it is used, and therefore it has utility in a cash balance. Such utility amounts to more than speculation on a rise in the PPM. This is demonstrated by the fact that people do hold cash even when they anticipate a fall in the PPM. Such holdings may be reduced, but they still exist, and as we have seen, this must be so in an uncertain world. In fact, without willingness to hold cash, there could be no monetary-exchange economy whatever.

The speculative demand therefore anticipates the underlying nonspeculative demands, whatever their source or inspiration. Suppose, then, that there is a general anticipation of a rise in the PPM (a fall in prices) not reflected in underlying supply and demand. It is true that, at first, this general anticipation raises, ceteris paribus, the demand for money and the PPM. But this situation does not last. For now that a pseudo “equilibrium” has been reached, the speculative anticipators, who did not “really” have an increased demand for money, sell their money (buy goods) to reap their gains. But this means that the underlying demand comes to the fore, and this is less than the money stock at that PPM. The pressure of spending then lowers the PPM again to the true equilibrium point. This may be diagramed as in Figure 77.

Money stock is 0S; the true or underlying money demand is DD, with true equilibrium point at A. Now suppose that the people on the market erroneously anticipate that true demand will be such in the near future that the PPM will be raised to 0E. The total demand curve for money then shifts to DsDs, the new total demand curve including the speculative demand. The PPM does shift to 0E as predicted. But now the speculators move to cash in their gain, since their true demand for money really reflects DD rather than DsDs. At the new price 0E, there is in fact an excess of money stock over quantity demanded, amounting to CF. Sellers rush to sell their stock of money and buy goods, and the PPM falls again to equilibrium. Hence, in the field of money as well as in that of specific goods, speculative anticipations are self-correcting, not “self-fulfilling.” They speed the market process of adjustment.

Long-run influences on the demand for money in a progressing economy will tend to be manifold, and in both directions. On the one hand, an advancing economy provides ever more occasions for new exchanges as more and more commodities are offered on the market and as the number of stages of production increases. These greater opportunities tend greatly to increase the demand-for-money schedule. If an economy deteriorates, fewer opportunities for exchange exist, and the demand for money from this source will fall.

The major long-run factor counteracting this tendency and tending toward a fall in the demand for money is the growth of the clearing system.8 Clearing is a device by which money is economized and performs the function of a medium of exchange without being physically present in the exchange.

A simplified form of clearing may occur between two people. For example, A may buy a watch from B for three gold ounces; at the same time, B buys a pair of shoes from A for one gold ounce. Instead of two transfers of money being made, and a total of four gold ounces changing hands, they decide to perform a clearing operation. A pays B two ounces of money, and they exchange the watch and the shoes. Thus, when a clearing is made, and only the net amount of money is actually transferred, all parties can engage in the same transactions at the same prices, but using far less cash. Their demand for cash tends to fall.

There is obviously little scope for clearing, however, as long as all transactions are cash transactions. For then people have to exchange one another’s goods at the same time. But the scope for clearing is vastly increased when credit transactions come into play. These credits may be quite short-term. Thus, suppose that A and B deal with each other quite frequently during a year or a month. Suppose they agree not to pay each other immediately in cash, but to give each other credit until the end of each month. Then B may buy shoes from A on one day, and A may buy a watch from B on another. At the end of the period, the debts are canceled and cleared, and the net debtor pays one lump sum to the net creditor.

Once credit enters the picture, the clearing system can be extended to as many individuals as find it convenient. The more people engage in clearing operations (often in places called “clearinghouses”) the more cancellations there will be, and the more money will be economized. At the end of the week, for example, there may be five people engaged in clearing, and A may owe B ten ounces, B owe C ten ounces, C owe D, etc., and finally E may owe A ten ounces. In such a case, 50 ounces’ worth of debt transactions and potential cash transactions are settled without a single ounce of cash being used.

