Chapter XVIII THE GENERAL THEORY RESTATED

1. Economic Interrelationships

Keynes’s Chapter 18 is called “The General Theory of Employment Restated.” The “restatement” turns out to be confusion worse confounded.

On the assumption that “we have now reached a point where we can gather together the threads of our argument,” Keynes thinks “it may be useful to make clear which elements in the economic system we usually take as given, which are the independent variables of our system and which are the dependent variables” (p. 245).

Now economics is concerned with human valuations, human decisions, and human action. Everything in the system is a variable. No relationship (unless it is merely two ways of saying the same thing) is a constant. Nothing is permanently “given.” Almost anything can be an “independent” variable, in the sense that a change can originate at that point. When a change has originated at any point, then the relationship of nearly all the factors is one of mutual dependence, of interdependence.

We take as given [Keynes continues] the existing skill and quantity of available labor, the existing quality and quantity of available equipment, the existing technique, the degree of competition, the tastes and habits of the consumer... the social structure including the forces... which determine the distribution of the national income. This does not mean that we assume these factors to be constant; but merely that, in this place and context, we are not considering or takinginto account the effects and consequences of changes in them. (Italics supplied, p. 245.)

David McCord Wright contends that this is actually the first point in the General Theory where Keynes states “the basic assumptions of his fundamental model”; and he uses the foregoing italics to stress the point that “in the basic model” on which Keynes’s system rests, “virtually all the dynamic social forces are omitted.”¹

Frank H. Knight, after quoting the same passage, as well as a passage on the following two pages (246-247), in which Keynes declares: “Thus we can sometimes regard our ultimate independent variables as consisting of”... etc., follows his quotations by a sweeping comment on the whole Keynesian system:

It would surely appear that if one is willing to make assumptions of this sort—along with those already pointed out, namely, that there is unemployment, that wages and prices cannot fall (but are free to rise), that wages are uninfluenced by the supply-offering of labor, that the price of capital-service is dependent only on the speculative attitude of the public toward money (i.e., toward general prices) and the quantity of money fixed by the arbitrary fiat of a central banking authority entirely uninfluenced either by saving or by the demand for capital—one should indeed find little difficulty in revolutionizing economic theory in any manner or degree or in rationalizing any policy which one might find appealing.²

On the same page, Keynes continues: “The division of the determinants of the economic system into the two groups of given factors and independent variables is, of course, quite arbitrary from any absolute standpoint” (p. 247). This is entirely true; and if Keynes had recognized this clearly and consistently, the whole General Theory might not have been written. What is “given,” what is an “independent variable,” and what is a “dependent variable,” depends entirely on the problem with which we are dealing.

Economic analysis continually involves the setting up and testing of hypotheses. It asks, for example, if a and b are given, what will be the value of c, or if a and c change, what will be the effect on b, etc.

The basic illustration is, of course, the relationship of supply, demand, and price. If “supply” is used in the sense of quantity supplied, and “demand” in the sense of quantity demanded, then a change originating in any one of these three factors will change another. In other words, if any two of these three factors are, by hypothesis or by assumption, the “independent variables,” then the other becomes, for the purpose of solving the particular problem under consideration, the “dependent variable.”

If supply is used in the sense of supply “schedule” or “curve,” and demand in the sense of demand “schedule” or “curve,” then orthodox economic analysis would say that a change in either one does not necessarily change the other, though a change in either would change price; and that under conditions of perfect competition price could not change independently, but only as a consequence of a change in the supply curve, or the demand curve, or both. This, it may be pointed out, is merely a consequence of the meaning of our terms. The full name for the “demand curve,” for example, is the curve of price-and-quantity-demanded.

In any case, it is characteristic of economic problem-solving that what is “given” is determined by the nature of the problem. Conclusions regarding what is dependent and what is independent, what is cause and what effect, are determined by our arbitrarily selected starting point.

In commenting upon Keynes’s Chapter 18, therefore, I shall not make again any detailed analysis of the factors that Keynes regards as “independent variables” and “dependent variables” respectively, what he regards as cause and what effect. It is enough merely to make the general point that his analysis is arbitrary and implausible, and sometimes clearly reverses cause and effect.

A few comments upon some particular sentences or passages, however, seem called for.

Within the economic framework which we take as given, the national income depends on the volume of employment, i.e. on the quantity of effort currently devoted to production, in the sense that there is a unique correlation between the two (p. 246). Our present object is to discover what determines at any time the national income of a given economic system and (which is almost the same thing) the amount of its employment (p. 247).

The national income is certainly not the same thing as the amount of employment. Nor is there a “unique correlation” between them. The United States with heavy unemployment would have an immensely higher income, either total or per capita, than India or China with full employment. And even within the same nation, say the United States, employment and income do not necessarily rise and fall proportionately. As employment gets fuller, production per man employed tends to fall. As unemployment rises, production per man employed tends to rise. This is partly because, when unemployment sets in, it is the least efficient workers that tend to be dropped first, and when employment rises, it is the less efficient (than those already employed) that must be hired. Moreover, when employment is assured, and other jobs are easy to obtain, there tends to be relaxation of effort on the part of workers, whereas when jobs are insecure, there is an increase of individual effort.

