by Friedrich A. Hayek, 1941.
THE SCOPE OF THE INQUIRY
The subject of this study is indicated in the title by the heading under which it is conventionally treated. It gives, however, no indication of the approach which we shall adopt. The contents of the following pages would perhaps have been more appropriately described as an Introduction to the Dynamics of Capitalistic Production, provided the emphasis were laid on the word Introduction, and provided that it were clearly understood that it deals only with a part of the wider subject to which it is merely a preliminary. The whole of the present discussion is essentially preparatory to a more comprehensive and more realistic study of the phenomena of capitalistic production, and it stops deliberately short of some of the most important problems that fall within that wider context.
The central aim of this study is to make a systematic survey of the interrelations between the different parts of the material structure of the process of production, and the way in which it will adapt itself to changing conditions. In so far as these complex problems have been explicitly discussed in the past they have been treated as part of the theory of capital and interest. Here they will be treated from a somewhat different viewpoint. Our main concern will be to discuss in general terms what type of equipment it will be most profitable to create under various conditions, and how the equipment existing at any moment will be used, rather than to explain the factors which determined the value of a given stock of productive equipment and of the income that will be derived from it. As will appear presently, there are in this field a number of fairly important and difficult problems which fall into what is usually regarded as the sphere of equilibrium analysis, but which have not yet received adequate attention. By far the greater part of the present investigation will be confined to that part of the subject which belongs to equilibrium analysis proper. A full treatment of the economic process as it proceeds in time, and of the monetary problems that are connected with this process, is outside the compass of this book. The discussion in justification of the distinction that is involved here, and of the methodological issues underlying it, will be reserved for the two following chapters. All that I wish to explain at this point is why the task of merely putting those elements of the theory of capital which are commonly treated as belonging to general equilibrium analysis into a form in which they will prove useful for the analysis of the monetary phenomena of the real world, is important enough to merit a separate study.
It may at first be somewhat disconcerting to be told that the theory of a subject which has been so widely and so vigorously discussed right from the beginning of economic science as the theory of capital, should need almost complete recasting as soon as we try to use its resuIts in the analysis of the more complex phenomena of the real world. But there are very good reasons why the theory of capital in the form in which it now exists has proved less useful than we should wish for the purposes for which we now need it. The fact is that the problems of capital as here understood, that is, the problems arising out of the dependence of production on the availability of “capital” in certain forms and quantities, have hardly ever been studied for their own sake and importance. And, as we shall see, the theory of stationary equilibrium, within which they were treated, did not really offer any opportunity for their explicit discussion. Such analysis as they have received has been almost entirely subordinate to another problem, the problem of explaining interest. And the treatment of the theory of capital as an adjunct to the theory of interest has had somewhat unfortunate effects on its development. This for two reasons.
Firstly, it was carried only just so far as seemed necessary for the main purpose of explaining interest, and this explanation aimed at illustrating a general principle by the simplest imaginable cases rather than at providing an adequate account of the interrelationships under more complex conditions.
Secondly, and this is even more important, the attempts to explain interest, by analogy with wages and rent, as the price of the services of some definitely given “factor” of production,  has nearly always led to a tendency to regard capital as a homogeneous substance the “quantity” of which could be regarded as a “datum”, and which, once it had been properly defined, could be substituted, for purposes of economic analysis, for the fuller description of the concrete elements of which it consisted. It was inevitable in these circumstances that different authors should have singled out different aspects of the same phenomenon as the relevant ones, and the consequences of this were those unending discussions about the “nature” of capital which are among the least edifying chapters of economic science.
 Cf. Armstrong, 1936, p. 3: “… the treatment of capital … as a factor of production on a par with land and labour has led to many erroneous conclusions”. (The full titles of the publications referred to in this manner will be found in the Bibliography at the end of this volume.)
There were of course praiseworthy exceptions, the most notable of which are to be found in the works of Jevons, Böhm-Bawerk, and Wicksell, who did at least begin with the analysis of the process of production and the role of capital in it, instead of with a concept of capital defined as some quasi-homogeneous magnitude. But even these faclor authors and their followers used this analysis only in order to arrive ultimately at some single definition which, for the purposes of further analysis, lumped together as one quasi-homogeneous mass all or most of the different items of man-made wealth; and this definition was then used in the place of the fuller description from which they had started.
