Debunking the Keynesian Accelerator Theory

Skousen, M., ed., (1992). Dissent on Keynes: A Critical Appraisal of Keynesian Economics. New York and London: Praeger.

Chapter 4 : The Myths of the Multiplier and the Accelerator, by Jeffrey M. Herbener

The Accelerator

While interest in the trade cycle stimulated development of the acceleration principle, it is fundamentally a theory about the use of capital goods in the production of consumption goods. Thus, before analyzing the accelerationist doctrine, a theory of production is briefly reviewed below (Rothbard [1962] 1970: 40-61, 273-308).

The production of consumption goods on the free market occurs in stages where factors of production are mixed together to create capital goods in one stage that are then used in a subsequent stage to create other capital goods until, eventually, a consumption good is produced. The factors of production necessary to create capital along the stages come from setting aside wealth (saving) out of current production. The amount of capitalist saving, and thus the amount of created capital goods (investment), is determined by time preference — the desire for consumption goods sooner as opposed to later.

When individuals have a lower time preference, they save and invest relatively greater amounts and thus make larger additions to the size of the production structure, which, in turn, allows for greater future amounts of production and consumption. On the free market, this process of saving and investing is greatly enhanced (as all productive processes are) by specialization and the use of money. Those who specialize in saving are coordinated with those who specialize in investing by entrepreneurial arbitrage. [...]

Not only does entrepreneurial arbitrage lead to the amount of saving and investing that individuals desire, it also allocates this saved wealth into the highest-valued investment projects. This results from the feasibility of making meaningful profit/loss calculations and from the fact that action based on such calculations will earn a profit or suffer a loss. Economic calculation also allows individuals to fit different production processes (within each stage and between stages) together into an orderly structure.

Finally, because of free exchange based on private property, this structure of production is both flexible and complex. Flexibility (both within and between stages) comes from the fact that when individuals, as consumers or resource owners, change their valuations, entrepreneurs can profit by changing the structure to satisfy them. Complexity occurs because each entrepreneur needs to plan production only with regard to the part of the structure that directly influences his profit. Thus, the structure can extend far beyond the ability of a single mind to plan its detail. That the structure is flexible and complex should be apparent from even cursory observations of a market economy.

In contrast, the acceleration principle envisions the role of capital in production as simple, fixed, and mechanical. Knox characterized this relationship by the equation: Ct = Kt - Kt-1 = a(Ot - Ot-1), which relates the current production of capital (Ct) to a firm's current increase in the stock of capital (Kt - Kt-1) to the constant ratio (a) between that amount and the current increase in production of consumption goods (Ot - Ot -1). According to the equation, there are two parts to the acceleration principle: one is the identity between capital produced and a change in the stock of capital; the other is a theory of investment: that there is a fixed ratio between changes in capital stock and changes in output (Knox 1952: 271).

From the first part of the theory, Knox derived two important implications of his assertion that the stock of capital exceeds the current output of capital. The first, Knox stated, is the possibility that "the stock of equipment is rising, but not at a sufficiently fast rate to prevent a decline in the output of the capital goods industries" (ibid.: 272). Second, "a given percentage fluctuation in stock means a more than proportionate fluctuation in output; and the more durable the stock the greater is the relative violence of the movements in output" (ibid.). Knox continued:

The effects of this purely technical relationship in matters of timing and amplitude are very useful in the analysis of business cycles; and its validity is unquestionable. This paper, therefore, is concerned with the second part of the acceleration principle: the theory of investment. So long as the accelerator is constant, the conclusions about the relationship between K and C apply also, once due allowances have been made for lags, to the timing and amplitude of fluctuations in K and in O. The crucial problem of the acceleration principle is whether the accelerator is constant. . . . The equation Kt - Kt-1 = a(Ot - Ot-1) states the acceleration principle in its most uncompromising form: net induced investment is solely a function of the rate of growth of final output. (Ibid.: 273) Furthermore, this relationship is assumed to hold for an entire economy.