Again, either insistence on excessive wage-rates, or new inventions and improvements, may force the substitution of machinery for workers. In one case there may be a temporary fall in employment without any corresponding fall in production (or total income). In the other case there may be no net change in employment but a significant rise in production (and real income). The “volume of employment” does not necessarily mean “the quantity of effort currently devoted to production.” Part of “the effort devoted to production” consists in capital improvement, better management, a better balance of production, etc. “Full employment” can conceal gross inefficiencies in production, malinvestment, unbalanced output of consumer goods, and laxity. All of which Keynes consistently ignores.

“Changes in the rate of consumption are, in general, in the same direction (though smaller in amount) as changes in the rate of income.” (Keynes’s italics, p. 248.) In other words, when a man’s income rises, he consumes more; the more his income rises, the more he tends to consume; and when a man’s income falls, he consumes less! Tremendous discovery, which deserves all the italics that Keynes can give it.

2. “Stable” Unemployment

Keynes’s reasoning leads to the logical conclusion that there must be violent fluctuations in prices and employment. But these violent fluctuations do not, in fact, seem to occur. Instead of concluding, however, that there must be something wrong in his own analysis, Keynes concludes that there must be something illogical about economic realities. He develops a theory of mysterious stabilizing forces.

In particular, it is an outstanding characteristic of the economic system in which we live that, whilst it is subject to severe fluctuations in respect of output and employment, it is not violently unstable. Indeed it seems capable of remaining in a chronic condition of sub-normal activity for a considerable period without any marked tendency either towards recovery or towards complete collapse. Moreover, the evidence indicates that full, or even approximately full, employment is of rare and short-lived occurrence (pp. 249-250).

This is a sweeping generalization from a comparatively short and special experience. The condition of comparatively “stabilized unemployment” existed in the United States from about 1931 to 1939. It began sooner in Britain, from about 1925. And in both cases the reason was the same. The British pound sterling, off gold, had fallen from a parity of $4.86 to a low of $3.18 in February of 1920; it had recovered strongly and in late 1924 and early 1925 stood at approximately 10 per cent below the gold parity. Prices and wages had adjusted themselves upward, however, to a lower value for the pound. In April of 1925 Britain decided to return to a gold standard at the old parity of $4.86. This decision would not have been disastrous if British business and labor had recognized its implications, which was that wage-rates and prices would have to readjust downward again to compensate for the domestic and international rise in the value of the pound. But organized labor in Britain remained adamant against accepting any cut in wage-rates. It was precisely because organized labor in Britain followed the very course during and after 1925 that Keynes applauds in the General Theory that it brought about the “stable unemployment” that he deplores and regards as a permanent attribute of “the economic system in which we live.”

The same thing is true in the United States. Prolonged mass unemployment was specifically a phenomenon of the 1930’s. As a result of the inflation of World War I, wholesale prices in May of 1920 had reached a peak at 248 per cent of the 1913 level. Then came the most violent price break on record for such a period. By August of the following year, 1921, the index of wholesale prices had dropped to 141. This resulted, temporarily, in heavy unemployment. But wage-rates were fortunately still flexible. As compared with wholesale prices, their decline was indeed comparatively small. If we compare average wholesale prices with average hourly wages in 1920 and 1922, we find that whereas prices fell an average of 38 per cent between 1920 and 1922 hourly wages fell an average of only 11 per cent. But this was enough to permit readjustment. By the spring of 1923 the United States had reached new high levels in industrial production and there were labor shortages in many lines.³

In brief, the “stabilized” unemployment in the United States in the thirties, and in Britain in the late twenties and the thirties, was not a permanent characteristic of “the economic system in which we live.” It was a temporarily frozen situation due to the very wage-inflexibility-down-wards that Keynes advocates. It was not the result of laissez faire, but the result of labor-union policy supported by government policy. And it was not an “unemployment equilibrium,” which is a contradiction in terms, but an unemployment frozen by policy, by a refusal to adjust.

3. The Demand for Labor is Elastic

“When there is a change in employment, money-wages tend to change in the same direction as, but not in great disproportion to, the change in employment; i.e. moderate changes in employment are not associated with very great changes in money-wages” (p. 251).