As we shall see, it is more than doubtful whether the discussion of “capital” in terms of some single magnitude, however defined, was fortunate even for its immediate purpose, i.e. the explanation of interest. And there can be no doubt that for the understanding of the dynamic processes it was disastrous. The problems that are raised by any attempt to analyse the dynamics of production are mainly problems connected with the interrelationships between the different parts of the elaborate structure of productive equipment which man has built to serve his needs. But all the essential differences between these parts were obscured by the general endeavour to subsume them under one comprehensive definition of the stock of capital. The fact that this stock of capital is not an amorphous mass but possesses a definite structure, that it is organised in a definite way, and that its composition of essentially different items is much more important than its aggregate “quantity”, was systematically disregarded. Nor did it help much further when it was occasionally emphasised that capital was an “integrated organic conception”,  so long as such hints were not followed up by a careful analysis of the way in which the different parts were made to fit together.
 Knight, 1935a, p. 83.
This concentration on a particular capital concept to the neglect of all the multitudinous meanings which attach to the word capital in everyday speech has a further disadvantage. It is not only that the term capital in any of its “real” senses does not refer to a homogeneous substance. There is the further difficulty that even if we describe physically all the items of which the real structure of production is composed we have not described all the factors which will dictate their mode of utilisation. The various meanings of the term capital in everyday speech are an unconscious tribute to the complexity of the problem, and it has been unfortunate that the majority of authors seem to have assumed that somewhere or other there was some single substance corresponding to the singleness of the term which had discharged so many functions.
In fact there are at least two kinds of relevant magnitudes or rather proportions which must be taken into account if we want to understand the working of the price mechanism in this field; neither of them is a simple “quantity”, and neither of them stands in a unique relationship to the rate of interest except through its relation to the other. The first is the dimensions of the real structure of productive equipment, describing how it is organised for, or capable of, yielding various quantities of final output at different dates. The second is the proportional demands, or the relative prices, which are expected to rule for these different quantities of output at different dates. The first of these two quantitative relationships describes the proportions between the existing quantities of concrete resources in terms of their relative costs, while the second describes the relative demand for the two kinds of resources. But only together do these two sets of quantitative relationships or proportions determine what is usually regarded as the supply of capital in value terms.
The treatment of the capital problem in terms of the demand for and supply of one single magnitude is only possible on the assumption that the proportions just described stand in a certain equilibrium relationship to one another. On this assumption the result of a given supply of concrete capital goods meeting an exactly corresponding demand for them could be represented as a single-value magnitude, a quantity of capital in the abstract which could be set against a marginal productivity schedule for capital as such; and in this sense there would be a unique correlation between “the” quantity of capital and the rate of interest. As a first explanation of the rate of interest, the consideration of such an imaginary state of ultimate equilibrium may have certain advantages. There can be little doubt that the traditional theories of interest do little more than describe the conditions of such a long-term stationary equilibrium. Since this concept of long-term equilibrium assumes that the quantities of the individual resources measured in terms of costs are in perfect correspondence with their respective values, the description of capital in terms of an aggregate of value is sufficient. Even for the purposes of what is sometimes called “comparative statics”, that is the comparison of alternative states of stationary equilibrium, it is still possible to assume that the two magnitudes move in step with each other from one position of equilibrium to another, so that it never becomes necessary to distinguish between them.
The problem takes on a different complexion, however, as soon as we ask how a state of stationary equilibrium can ever be brought about, or what will be the reaction of a given system to an unforeseen change. It is then no longer possible to treat the different aspects of capital as one, and it becomes evident that the “quantity of capital” as a value magnitude is not a datum,  but only a result, of the equilibrating process.
 Cf. Wicksell, Lectures on Political Economy, vol. i, p. 202: “But it would clearly be meaningless – if not altogether inconceivable – to maintain that the amount of capital is already fixed before equilibrium between production and consumption has been achieved”.
With the disappearance of stationary equilibrium, capital splits into two different entities whose movements have to be traced separately and whose interaction becomes the real problem. There is no longer one supply of a single factor, capital, which can be compared with the productivity schedule of capital in the abstract: and the terms demand and supply, as referring to magnitudes which affect the rate of interest, take on a new meaning. It is the existing real structure of productive equipment (which in long-term equilibrium is said to represent the supply) which now determines the demand for capital; and to describe what constitutes the supply, writers have usually been compelled to introduce such vague and usually undefined terms as “free” or “disposable” capital. Even those writers who at earlier stages of their exposition have most emphatically decided in favour of only one of the meanings of the term capital, and that a “real” capital concept, later find it necessary either to use the word “capital” in another sense, or to introduce some new term for something which in ordinary language is also called capital. The consequent ambiguity of the term capital has been the source of unending confusion, and the suggestion has often been made  (and in one or two instances even put into practice ), that the term should be banned entirely from scientific usage. But much as there may be to be said in favour of this procedure, it seems on the whole preferable to use the expression as a technical term for one of the magnitudes in question, without, however, ignoring the other magnitudes which are sometimes denoted by this term. As will be more fully explained below (Chapter IV), we shall use the term capital as a name for the total stock of the non-permanent factors of production.