As Paul Samuelson and William Nordhaus explain:

According to this law, society's needed stock of capital, whether inventory or equipment, depends primarily upon the level of income or production. Additions to the stock of capital, or what we customarily call net investment, will take place only when income is growing. As a result, a prosperity period may come to an end, not simply because consumption sales have gone down, but merely because sales have leveled off at a high level (or have continued to grow but at a lower rate than previously). (Samuelson and Nordhaus 1989: 215)

On this last contention, they continue:

A simplified arithmetical example will make this clear. Imagine a typical textile-manufacturing firm whose stock of capital equipment is always kept equal to about two times the value of its yearly sales of cloth. Thus, when its sales have remained at $30 million per year for some time, its balance sheet will show $60 million of capital equipment, consisting of perhaps 20 machines of different ages, with one wearing out each year and being replaced. Because replacement just balances depreciation, there is no net investment or saving being done by the corporation. Gross investment takes place at the rate of $3 million per year, representing the yearly replacement of one machine. (The other $27 million of sales may be assumed to be wages and dividends.). . . Now let us suppose, that, in the fourth year, sales rise 50 percent — from $30 million to $45 million. Then the number of machines must also rise 50 percent, or from 20 to 30 machines. In the fourth year, instead of one machine, eleven machines must be bought — ten new ones in addition to the replacement of the worn-out one. (Ibid., p. 216) According to the acceleration principle, consumption has to continue to keep increasing in order for investment to stand still. (Ibid.: 215-16)

This result renders a more robust theory of recessions and expansions than Keynes offered and gives further justification for corrective fiscal and monetary policy designed to continue the boom period. Yet both parts of the accelerator principle described by Knox are subject to criticism (Hutt 1963: 289-339).

The first part of the accelerator principle maintains that the current production of capital is identical to the current increase in the stock of capital. Yet this identity never holds when the stock of capital goods depreciates. The left-hand side of Knox's equation is gross investment, while the right-hand side is net investment. To create the identity, depreciation must be added to the right-hand side, which, in turn, destroys the second part of the equation. Moreover, Knox's equation is stated in physical terms and is intended to describe the entire economy. But using physical amounts for this purpose is absurd because different types of capital goods cannot be added together. If one uses money as a common denominator and adds the monetary value of different capital goods, the meaning of the second part of the equation is destroyed. The only alternative that allows adding up capital (and output) is to assume homogeneous capital goods (and output). However, this assumption denies reality, that is, the existence of the production structure (Mises [1949] 1966: 200-211).

Granting the existence of the identity, Knox derived his two implications from the assertion that the stock of capital goods exceeds the current production of capital goods. Yet merely writing down the identity does not make it so. When a group of individuals first begins the process of capital formation (and thus, civilization), it is not impossible for current saving and investing to exceed the existing amount of capital. In general, this claim could stem from the fortunate occurrence of a relatively large degree of past saving and investing or the unfortunate one of a relatively small degree of current saving and investing.

The first implication that Knox derived from this claim misstates the time sequence of the process of capital formation. The degree of current saving and investing is determined by time preference, which, in turn, leads to a change in the stock of capital goods. Individuals desire as an (intermediate) end a stock of capital (and its efficient allocation); they use saving and investing as means to that end. If individuals desire a smaller capital stock (or a slower increase in the stock), then it is beneficial to reduce capital production.

The acceleration principle turns this around, implying that the end is an increasing amount of capital production, which, in turn, is determined by the stock of capital "necessary" to produce output. Additionally, the process of producing capital (Ct) is carried out by firms specializing in that activity, who thus are separate from those firms adding this capital to their stock (Kt - Kt-1).

If additions to the stock of capital are mechanically linked to additions to final output, what happens to capital goods created to produce other capital goods? Because most capital goods are devoted to the higher stages of production, Ct in Knox's equation applies only to a small segment of the production structure, that is, the first stage. If C, is meant to include all capital, then the second part of the equation is destroyed.

