Investment, though the dynamic and volatile factor in the Keynesian system, is also the Keynesian stepchild. Keynesians have differed on the causal determinants of investment. Originally, Keynes determined it by the interest rate as compared with the marginal efficiency of capital, or prospect for net return. The interest rate is supposed to be determined by the money relation; we have seen that this idea is fallacious. Actually, the equilibrium net rate of return is the interest rate, the natural rate to which the bond rate conforms. Rather than changes in the interest rate causing changes in investment, as we have seen before, changes in time preference are reflected in changes in consumption-investment decisions. Changes in the interest rate and in investment are two sides of a coin, both determined by individual valuations and time preferences.

The error of calling the interest rate the cause of investment changes, and itself determined by the money relation, is also adopted by such “critics” of the Keynesian system as Pigou, who asserts that falling prices will release enough cash to lower the interest rate, stimulate investment, and thus finally restore full employment.

Modern Keynesians have tended to abandon the intricacies of the relation between interest and investment and simply declare themselves agnostic on the factors determining investment. They rest their case on an alleged determination of consumption.70

• 70Some Keynesians account for investment by the “acceleration principle” (see below). The Hansen “stagnation” thesis—that investment is determined by population growth, the rate of technological improvement, etc.—seems happily to be a thing of the past.

If Keynesians are unsure about investment, they have, until very recently, been very emphatic about consumption. Investment is a volatile, uncertain expenditure. Aggregate consumption, on the other hand, is a passive, stable “function” of immediately previous social income. Total net expenditures determining and equaling total net income in a period (gross expenditures between stages of production are unfortunately removed from discussion) consist of investment and consumption. Furthermore, consumption always behaves so that below a certain income level consumption will be higher than income, and above that level consumption will be lower. Figure 82 depicts the relations among consumption, investment, expenditure, and social income.

The relation between income and expenditure is the same as shown in Figure 78. Now we see why the Keynesians assume the expenditure curve to have a smaller slope than income. Consumption is supposed to have the identical slope as expenditures; for investment is unrelated to income, as the determinants are unknown. Hence, investment is depicted as having no functional relation to income and is represented as a constant gap between the expenditure and consumption lines.

The stability of the passive consumption function, as contrasted with the volatility of active investment, is a keystone of the Keynesian system. This assumption is replete with so many grave errors that it is necessary to take them up one at a time.

(a) How do the Keynesians justify the assumption of a stable consumption function with the shape as shown above? One route was through “budget studies”—cross-sectional studies of the relation between family income and expenditure by income groups in a given year. Budget studies such as that of the National Resources Committee in the mid-1930’s yielded similar “consumption functions” with dishoardings increasing below a certain point, and hoardings above it (i.e., income below expenditures below a certain point, and expenditures below income above it).

This is supposed to intimate that those doing the “dissaving,” i.e., the dishoarding, are poor people below the subsistence level who incur deficits by borrowing. But how long is this supposed to go on? How can there be a continuous deficit? Who would continue to lend these people the money? It is more reasonable to suppose that the dishoarders are decumulating their previously accumulated capital, i.e., that they are wealthy people whose businesses suffered losses during that year.

(b) Aside from the fact that budget studies are misinterpreted, there are graver fallacies involved. For the curve given by the budget study has no relation whatever to the Keynesian consumption function! The former, at best, gives a cross section of the relation between classes of family expenditure and income for one year; the Keynesian consumption function attempts to establish a relation between total social income and total social consumption for any given year, holding true over a hypothetical range of social incomes. At best, one entire budget curve can be summed up to yield only one point on the Keynesian consumption function. Budget studies, therefore, can in no way confirm the Keynesian assumptions.

(c) Another very popular device to confirm the consumption function reached the peak of its popularity during World War II. This was historical-statistical correlation of national income and consumption for a definite period of time, usually the 1930’s. This correlation equation was then assumed to be the “stable” consumption function. Errors in this procedure were numerous. In the first place, even assuming such a stable relation, it would only be an historical conclusion, not a theoretical law. In physics, an experimentally determined law may be assumed to be constant for other identical situations; in human action, historical situations are never the same, and therefore there are no quantitative constants! Conditions and valuations could change at any time, and the “stable” relationship altered. There is here no proof of a stable consumption function. The dismal record of forecasts (such as those of postwar unemployment) made on this assumption should not have been surprising.