 E.g. by Schumpeter, Handwürterbuch der Staatswissenschaften, 4. Auti., vol. 5, p. 582.
 E.g. by Cannan, Elementary Political Economy (1888).
We cannot go into too many details at this stage. But it may be helpful to add a few words, by way of illustration, about the reasons for the general failure seriously to take account of the essentially non-homogeneous nature of the different capital items, and about the consequences of this failure. Two ideas in particular have had a very harmful effect on the whole theory of capital. The first is the idea that particular capital items represented a definite value independently of the use that could be made of them, a value which was apparently thought to be determined by the amounts “invested” in these items. This idea is a remnant of the old cost-of-production theories of value whose influence has lingered longer in the theory of capital than perhaps anywhere else in economic theory.  The second is the conception that additions to the stock of capital always mean additions of new items similar to those already in existence, or that an increase of capital normally takes the form of a simple multiplication of the instruments used before, and that consequently every addition is complete in itself and independent of what existed previously. This treatment of capital as if it consisted of a single sort of instrument or a collection of certain kinds of instruments in fixed proportions – a treatment which has won favour from the fact that it has sometimes been used explicitly as a supposed simplification – is perhaps more than anything else responsible for the idea that capital may be regarded as a simple, physically determined quantity, and that the rate of interest may be explained as a simple (decreasing) function of this quantity. It would of course follow from these assumptions that the rate of interest must steadily and continuously fall in the course of economic progress since every addition to the stock of capital would tend to lower it; and the familiar fact that the rate of interest fluctuates widely over comparatively short periods would appear to be without any foundation in the real facts and would therefore have to be ascribed entirely to the influence of monetary factors.
 Cf. Knight, 1935c, p. 45: “Historically, this notion goes back to the classical theory of capital as the product of labour, hence is an indirect consequence of that fountainhead of error, the labour theory of value.”
The organisation of the structure of real resources corresponding to any expected aggregate value of the existing stock of capital will of course depend on the kind of productive technique that is possible with that amount of capital. And the assertion that under equilibrium conditions a different structural organisation will be associated with a different value of the stock of capital means that changes in the supply of capital will bring about changes in the productive technique. The widely held idea that capital consists of (or is) a definite collection of instruments combined in fixed proportions, and the corollary of this idea, that there is at anyone time only one practicable productive technique (which is supposed to be determined either by the state of technological knowledge or by the already existing durable instruments) leads to another fallacy. This fallacy, which may be conveniently described as the “theory of derived demand”, has played an important role in recent discussions of trade cycle problems.
The error inherent in this view is of course not the mere assertion that the demand for productive equipment is derived from the expected demand for consumers’ goods, which is quite correct, but the idea that the amount of productive equipment which is required in order to satisfy an additional demand for consumers’ goods is uniquely determined by the “existing state of technique”. If the productive technique to be employed were fixed by extraeconomic factors, and particularly if it were assumed to be independent of the rate of interest, then a given change in the demand for consumers’ goods would indeed automatically be transmitted at a given rate to the earlier stages of production, and be transformed there into a demand for a uniquely determined quantity of equipment. This is a conclusion uniformly arrived at by authors who are able to think of an increase of capital only in terms of a simple duplication of equipment of the type already in existence,  and who completely disregard the changes in productive technique connected with the transition from less to more “capitalistic” methods of production and vice versa. This view has become widely known in the discussion of trade cycle problems as the “acceleration principle of derived demand”. It derives a certain specious plausibility from the fact that under certain monetary conditions things may for a time work in accordance with it.  But, as we shall see, the fact that monetary influences may sometimes temporarily obscure, or even reverse, the more permanent influences of the underlying real factors, is one of the main reasons why it is essential to make a systematic study of the significance of these real factors. A last instance may be mentioned of the unfortunate effects which these simplified ideas on capital have exerted on the analysis of dynamic phenomena such as industrial fluctuations. I refer to the crude investment distinction which is commonly made between current production and new investment, or between the reproduction of the existing stock of capital and additions to that stock, and the even cruder distinction between the gross production of capital goods and the production of consumers’ goods. Here too the idea that the growth of capital takes place in such a form that new items of a similar nature to those previously in existence are added to an otherwise unchanged stock, has been responsible for a good deal of confusion in contemporary discussion. The same applies to the cognate idea that for purposes of analysis the whole capital problem can be adequately dealt with by dividing industries into two groups, those producing consumers’ goods and those producing investment goods. But the problems involved here are obviously too complex to allow more than a mere mention at this stage. They are to some extent connected with the distinction between long and short periods, and the various concepts of equilibrium, which will be discussed in the next chapter.