The second implication that Knox derived from the first part of the acceleration principle is subject to the same criticisms made above. The acceleration principle reverses cause and effect; individuals change their degree of saving and investing in order to change their stock of capital goods, not vice versa. Furthermore, no significance can be attached to the fact that the percentage change in the stock is smaller than the percentage change in the amount of current capital production. The only meaningful calculation in capital formation is profit and loss.

If entrepreneurs anticipate that changing production of some good to match the variation in its demand will be less profitable, then they will even out production over time by "overproducing" in periods of low demand to provide more of that good in periods of high demand. This entrepreneurial arbitrage over time is much easier for durable goods, such as capital goods, than for nondurable ones. Finally, if Knox's implication has significance, why not develop an accelerator for all goods where the existing stock exceeds the level of current production? Why limit the principle to first-order capital goods?

Contrary to Knox's belief, the first part of the acceleration principle — the identity between capital produced and a change in the stock of capital — is not "unquestionable." It is insignificant, and thus his two implications have no usefulness in analyzing the business cycle. The fact that investment varies more than consumption over the business cycle is not a consequence of the acceleration principle, but can be explained by the Misesian theory of the business cycle (Mises [1949] 1966: 538-86; Rothbard [1963] 1983a: 11-38).

Regarding the second part of the acceleration principle — the theory of investment — two general criticisms can be made. First, the addition to the capital stock, net investment, is not a function (solely or partially) of the rate of growth of final output. It is the means by which individuals increase the production of future consumption goods at the expense of the production of current consumption goods (as well as changing the numbers of specific consumption goods across time).

Stated differently, current net investment causes the number of future consumption goods to increase because individuals use capital formation to accomplish exactly that outcome. Second, contrary to the accelerator principle, no relationship exists, in either direction, between current changes in output and current changes in the stock of capital. In a world of scarcity, one must precede the other. Thus, the later one can only bring about future changes in output.

And the increase in current output (from existing factors of production) can lead only to a future increase in the capital stock. If existing factors of production are used to increase current output, they cannot simultaneously be used to increase the current capital stock.

In addition to these general criticisms of the basic statement of the theory of investment, the presentation of the principle by Samuelson and Nordhaus (1989) faces specific criticisms. First, society has no "need" for a certain stock of capital. Individuals choose the amount of capital that they prefer — more with low time preference and less with high time preference. Second, prosperity is not equivalent to growing consumption sales. Prosperity means fulfilling valuable ends, one of which may be reducing current consumption to create capital goods for greater future consumption.

Third, why would a firm always keep a stock of capital goods equal to two times the value of yearly sales? By this statement Samuelson and Nordhaus must mean the value of the stock since the physical amount of capital cannot be compared with the value of output. However, the fixed mechanical accelerator is understandable only in physical terms. This fact is illustrated by their statement that a 50 percent increase in sales requires a 50 percent increase in the number of machines. But they must either compare the value of output with the value of the machines or the amount of output with the amount of machinery.

If they mean the former, then they must demonstrate why the market process of imputing value to factors of production would lead to equal proportional changes in the values of output and capital. It is certainly true that if an increase in demand for the consumption good leads to greater profit, entrepreneurs will bid more intensely for the factors (including capital) used to produce this good. But the resulting changes in prices (and total market values) are determined by the highly complex set of interactions among individuals. No implication of equal proportional changes in value can be derived. On the other hand, if Samuelson and Nordhaus mean to compare amounts of capital with amounts of output, then they must disprove the existence of the production structure.

As demonstrated above, the production of any particular consumption good is flexible since many different ways exist to combine factors of production. Entrepreneurs will gain profit by continually adjusting the production technique as subjective values change. Firms which refuse to alter their methods will take losses and eventually disappear. Furthermore, even with a given set of capital goods, every level of sales from zero to maximum capacity can be produced. Thus, the acceleration principle cannot exist with unused capacity.