Moreover, a stable relation was not even established. Income was correlated with consumption and with investment. Since consumption is a much larger magnitude than (net) investment, no wonder that its percentage deviations around the regression equation were smaller! Furthermore, income is here being correlated with 80–90 percent of itself; naturally, the “stability” is tremendous. If income were correlated with saving, of similar magnitude as investment, there would be no greater stability in the income-saving function than in the “income-investment function.”

Thirdly, the consumption function is necessarily an ex ante relation; it is supposed to tell how much consumers will decide to spend given a certain total income. Historical statistics, on the other hand, record only ex post data, which give a completely different story. For any given period of time, for example, hoarding and dishoarding cannot be recorded ex post. In fact, ex post, on double-entry accounting records, total social income is always equal to total social expenditures. Yet, in the dynamic, ex ante, sense, it is precisely the divergence between total social income and total social expenditures (hoarding or dishoarding) that plays the crucial role in the Keynesian theory. But these divergences can never be revealed, as Keynesians believe, by study of ex post data. Ex post, in fact, saving always equals investment, and social expenditure always equals social income, so that the ex post expenditure line coincides with the income line.71

(d) Actually, the whole idea of stable consumption functions has now been discredited, although many Keynesians do not fully realize this fact.72 In fact, Keynesians themselves have admitted that, in the long run, the consumption function is not stable, since total consumption rises as income rises; and that in the short run it is not stable, since it is affected by all sorts of changing factors. But if it is not stable in the short run and not stable in the long run, what kind of stability does it have? Of what use is it? We have seen that the only really important runs are the immediate and the long-run, which shows the direction in which the immediate is tending. There is no use for some sort of separate “intermediate” situation.

(e) it is instructive to turn now to the reasons that Keynes himself, in contrast to his followers, gave for assuming his stable consumption function. It is a confused exposition indeed.73 The “propensity to consume” out of given income, according to Keynes, is determined by two sets of factors, “objective” and “subjective.” It seems clear, however, that these are purely subjective decisions, so that there can be no separate objective determinants. In classifying subjective factors, Keynes makes the mistake of subsuming hoarding and investing motivations under categories of separate “causes”: precaution, foresight, improvement, etc. Actually, as we have seen, the demand for money is ultimately determined by each individual for all sorts of reasons, but all tied up with uncertainty; motives for investment are to maintain and increase future standards of living. By a sleight of hand completely unsupported by facts or argument Keynes simply assumes all these subjective factors to be given in the short run, although he admits that they will change in the long run. (If they change in the long run, how can his system yield an equilibrium position?) He simply reduces the subjective motives to current economic organization, customs, standards of living, etc., and assumes them to be given.74 The “objective factors” (which in reality are subjective, such as time-preference changes, expectations, etc.) can admittedly cause short-run changes in the consumption function (such as windfall changes in capital values). Expectations of future changes in income can affect an individual’s consumption, but Keynes simply asserts without discussion that this factor “is likely to average out for the community as a whole.” Time preferences are discussed in a very confused way, with interest rate and time preference assumed to be apart from and influencing the propensity to consume. Here again, short-run fluctuations are assumed to have little effect, and Keynes simply leaps to the conclusion that the propensity to consume is, in the short run, a “fairly” stable function.75

(f) The failure of the consumption-function theory is not only the failure of a specific theory. It is a profound epistemological failure as well. For the concept of a consumption function has no place in economics at all. Economics is praxeological, i.e., its propositions are absolutely true given the existence of the axioms—the basic axiom being the existence of human action itself. Economics, therefore, is not and cannot be “empirical” in the positivist sense, i.e., it cannot establish some sort of empirical hypothesis which could or could not be true, and at best is only true approximately. Quantitative, empiricohistorical “laws” are worthless in economics, since they may only be coincidences of complex facts, and not isolable, repeat-able laws which will hold true in the future. The idea of the consumption function is not only wrong on many counts; it is irrelevant to economics.

Furthermore, the very term “function” is inappropriate in a study of human action. Function implies a quantitative, determined relationship, whereas no such quantitative determinism exists. People act and can change their actions at any time; no causal, constant, external determinants of action can exist. The term “function” is appropriate only to the unmotivated, repeat-able motion of inorganic matter.