Clearing, then, is a process of reciprocal cancellations of money debts. It permits a huge quantity of monetary exchanges without actual possession and transfer of money, thereby greatly reducing the demand for money. Clearing, however, cannot be all-encompassing, for there must be some physical money which could be used to settle the transaction, and there must be physical money to settle when there is no 100-percent cancellation (which rarely occurs).

• 8On the clearing system, see Mises, Theory of Money and Credit, pp. 281–86.

A popular fallacy rejects the concept of “demand for money” because it is allegedly always unlimited. This idea misconceives the very nature of demand and confuses money with wealth or income. It is based on the notion that “people want as much money as they can get.” In the first place, this is true for all goods. People would like to have far more goods than they can procure now. But demand on the market does not refer to all possible entries on people’s value scales; it refers to effective demand, to desires made effective by being “demanded,” i.e., by the fact that something else is “supplied” for it. Or else it is reservation demand, which takes the form of holding back the good from being sold. Clearly, effective demand for money is not and cannot be unlimited; it is limited by the appraised value of the goods a person can sell in exchange and by the amount of that money which the individual wants to spend on goods rather than keep in his cash balance.

Furthermore, it is, of course, not “money” per se that he wants and demands, but money for its purchasing power, or “real” money, money in some way expressed in terms of what it will purchase. (This purchasing power of money, as we shall see below, cannot be measured.) More money does him no good if its purchasing power for goods is correspondingly diluted.

We have been discussing money, and shall continue to do so in the current section, by comparing equilibrium positions, and not yet by tracing step by step how the change from one position to another comes about. We shall soon see that in the case of the price of money, as contrasted with all other prices, the very path toward equilibrium necessarily introduces changes that will change the equilibrium point. This will have important theoretical consequences. We may still talk, however, as if money is “neutral,” i.e., does not lead to such changes, because this assumption is perfectly competent to deal with the problems analyzed so far. This is true, in essence, because we are able to use a general concept of the “purchasing power of money” without trying to define it concretely in terms of specific arrays of goods. Since the concept of the PPM is relevant and important even though its specific content changes and cannot be measured, we are justified in assuming that money is neutral as long as we do not need a more precise concept of the PPM.

We have seen how changes in the money relation change the PPM. In the determination of the interest rate, we must now modify our earlier discussion in chapter 6 to take account of allocating one’s money stock by adding to or subtracting from one’s cash balance. A man may allocate his money to consumption, investment, or addition to his cash balance. His time preferences govern the proportion which an individual devotes to present and to future goods, i.e., to consumption and to investment. Now suppose a man’s demand-for-money schedule increases, and he therefore decides to allocate a proportion of his money income to increasing his cash balance. There is no reason to suppose that this increase affects the consumption/investment proportion at all. It could, but if so, it would mean a change in his time preference schedule as well as in his demand for money.

If the demand for money increases, there is no reason why a change in the demand for money should affect the interest rate one iota. There is no necessity at all for an increase in the demand for money to raise the interest rate, or a decline to lower it—no more than the opposite. In fact, there is no causal connection between the two; one is determined by the valuations for money, and the other by valuations for time preference.

Let us return to the section in chapter 6 on Time Preference and the Individual’s Money Stock. Did we not see there that an increase in an individual’s money stock lowers the effective time-preference rate along the time-preference schedule, and conversely that a decrease raises the time-preference rate? Why does this not apply here? Simply because we were dealing with each individual’s money stock and assuming that the “real” exchange-value of each unit of money remained the same. His time-preference schedule relates to “real” monetary units, not simply to money itself. If the social stock of money changes or if the demand for money changes, the objective exchange-value of a monetary unit (the PPM) will change also. If the PPM falls, then more money in the hands of an individual may not necessarily lower the time-preference rate on his schedule, for the more money may only just compensate him for the fall in the PPM, and his “real money stock” may therefore be the same as before. This again demonstrates that the money relation is neutral to time preference and the pure rate of interest.