This is a typical instance of Keynes’s reversal of typical or normal cause and effect. The significant thing, in most situations, is the effect of changes in wage-rates on employment. Looked at from this side, employment tends, of course, to change in the opposite direction from wage-rates. If there has been prolonged mass unemployment, as a result of labor-union insistence on excessive hourly wage-rates (in relation to prices and marginal labor productivity), then a fall of these wage-rates toward the equilibrium point will mean a rise in employment. If, of course, it is prices rather than wage-rates that have been above the equilibrium level, or if for some reason wage-rates have temporarily fallen below the equilibrium level, then an increase in the demand for goods due to a fall in prices, or some other change, or an increase in the demand for labor due to the low wage-rate, will mean an increase in both employment and wage-rates. In this special case the relationship stated above by Keynes would hold. But this is a comparatively rare and short-lived situation. Much more frequently, it is a downward adjustment in wage-rates (or a gradual rise in man-machine-hour productivity) that will bring a rise in employment.

What will happen, in short, depends upon the initial situation from which we start; upon the assumptions we make regarding the previous state of disequilibrium. But Keynes almost never explicitly states his initial assumptions. He persistently treats abnormal situations as normal ones, or hopelessly confuses everything by calling a state of disequilibrium a state of equilibrium.

Keynes is correct, though not for the reasons he gives, in declaring that “moderate changes in employment are not associated with very great changes in money-wages” (p. 251). A much more enlightening way to state this is to say that moderate changes in wage-rates can bring about much larger changes in employment. Paul Douglas, as a result of elaborate statistical studies, came to the conclusion that the demand for labor is highly elastic—that a 1 per cent decline in wages can mean a 3 or 4 per cent increase in employment, when wages have been held above the point of marginal productivity.⁴ (This could mean, conversely, that a rise of 1 per cent in wage-rates, under similar conditions, could mean a 3 or 4 per cent decrease in employment.) A. C. Pigou independently came to a similar conclusion.⁵

(I do not personally believe that it is possible to measure, either by statistics or mathematical deduction, the precise “elasticity” of demand for any service or commodity. A better name for “elasticity” of demand is responsiveness of demand. The latter phrase at least makes it clearer that what we are talking about is the decisions and actions of employers or consumers, and not some inherent quality in the service or commodity itself. But as changes in price can never be assumed to be the sole reason for changes in the quantity demanded, and as “other conditions” [including the “demand curve” itself] can never safely be assumed to be precisely the same for any two years, two days, or two moments in succession, it follows that the “elasticity” or responsiveness of demand is never precisely measureable. On what reasonably appear to be fairly persistent relationships, however, we may be reasonably justified in basing practical policies.)

4. Stabilize Wage-Rates—or Employment?

If competition between unemployed workers always led to a very great reduction in money-wage, there would be violent instability in the price level.... The wage-unit might have to fall without limit until it reached a point where the effect of the abundance of money in terms of the wage-unit on the rate of interest was sufficient to restore a level of full employment (p. 253).

There are more fallacies in this passage than the reader is likely to have patience to examine. Keynes is apparently trying to prove that if there were free competition among workers, instead of union-enforced or law-enforced inflexibility downwards, the result would be intolerably and limitlessly violent oscillations in prices.

The proposition is just as absurd as it sounds. Price changes normally come first, and determine wage-rate changes, rather than vice versa. It is far better, when the choice must be made, to have wide oscillations in prices than wide oscillations in production and employment. The attempt to “stabilize” farm prices at levels above those that would be set by a free, competitive market, as American experience has so dramatically proved, merely leaves unsold farm “surpluses” that pile up in government warehouses. The attempt to stabilize wages at levels above those that would be set by a free, competitive market leaves unemployed surpluses of labor that pile up on government unemployment insurance or relief rolls. We do not stabilize the economy by trying to hold up wages regardless of what happens to prices. We unstabilize it, and create the very mass unemployment that Keynes professes to wish to cure.

It is significant that the Keynesians do not dare to apply their theory both ways. They do not urge that wage-rates be held down when prices soar, in order to stabilize prices by bringing them down again.

Keynes’s wage theories are useful only as labor-union propaganda. Their “scientific” pretensions are pure quackery.

In the passage quoted above from page 253 of the General Theory, Keynes drags in the effect of a reduction of wage-rates on the interest rate. Of course, the interconnection of all prices (and both wage-rates and interest rates are “prices” in the broadest sense) is such that there is some interrelationship between wage-rates and interest rates. But the interrelationship is so complex and for the most part so indirect that a lengthy discussion of this point would be largely irrelevant digression.

We have already seen that Keynes had a false theory of interest. We shall soon see that he had also a false theory of wage-rates, a false theory of money and credit, and a false theory of prices.

¹Science, November 21, 1958, p. 1259.

²The Canadian Journal of Economics and Political Science, February, 1937, pp. 120-121.

³ For a fuller account of what happened to prices, wages, and production in both Britain and America in the twenties and the thirties, the reader may consult Benjamin M. Anderson, Economics and the Public Welfare (New York: Van Nostrand, 1949).

⁴ Paul H. Douglas, The Theory of Wages (New York: Macmillan, 1934), pp. 113-158 and 501-502.

⁵ A. C. Pigou, The Theory of Unemployment (London: Macmillan, 1933).