Fourth, even if a fixed relationship exists, as Samuelson and Nordhaus believe, they have not made a case for the fixed accelerator. In their example, when sales of output increase by 50 percent, it implies nothing certain about the amount of output sold. Typically, as demand for a good increases, both price and quantity increase. Since the amount of output increases by less than 50 percent, the number of machines used to produce the additional output will increase by less than 50 percent. In fact, if the 50 percent rise in sales came solely from a higher price (not likely), the percent change in the number of machines would be zero since no additional output would be produced. Additionally, Samuelson and Nordhaus's selection of a 50 percent increase appears to be a deliberate exaggeration designed to make their case seem more plausible.

They surely do not mean to imply a 50 percent increase in aggregate consumption, yet they do imply that the accelerator is an aggregate concept. Selecting a more reasonable 10 percent increase in sales for a single firm, or perhaps 3 percent for the entire economy, makes producing this additional amount from unused capacity appear quite feasible. Even if a firm actually had a 50 percent increase in the amount of output sold, that still would not imply a proportionate increase in the number of new machines produced. Such a firm could use the existing capacity of other firms by subcontracting the work or by purchasing used capital goods, and so on. Even if the increase in sales could only be satisfied by an increase in the production of new machinery, the acceleration principle would not apply to an entire economy. Increasing the amount of textile machinery requires resources which, in the absence of additional savings, must be transferred from the production of other capital goods. The total production of all capital goods can only increase when total consumption falls, that is, when saving increases.

Fifth, if the resources needed to produce more capital goods in total can be acquired out of thin air, then Samuelson and Nordhaus understate the acceleration principle. Since the production of each capital good is linked to others in the structure of production, the accelerator must proceed from stage to stage in a fashion similar to the multiplier. However, with the accelerator, each "round" can potentially become larger instead of smaller, as with the multiplier. In their example, a 50 percent increase in the sale of cloth causes a 1000 percent increase in the production of textile machinery at the first stage of production. What percentage increase is dictated for the capital goods "needed" to produce the textile machinery, that is, second-stage capital goods? Depending on the size of the accelerator coefficient (a) for the relationship between second-stage and first-stage capital goods, second-stage capital production could increase by more than 1000 percent. And so would the process go for the other stages of production.

Simply writing down the accelerator equation does not preclude this possibility; the equation cannot tell if Ot stands for a consumption good or an intermediate capital good. If C, is meant to stand for all capital goods, then the equation is nonsense since there is no method by which to add different types of capital goods together. If the equation implicitly assumes that all capital goods are homogeneous (and thus, can be added up), then it denies the existence of the production structure.

Sixth, even if the acceleration principle operates just as Samuelson and Nordhaus state, it is only temporary. In the third year, one additional machine is produced and used to replace the one machine out of the total twenty that wears out. Then in the fourth year, when sales increase by 50 percent, eleven machines are produced — ten new ones plus one for replacement (a 1000% annual increase in gross investment).

But in the fifth year, with sales the same as in the fourth year, the firm will need only one and one-half machines for replacement and no new machines. Thus, except for the fourth year, the production of capital increases by 50 percent, the same percentage increase as in sales. This effect is not only not accelerated, it is not even an upward trend, but just a one-year blip. Furthermore, entrepreneurial activity would smooth out this blip in capital-goods production over time, if it were profitable. Thus, Samuelson and Nordhaus's final contention — that consumption must continue to increase in order for investment to stand still — is incorrect.

Seventh, their comments about gross investment, net investment, and saving are irrelevant for the acceleration principle. The number of capital goods produced depends only on gross investment and not at all on net investment. Also, the fact that net investment is zero implies nothing about the level of saving attained by the firm. Since saving can be a specialized process separate from investing, the firm does not (and probably would not) have to save in the form of the capital it uses in production. However, most firms are net savers because they save in the process of production by supplying current wealth to resource owners in advance of receiving wealth from the sale of output (Rothbard [1962] 1970: 273-312).