In conclusion, there is no reason whatever to assume that at some point, expenditures will be below income, while at lower points it will be above income. Economics does not and cannot know what ex ante expenditure will ever be in relation to income; at any point, it could be equal, or there could be net hoarding or dishoarding. The ultimate decisions are made by the individuals and are not determinable by science. There is, therefore, no stable expenditure function whatever.

• 71See Lindahl, “On Keynes’ Economic System—Part I,” p. 169 n. Lindahl shows the difficulties of mixing an ex post income line with ex ante consumption and spending, as the Keynesians do. Lindahl also shows that the expenditure and income lines coincide if the divergence between expected and realized income affects income and not stocks. Yet it cannot affect stocks, for, contrary to Keynesian assertion, there is no such thing as hoarding or any other unexpected event leading to “unintended increase in inventories.” An increase in inventories is never unintended, since the seller has the alternative of selling the good at the market price. The fact that his inventory increases means that he has voluntarily invested in larger inventory, hoping for a future price rise.
• 72Summing up disillusionment with the consumption function are two significant articles: Murray E. Polakoff, “Some Critical Observations on the Major Keynesian Building Blocks,” Southern Economic Journal, October, 1954, pp. 141–51; and Leo Fishman, “Consumer Expectations and the Consumption Function,” ibid., January, 1954, pp. 243–51.
• 73Keynes, General Theory, pp. 89–112.
• 74Ibid., pp. 109–10.
• 75What is “fairly” supposed to mean? How can a theoretical law be based on “fair” stability? More stable than other functions? What are the grounds for this assumption, particularly as a law of human action? Ibid., pp. 89–96.

The once highly esteemed “multiplier” has now happily faded in popularity, as economists have begun to realize that it is simply the obverse of the stable consumption function. However, the complete absurdity of the multiplier has not yet been fully appreciated. The theory of the “investment multiplier” runs somewhat as follows:

Social Income = Consumption + Investment

Consumption is a stable function of income, as revealed by statistical correlation, etc. Let us say, for the sake of simplicity, that Consumption will always be .80 (Income).76 In that case,

Income = .80 (Income) + Investment.
.20 (Income) = Investment; or
Income = 5 (Investment).

The “5” is the “investment multiplier.” It is then obvious that all we need to increase social money income by a desired amount is to increase investment by 1/5 of that amount; and the multiplier magic will do the rest. The early “pump primers” believed in approaching this goal through stimulating private investment; later Keynesians realized that if investment is an “active” volatile factor, government spending is no less active and more certain, so that government spending must be relied upon to provide the needed multiplier effect. Creating new money would be most effective, since the government would then be sure not to reduce private funds. Hence the basis for calling all government spending “investment”: it is “investment” because it is not tied passively to income.

The following is offered as a far more potent “multiplier,” on Keynesian grounds even more potent and effective than the investment multiplier, and on Keynesian grounds there can be no objection to it. It is a reductio ad absurdum, but it is not simply a parody, for it is in keeping with the Keynesian method.

Social Income = Income of (insert name of any person, say the reader) + Income of everyone else.

Let us use symbols:

Social income = Y
Income of the Reader = R
Income of everyone else = V

We find that V is a completely stable function of Y. Plot the two on coordinates, and we find historical one-to-one correspondence between them. It is a tremendously stable function, far more stable than the “consumption function.” On the other hand, plot R against Y. Here we find, instead of perfect correlation, only the remotest of connections between the fluctuating income of the reader of these lines and the social income. Therefore, this reader’s income is the active, volatile, uncertain element in the social income, while everyone else’s income is passive, stable, determined by the social income.

Let us say the equation arrived at is:

This is the reader’s own personal multiplier, a far more powerful one than the investment multiplier. To increase social income and thereby cure depression and unemployment, it is only necessary for the government to print a certain number of dollars and give them to the reader of these lines. The reader’s spending will prime the pump of a 100,000-fold increase in the national income.77

• 76Actually, the form of the Keynesian function is generally “linear,” e.g., Consumption = .80 (Income) + 20. The form given in the text simplifies the exposition without, however, changing its essence.
• 77Also see Hazlitt, Failure of the “New Economics,” pp. 135–55.