 Although a great deal of the current discussion of trade cycle problems is to some extent affected by this idea, there is probably no other book by a reputable economist where it is used so crudely as in H. G. Moulton’s Formation of Capital (Washington, 1935), a book which is also, apart from this particular point, a veritable treasure-box of most of the current fallacies connected with capital.
 See Part IV below, and Hayek, 1939, where the significance of the “acceleration principle of derived demand” is discussed in some detail.
EQUILIBRIUM ANALYSIS AND THE CAPITAL PROBLEM
It was suggested in the first chapter that most of the shortcomings of the theory of capital in its present form are due to the fact that it has in effect only been studied under the assumptions of a stationary state, a stationary state where most of the interesting and important capital problems are absent. This is so largely because the characteristic problems of capital theory are problems of the interdependence of different industries and consequently only arise in connection with a theory of general equilibrium, and because most of the current systems of economic theory (particularly the most influential, that of Marshall) do not really consider any state of general equilibrium which is not at the same time stationary. The so-called short-term equilibria, if this concept is to have any meaning, must necessarily be conceived as partial equilibria.  And the long-period equilibrium, which alone is a general equilibrium, is (as Marshall himself has pointed out) identical with “the supposition of a stationary state of industry”. 
 The reason for this will become clear as we proceed. Here it need only be pointed out that the method of short-term equilibrium essentially consists in disregarding all these consequences of a given change whose significance, for the problem immediately under consideration, is of the second order of smalls. This means that we deliberately neglect consequences because they do not affect the parts of the system with which we are mainly concerned – a procedure which is clearly inadmissible when we are interested in the equilibrium of the system as a whole.
 Cf. A. Marshall, Principles of Economics, 7th ed., p. 379 note: “But in fact a theoretically. long period must give time enough to enable not only the factors of production of the commodity to be adjusted to demand, but also the factors of production of those factors of production to be adjusted, and so on; and this, when carried to its logical consequences, will be found to involve the supposition of a stationary state of industry in which the requirements of a future age can be anticipated an indefinite time beforehand. . . . Relatively short and long periods go generally on similar lines. In both use is made of that paramount device, the partial or total isolation for special study of some set of relations.” See also ibid. p. 367, where the stationary state is described as a state in which “the same amount of things per head of the population will have been produced in the same ways by the same classes of people for many generations together; and therefore this supply of the appliances for production will have had full time to be adjusted to the steady demand.”
An effective discussion of the problems of capital theory must, however, move precisely in that neglected field which deals with general equilibria that are not at the same time stationary states. It must proceed by way of a theory of general equilibrium because it deaIs with the interrelationships between groups of industries, and in particular with those effects of changes in one industry on another which are deliberately neglected when we study the particular short-period equilibrium of a special industry or group of industries. And it must not be confined to the stationary state, because here ex definitione most of the problems with which the theory of capital must be concerned have disappeared.  The main problems are to explain what types of instruments will be produced under given conditions, and what will be the consequences of producing particular instruments. And these problems will of course be non-existent if we assume from the beginning that the same stock of instruments will be constantly reproduced. The impossibility of treating the problems of capital adequately within the framework of a stationary equilibrium becomes, of course, even more obvious as soon as we include, as we must, the problems relating to what are usually described as “saving” and (new) “investment”, since these are activities which imply by definition that the persons undertaking them want to alter their future position, and consequently will do in the future something different from what they are doing in the present.
 Cf. W. E. Armstrong, 1936, p. 1: “All that is significant and vital in the concept of Waiting (as the equivalent of Capital) belongs to the economics of the developing community, and cannot without violent wrenching of ideas outside their proper context be transferred to the study of Stationary States”.