An increased demand for money, then, tends to lower prices all around without changing time preference or the pure rate of interest Thus, suppose total social income is 100, with 70 allocated to investment and 30 to consumption. The demand for money increases, so that people decide to hoard a total of 20. Expenditure will now be 80 instead of 100, 20 being added to cash balances. Income in the next period will be only 80, since expenditures in one period result in the identical income to be allocated to the next period.9 If time preferences remain the same, then the proportion of investment to consumption in the society will remain roughly the same, i.e., 56 invested and 24 consumed. Prices and nominal money values and incomes fall all along the line, and we are left with the same capital structure, the same real income, the same interest rate, etc. The only things that have changed are nominal prices, which have fallen, and the proportion of total cash balances to money income, which has increased.

A decreased demand for money will have the reverse effect. Dishoarding will raise expenditure, raise prices, and, ceteris paribus, maintain the real income and capital structure intact. The only other change is a lower proportion of cash balances to money income.

The only necessary result, then, of a change in the demand-for-money schedule is precisely a change in the same direction of the proportion of total cash balances to total money income and in the real value of cash balances. Given the stock of money, an increased scramble for cash will simply lower money incomes until the desired increase in real cash balances has been attained.

If the demand for money falls, the reverse movement occurs. The desire to reduce cash balances causes an increase in money income. Total cash remains the same, but its proportion to incomes, as well as its real value, declines.10

• 9Since no one can receive a money income unless someone else makes a money expenditure on his services. (See chapter 3 above.)
• 10Strictly, the ceteris paribus condition will tend to be violated. An increased demand for money tends to lower money prices and will therefore lower money costs of gold mining. This will stimulate gold mining production until the interest return on mining is again the same as in other industries. Thus, the increased demand for money will also call forth new money to meet the demand. A decreased demand for money will raise money costs of gold mining and at least lower the rate of new production. It will not actually decrease the total money stock unless the new production rate falls below the wear-and-tear rate. Cf. Jacques Rueff, “The Fallacies of Lord Keynes’ General Theory” in Henry Hazlitt, ed., The Critics of Keynesian Economics (Princeton, N.J.: D. Van Nostrand, 1960), pp. 238–63.

Many economists, beginning with Irving Fisher, have asserted that the market rate of interest, in addition to containing specific entrepreneurial components superimposed on the pure rate of interest, also contains a “price” or a “purchasing-power component.” When the purchasing power of money is generally expected to rise, the theory asserts that the market rate of interest falls correspondingly; when the PPM is expected to fall, the theory declares that the market rate of interest rises correspondingly.

These economists erred by concentrating on the loan rate rather than on the natural rate (the rate of return). The reasoning behind this theory was as follows: If the purchasing power of money is expected to change, then the pure rate of interest (determined by time preference) will no longer be the same in “real terms.” Suppose that 100 gold ounces exchange for 105 gold ounces a year from now—i.e., that the rate of interest is 5 percent. Now, suddenly, let there be a general expectation that the purchasing power of money will increase. In that case, a lower amount to be returned, say 102 ounces, may yield 5 percent real interest in terms of purchasing power. A general expectation of a rise in purchasing power, therefore, will lower the market rate of interest at present, while a general expectation of a fall in purchasing power will raise the rate.28

There is a fatal defect in this generally accepted line of reasoning. Suppose, for example, that prices are generally expected to fall by 50 percent in the next year. Would someone lend 100 gold ounces to exchange for 53 ounces one year from now? Why not? This would certainly preserve the real interest rate at 5 percent. But then why should the would-be lenders not simply hold on to their money and double their real assets as a result of the price fall? And that is precisely what they would do; they certainly would not give money away, even though their real assets would be greater than before. Fisher simply shrugged off this point by saying that the purchasing-power premium could never make the interest rate negative. But this flaw vitiates the entire theory.