Finally, Samuelson and Nordhaus do not explain where the funds come from to increase the sales of cloth. If they are transferred from the sale of other consumption goods, then a negative acceleration would occur there, and it would offset the positive acceleration in cloth. If the funds come from reduced saving, then fewer resources exist to produce capital goods and no acceleration can occur. If they come from monetary inflation via credit expansion, then capital production increases before the funds become available for consumption purchases.

This latter case is part of the Misesian theory of the business cycle and fully explains why capital-goods production varies more over the cycle than the production of consumption goods (Mises [1949] 1966: 538-86). In summary, the acceleration principle misconceives the economic process of capital formation. Its mathematical premises lead only to fallacious inferences that are inconsistent with the theory of production and, thus, are unable to advance our understanding of business cycles.

Rothbard, Murray N. ([1962] 2009). Man, Economy, and State: A Treatise on Economic Principles with Power and Market: Government and the Economy. Scholar’s Edition. 2nd ed. Auburn, AL: Ludwig von Mises Institute.

Chapter 11 — Money and Its Purchasing Power

18. The Fallacy of the Acceleration Principle

The “acceleration principle” has been adopted by some Keynesians as their explanation of investment, then to be combined with the “multiplier” to yield various mathematical “models” of the business cycle. The acceleration principle antedates Keynesianism, however, and may be considered on its own merits. It is almost always used to explain the behavior of investment in the business cycle.

The essence of the acceleration principle may be summed up in the following illustration:

Let us take a certain firm or industry, preferably a first-rank producer of consumers’ goods. Assume that the firm is producing an output of 100 units of a good during a certain period of time and that 10 machines of a certain type are needed in this production. If the period is a year, consumers demand and purchase 100 units of output per year. The firm has a stock of 10 machines. Suppose that the average life of a machine is 10 years. In equilibrium, the firm buys one machine as replacement every year (assuming it had bought a new machine every year to build up to 10). 78

78. It is usually overlooked that this replacement pattern, necessary to the acceleration principle, could apply only to those firms or industries that had been growing in size rapidly and continuously.

Now suppose that there is a 20-percent increase in the consumer demand for the firm’s output. Consumers now wish to purchase 120 units of output. Assuming a fixed ratio of capital investment to output, it is now necessary for the firm to have 12 machines (maintaining the ratio of one machine: 10 units of annual output). In order to have the 12 machines, it must buy two additional machines this year. Add this demand to its usual demand of one machine, and we see that there has been a 200-percent increase in demand for the machine. A 20-percent increase in demand for the product has caused a 200-percent increase in demand for the capital good. Hence, say the proponents of the acceleration principle, an increase in consumption demand in general causes an enormously magnified increase in demand for capital goods. Or rather, it causes a magnified increase in demand for “fixed” capital goods, of high durability. Obviously, capital goods lasting only one year would receive no magnification effect. The essence of the acceleration principle is the relationship between the increased demand and the low level of replacement demand for a durable good. The more durable the good, the greater the magnification and the greater, therefore, the acceleration effect.

Now suppose that, in the next year, consumer demand for output remains at 120 units. There has been no change in consumer demand from the second year (when it changed from 100 to 120) to the third year. And yet, the accelerationists point out, dire things are happening in the demand for fixed capital. For now there is no longer any need for firms to purchase any new machines beyond what is necessary for replacement. Needed for replacement is still only one machine per year. As a result, while there is zero change in demand for consumers’ goods, there is a 200-percent decline in demand for fixed capital. And the former is the cause of the latter. In the long run, of course, the situation stabilizes into an equilibrium with 120 units of output and one unit of replacement. But in the short run there has been consequent upon a simple increase of 20 percent in consumer demand, first a 200-percent increase in the demand for fixed capital, and next a 200-percent decrease.

To the upholders of the acceleration principle, this illustration provides the key to some of the main features of the business cycle: the greater fluctuations of fixed capital-goods industries as compared with consumers’ goods, and the mass of errors revealed by the crisis in the investment goods industries. The acceleration principle leaps boldly from the example of a single firm to a discussion of aggregate consumption and aggregate investment. Everyone knows, the advocates say, that consumption increases in a boom. This increase in consumption accelerates and magnifies increases in investment. Then, the rate of increase of consumption slows down, and a decline is brought about in investment in fixed capital. Furthermore, if consumption demand declines, then there is “excess capacity” in fixed capital — another feature of the depression.