The root of the difficulty consists in ignoring the natural rate of interest. Let us consider the interest rate in those terms. Then, suppose 100 ounces are paid for factors that will be transformed in one year into a product that sells for 105 gold ounces, for an interest gain of five and an interest return of 5 percent. Now a general expectation arises of a general halving of prices one year from now. The selling price of the product will be 53 ounces in a year’s time. What happens now? Will entrepreneurs buy factors for 100 and sell at 53 merely because their real interest rate is preserved? Certainly not. They will do so only if they do not at all anticipate the change in purchasing power. But to the extent that it is anticipated, they will hold money rather than buy factors. This will immediately lower factor prices to their expected future levels, say from 100 to 50.

What happens to the loan rate is analytically quite trivial. It is simply a reflection of the natural rate and depends on how the expectations and judgment of the people on the loan market compare with those on the stock and other markets. For the free economy, there is no point in separately analyzing the loan market. Analysis of the Fisher problem—the relation of the interest rate to price changes—should concentrate on the natural rate of interest. Discussion of the relation between price movements and the (natural) rate of interest should be divided into two parts: first, assuming “neutral money”—that all prices change equally and at the same time—and second, analyzing conditions where factor and product change at different rates. And these changes should first be analyzed without considering that they had been anticipated by people on the market.

Suppose, first, that all prices change equally and at the same time. Instead of thinking in terms of 100 ounces borrowed on the loan market, let us consider the natural rate. An investor buys factors in period 1 and then sells the product, say in period 3. Time, as we have seen, is the essence of the production structure. All the processes take time, and capitalists pay money to owners of factors in advance of production and sale. Since factors are bought before products are sold, what is the effect of a period of rising general prices (i.e., falling PPM)? The result is that the entrepreneur reaps an apparent extra profit. Suppose that he normally purchases original factors for 100 and then sells the product for 120 ounces two years later, for an interest return of 10 percent per annum. Now suppose that a decrease in the demand for money or an increase of money stock propels a general upward movement in prices and that all prices double in two years’ time. Then, just because of the passage of time, an entrepreneur who purchases factors for 100 now will sell for 240 ounces in two years’ time. Instead of a net return of 20 ounces, or 10 percent per annum, he reaps 140 ounces, or 70 percent per annum.

It would seem that a rise in prices creates an inherent tendency for large-scale profits that are not simply individual rewards for more accurate forecasting. However, more careful analysis reveals that this is not an extra profit at all. For the 240 ounces two years from now is roughly equivalent, in terms of purchasing power, to 120 ounces now. The real rate of net return, based on money’s services, is the same 10 percent as it has always been. It is clear that any lower net return would amount to a decline in real return. A return of a mere 120 ounces, for example, would amount to a drastic negative real return, for 100 ounces would then be invested for the equivalent gross return of only 60 ounces. It has often been shown that a period of rising prices misleads businessmen into thinking that their increased money profits are also real gains, whereas they only maintain real rates of return. Consider, for example, “replacement costs”—the prices which the businessmen will now have to pay for factors. The capitalist who earns 240 ounces on a 100-ounce investment neglects to his sorrow the fact that his factor bundle now costs 200 ounces instead of 100. Businessmen who under such circumstances treat their monetary profits as real profits and consume them soon find that they are really consuming their capital.

The converse occurs in the case of falling prices. The capitalist buys factors in period 1 and sells the product in period 3, when all-around prices are lower. If prices are to fall by a half in two years, an investment of 100, followed by a sale at 60, does not really involve the terrific loss that it appears to be. For the 60 return is equivalent in real terms, both in generalized purchasing power and in replacement of factors, to 120 previous ounces. His real rate of return remains the same. The consequence is that businessmen will be likely to overstate their losses in a period of price contraction. Perhaps this is one of the major reasons for the deep-seated belief of most businessmen that they always gain during a general price expansion and lose during a period of general contraction. This belief is purely illusory.