The acceleration principle is rife with error. An important fallacy at the heart of the principle has been uncovered by Professor Hutt. We have seen that consumer demand increases by 20 percent; but why must two extra machines be purchased in a year? What does the year have to do with it? If we analyze the matter closely, we find that the year is a purely arbitrary and irrelevant unit even within the terms of the example itself. We might just as readily take a week as the period of time. Then we would have to say that consumer demand (which, after all, goes on continuously) increases 20 percent over the first week, thereby necessitating a 200-percent increase in demand for machines in the first week (or even an infinite increase if the replacement does not precisely occur in the first week), followed by a 200-percent (or infinite) decline in the next week, and stability thereafter. A week is never used by the accelerationists because the example would then be glaringly inapplicable to real life, which does not see such enormous fluctuations in the course of a couple of weeks. But a week is no more arbitrary than a year. In fact, the only non arbitrary period to choose would be the life of the machine (e.g., 10 years). Over a ten-year period, demand for machines had previously been ten (in the previous decade), and in the current and succeeding decades it will be 10 plus the extra two, i.e., 12. In short, over the 10-year period the demand for machines will increase precisely in the same proportion as the demand for consumers’ goods — and there is no magnification effect whatever.

Since businesses buy and produce over planned periods covering the life of their equipment, there is no reason to assume that the market will not plan production suitably and smoothly, without the erratic fluctuations manufactured by the model of the acceleration principle. There is, in fact, no validity in saying that increased consumption requires increased production of machines immediately; on the contrary, it is only increased saving and investment in machines, at points of time chosen by entrepreneurs strictly on the basis of expected profit, that permits increased production of consumers’ goods in the future.

Secondly, the acceleration principle makes a completely unjustified leap from the single firm or industry to the whole economy. A 20-percent increase in consumption demand at one point must signify a 20-percent drop in consumption somewhere else. For how can consumption demand in general increase? Consumption demand in general can increase only through a shift from saving. But if saving decreases, then there are less funds available for investment. If there are less funds available for investment, how can investment increase even more than consumption? In fact, there are less funds available for investment when consumption increases. Consumption and investment compete for the use of funds.

Another important consideration is that the proof of the acceleration principle is couched in physical rather than monetary terms. Actually, consumption demand, particularly aggregate consumption demand, as well as demand for capital goods, cannot be expressed in physical terms; it must be expressed in monetary terms, since the demand for goods is the reverse of the supply of money on the market for exchange. If consumer demand increases either for one good or for all, it increases in monetary terms, thereby raising prices of consumers’ goods. Yet we notice that there has been no discussion whatever of prices or price relationships in the acceleration principle. This neglect of price relationships is sufficient by itself to invalidate the entire principle. The acceleration principle simply glides from a demonstration in physical terms to a conclusion in monetary terms.

Furthermore, the acceleration principle assumes a constant relationship between “fixed” capital and output, ignoring substitutability, the possibility of a range of output, the more or less intensive working of factors. It also assumes that the new machines are produced practically instantaneously, thus ignoring the requisite period of production.

Huerta de Soto, Jesús ([2006] 2009). Money, Bank Credit and Economic Cycles. 2nd ed. Auburn, Ala.: Ludwig von Mises Institute.

Chapter 7 : A Critique of Monetarist and Keynesian Theories

Criticism of the “Accelerator” Principle

Our theory on the impact of credit expansion on the structure of production rests on a capital theory we examined in detail in chapter 5. According to this theory, a healthy, permanent “lengthening” of the productive structure is contingent on a prior increase in saving. Therefore we must criticize the so-called “accelerator principle,” developed by the Keynesian School. Those who accept this principle assert that any increase in consumption leads to a more-than-proportional increase in investment, which is contrary to what our theory suggests.

In fact, according to the accelerator principle, a rise in the demand for consumer goods and services provokes an exaggerated upsurge in the demand for capital goods. The argument centers around the notion that a fixed relationship exists between the output of consumer goods and the number of machines necessary to produce them. Thus any rise in the demand for consumer goods and services causes a proportional increase in the number of machines necessary to produce them. When we compare this new number with that normally demanded to compensate for the customary depreciation of the machines, we see an upturn in the demand for capital goods which is far more than proportional to the rise in the demand for consumer goods and services. 80

80. Samuelson provides the following example to illustrate the accelerator principle:

Imagine a typical textile firm whose stock of capital equipment is always kept equal to about 2 times the value of its yearly sales of cloth. Thus, when its sales have remained at $30 million per year for some time, its balance sheet will show $60 million of capital equipment, consisting of perhaps 20 machines of different ages, with 1 wearing out each year and being replaced. Because replacement just balances depreciation, there is no net investment or saving being done by the corporation. Gross investment takes place at the rate of $3 million per year, representing the yearly replacement of 1 machine. . . . Now let us suppose that, in the fourth year, sales rise 50 per cent — from $30 to $45 million. Then the number of machines must also rise 50 per cent, or from 20 to 30 machines. In that fourth year, instead of 1 machine, 11 machines must be bought — 10 new ones in addition to the replacement of the worn-out one. Sales rose 50 per cent. How much has machine production gone up? From 1 machine to 11; or by 1,000 percent! (Samuelson, Economics, 11th ed. [New York: McGraw-Hill, 1980], pp. 246–47)

Interestingly, the analysis of the accelerator principle was eliminated from the 15th edition of the book, published in 1992.

We know that according to the accelerator principle, an increase in the demand for consumer goods and services brings about tremendously magnified growth in the demand for capital goods. However the principle also implies that if the demand for capital goods is to remain constant, the demand for consumer goods and services will have to continue to rise at a progressively increasing rate. This is due to the fact that a steady demand for consumer goods and services, i.e., a demand which does not increase, will provoke a marked contraction in the demand for equipment goods. The demand for these goods will return to the level necessary for replacements only. The accelerator principle clearly and perfectly fits the Keynesian prescriptions of an unlimited expansion of consumption and aggregate demand: indeed, the accelerator doctrine indicates that any rise in consumption causes a huge upsurge in investment, and that saving is of no importance!

The theory based on the accelerator not only omits the most elementary principles of capital theory; it was also developed based on a mechanistic, automatic and fallacious conception of economics. Let us analyze each of the reasons behind this assertion.

First, the accelerator theory excludes the real functioning of the entrepreneurial market process and suggests that entrepreneurial activities are nothing more than a blind, automatic response to momentary impulses in the demand for consumer goods and services. However entrepreneurs are not robots, and their actions are not mechanical. On the contrary, entrepreneurs predict the course of events, and with the purpose of obtaining a profit, they act in light of what they believe may happen. Hence no transmitter mechanism automatically and instantaneously determines that growth in the demand for consumer goods and services will trigger an immediate, proportional increase in the demand for capital goods. Quite the opposite is true. In view of potential variations in the demand for consumer goods and services, entrepreneurs usually maintain a certain amount of idle capacity in the form of capital equipment. This idle capacity allows them to satisfy sudden increases in demand when they occur. The accelerator principle proves to be much less sound when, as in real life, companies keep some capital goods in reserve.