In these examples, the natural interest rate on the market has contained a purchasing-power component, which corrects for real rates, positively in money terms during a general expansion, and negatively during a general contraction. The loan rate will be simply a reflection of what has been happening in the natural rate. So far, the discussion is similar to Fisher’s, except that these are the effects of actual, not anticipated, changes and the Fisher thesis cannot take account of the negative interest rate case. We have seen that rather than take a monetary loss, even though their real return will be the same, entrepreneurs will hold back their purchases of factors until factor prices fall immediately to their future low level. But this process of anticipatory price movement does not occur only in the extreme case of a prospective “negative” return. It happens whenever a price change is anticipated. Thus, suppose all entrepreneurs generally anticipate that prices will double in two years. The fact of an anticipated rise will lead to an increase in the price level now and an approach immediately toward a doubled price level. An anticipated fall will lead to an immediate fall in factor prices. If all changes were anticipated by everyone, there would be no room for a purchasing-power component to develop. Prices would simply fall immediately to their future level.

The purchasing-power component, then, is not the reflection, as has been thought, of expectations of changes in purchasing power. It is the reflection of the change itself; indeed, if the change were completely anticipated, the purchasing power would change immediately, and there would be no room for a purchasing-power component in the rate of interest. As it is, partial anticipations speed up the adjustment of the PPM to the changed conditions.

So far we have distinguished three components of the natural rate of interest (all reflected in the loan rate of interest). One is the pure rate of interest—the result of individual time preferences, tending to be uniform throughout the economy. Second are the specific entrepreneurial rates of interest. These differ from firm to firm and so are not uniform. They are anticipated in advance, and they are the rates that an investor will have to anticipate receiving before he enters the field. A particularly “risky” venture, if successful at all, will therefore tend to earn more in net return than what is generally anticipated to be a “safe” venture. The third component of the natural rate of interest is the purchasing-power component, correcting for general PPM changes because of the inevitable time lags in production. This will be positive in an expansion and negative in a contraction, but will be ephemeral. The more that changes in the PPM are anticipated, the less important will be the purchasing-power component and the more rapid will be the adjustment in the PPM itself.

There is still a fourth component in the natural rate of interest. This exists to the extent that money changes are not neutral (and they never are). Sometimes product prices rise and fall faster than factor prices, sometimes they rise and fall more slowly, and sometimes their behavior is mixed, with some factor prices and some product prices rising more rapidly. Whenever there is a general divergence in rates of movement between the prices of the product and of original factors, a terms-of-trade component emerges in the natural rate of interest.

Historically, it has often been the case that product prices rise more rapidly and fall more rapidly than the prices of original factors. In the former case, there is, during the period of transition, a favorable change in the terms of trade to the general run of capitalists. For selling prices are increasing faster than the buying prices of original factors. This will increase the general rate of return and constitute a general positive terms-of-trade component in the natural rate of interest. This, of course, will also tend to be reflected in the loan rate of interest. In the case of a contraction, a more sluggish fall in the prices of factors creates a general negative terms-of-trade component in the interest rate. The components are precisely the reverse whenever factor prices change more rapidly than product prices. Whenever there is no general change in the “terms of trade” to capitalist-entrepreneurs, no terms-of-trade component will appear in the interest rate.

Changes in terms of trade discussed here are only those resulting purely from differences in the speed of reaction to changing conditions. They do not include basic changes in the terms of trade resulting from changes in time preferences, such as we have discussed above.

It is clear that all the interest-rate components aside from the pure rate—entrepreneurial, purchasing power, and terms of trade—are “dynamic” and the result of uncertainty. None of these components would exist in the ERE, and therefore the market interest rate in the ERE would equal the pure rate determined by time preferences alone. In the ERE the only net incomes would be a uniform pure interest return and wages to labor (ground land rents being capitalized into an interest return).

• 28Irving Fisher, The Rate of Interest (New York, 1907), chap. v, xiv; idem, Purchasing Power of Money, pp. 56–59.