Therefore it is obvious that the accelerator principle would only be sound if capital goods were in full use, such that it would be impossible to raise the output of consumer goods at all without increasing the number of machines. Nevertheless, and second, the great fallacy of the accelerator principle is that it depends on the existence of fixed, unchanging proportions between capital goods, labor and the output of consumer goods and services. The accelerator principle fails to take into account that the same result in terms of consumer goods and services can be achieved using many different combinations of fixed capital, variable capital and especially, labor. The specific combination an entrepreneur may choose in any given case depends on the structure of relative prices. Hence, to assume fixed proportions exist between the output of consumer goods and services and the quantity of capital goods necessary to produce them is an error, and it contradicts the basic principles of the theory of prices in the factor market. Indeed, as we saw when we analyzed the “Ricardo Effect,” a drop in the relative price of labor will lead companies to produce consumer goods and services in a more labor-intensive manner, i.e., using fewer capital goods in relative terms. The reverse is also true: a rise in the relative cost of labor will trigger a relative increase in the use of capital goods. Because the accelerator principle rests on the assumption that fixed proportions exist between the factors of production, it totally excludes the role entrepreneurship, the price system and technological change play in market processes.

Furthermore, and third, even if, for the sake of argument, we suppose fixed ratios exist between consumption and capital equipment used, and we even assume there to be no idle capacity with respect to capital goods, we must ask ourselves the following question: How can the output of capital goods possibly rise in the absence of the saving necessary to finance such an investment? It is an insoluble logical contradiction to consider that an increase in the demand for consumer goods and services will automatically and instantaneously provoke a much-more-than-proportional rise in the output of capital goods, given that in the absence of excess capacity the production of these goods is contingent on growth in voluntary saving. Moreover such growth inevitably entails a momentary drop in the demand for consumer goods (which clearly contradicts the premise on which the accelerator theory is based). Therefore the accelerator theory contradicts the most fundamental principles of capital theory.

Fourth, it is important to realize that an investment in capital goods which is far more than proportional to the increase in the demand for consumer goods can only be financed if substantial credit expansion is initiated and sustained. In other words, the accelerator principle ultimately presupposes that the increase in credit expansion necessary to stimulate an enormously exaggerated investment in capital goods takes place. We are already familiar with the effects such credit expansion exerts on the productive structure and with the way in which the relative-price system invariably limits the expansion and forces a reversal that manifests itself in a crisis and recession. 82

82. [I]f, for the sake of argument, we were ready to admit that capitalists and entrepreneurs behave in the way that the disproportionality doctrines describe, it remains inexplicable how they could go on in the absence of credit expansion. The striving after such additional investments raises the prices of the complementary factors of production and the rate of interest on the loan market. These effects would curb the expansionist tendencies very soon if there were no credit expansion. (Mises, Human Action, p. 586)

Fifth, it is absurd to expect a rise in the demand for consumer goods and services to cause an instantaneous upsurge in the output of capital goods. We know that during the boom, which is financed by credit expansion, companies and industrial sectors devoted to the production of equipment and capital goods operate at maximum capacity. Orders pile up and companies are unable to satisfy the increased demand, except with very lengthy time lags and dramatic increases in the price of equipment goods. Therefore it is impossible to imagine that a rise in the output of capital goods could take place as soon as the accelerator principle presupposes.

Sixth, the accelerator theory rests on peculiar mechanistic reasoning by which an attempt is made to relate growth in the demand for consumer goods and services, measured in monetary terms, with a rise, in physical terms, in the demand for equipment and capital goods. Entrepreneurs never base their decisions on a comparison between monetary and physical magnitudes; instead they always compare estimated income and costs, measured strictly in monetary terms. To compare heterogeneous magnitudes is absurd and makes entrepreneurial economic calculation utterly impossible. Obviously, if the price of capital goods begins to increase, entrepreneurial decisions will not mechanically manifest themselves in “fixed proportions” of inputs. Instead entrepreneurs will carefully monitor the evolution of costs to determine the extent to which production will continue at the old proportions, or they will start using a higher proportion of alternative factors, specifically labor.

In short the accelerator principle rests on fallacious, mechanistic reasoning which excludes the most elementary principles of the market process, specifically the nature of entrepreneurship. The doctrine ignores the functioning and effects of the price system, the possibility of substituting certain inputs for others, the most essential aspects of capital theory and of the analysis of the productive structure, and finally, the microeconomic principles which govern the relationship between saving and the lengthening of the productive structure.