Perhaps the irrelevance of the stationary equilibrium construction for the treatment of capital problems comes out most clearly when we remember that this fictitious state could not conceivably be brought about at any given moment in society as it exists, but could be reached only after the lapse of a very long time.  The equipment which is given at any moment is always the inheritance from a past in which future developments have been foreseen only very imperfectly. And, as we shall see, it is precisely the existence of this equipment and its effect in determining what we can and what we cannot do for a very long time ahead, which constitutes the datum that creates the peculiar problem of capital. A theory which starts out by assuming that adjustments have proceeded to the point where no further changes are required is without relevance to our problems. What we need is a theory which helps us to explain the interrelations between the actions of different members of the community during the period (which is the only period of practical importance) before the material structure of productive equipment has been brought to a state which will make an unchanging, self-repeating process possible.
 Stationary equilibrium presupposes the existence of equilibrium relations between the existing things, that is, it assumes that the existing goods are of exactly the same kind as those which under existing conditions it will be profitable. to reproduce. It is not an equilibrium determined by the types of goods which happen to exist, but an equilibrium which has found expression in the past production of particular types and quantities of goods. For this reason it is without significance for the explanation of what happens prior to the time when all goods that are not permanent have been replaced by such goods as it will be advantageous to reproduce indefinitely in identical forms and quantities. It is supposed to be determined solely by the permanent resources and the vague concept of a given supply of free capital, and to be independent of the particular forms in which capital actually exists. The equilibrium in which we are interested here is not an equilibrium that is already embodied in the things, but an equilibrium between the different activities of creating new goods, as determined by the goods which happen to exist at the outset. This concept is in fact no less realistic than that of a stationary equilibrium: since in order to arrive at a stationary equilibrium it would be necessary to pass through a phase in which the changes required to bring about a stationary state were still going on but their results were correctly foreseen.
This extension of the technique of equilibrium analysis which we propose to use here is still somewhat unfamiliar. It may therefore be useful, before we proceed to develop it further, to throw some added light on to the difference between the two concepts of equilibrium involved, by a short discussion of a closely related ambiguity in the use of the concept of dynamics in economics. This concept has indeed two altogether different meanings according as it is used in contrast to the concept of a stationary state or in contrast to the wider concept of equilibrium. When it is used in contrast to equilibrium analysis in general, it refers to an explanation of the economic process as it proceeds in – time, an explanation in terms of causation which must necessarily be treated as a chain of historical sequences. What we find here is not mutual interdependence between all phenomena but a unilateral dependence of the succeeding event on the preceding one. This kind of causal explanation of the process in time is of course the ultimate goal of all economic analysis, and equilibrium analysis is significant only in so far as it is preparatory to this main task. But between the concept of a stationary state and the problems of dynamics in this sense, there is an intermediate field through which we have to pass in order to go from one to the other. The term dynamics is sometimes also applied to this intermediate field, but here it refers to phenomena which still come within the scope of equilibrium analysis in the wider sense. All that the use of the term dynamics means here is that we do not postulate the existence of a stationary state; but it says nothing about the method which we use. 
 It is at least questionable whether the introduction of the terms statics and dynamics into economics (by J. S. Mill following A. Comte’s similar division of sociology) which is responsible for this confusion was beneficial. It seems to me that the only relevant distinction is between two methods, that of logical analysis of the different plans existing at one moment (“equilibrium analysis”) and that of causal analysis of a process in time. For this distinction the terms statics and dynamics seem altogether inappropriate, and it would probably be better if they were to disappear entirely from economics.
Now as I have tried to show elsewhere,  the general idea of equilibrium, of which the stationary state is merely a particular instance, refers to a certain type of relationship between the plans of different members of a society. It refers, that is, to the case where these plans are fully adjusted to one another, so that it is possible for all of them to be carried out because the plans of anyone member are based on the expectation of such actions on the part of the other members as are contained in the plans which those others are making at the same time. This is clearly the case where people know exactly what is going to happen for the reason that the same operations have been repeated time after time over a very long period. But the concept as such can also be applied to situations which are not stationary and where the same correspondence between plans prevails, not because people just continue to do what they have been doing in the past, but because they correctly foresee what changes will occur in the actions of others. This sort of fictitious state of equilibrium which (irrespective of whether there is any reason to believe that it will actually come about) can be conceived to comprise any sort of planned change, is indispensable if we want to apply the technique of equilibrium analysis at all to phenomena which are ex definitione absent in a stationary state. It is in this sphere alone that we can usefully discuss equilibrium relations extending over time, and in which consequently the pure theory of capital mainly falls, and the latter might almost be said to be identical with the whole of this intermediate field between the theory of the stationary state and economic dynamics proper. Yet this field has never been systematically explored.
 In an article on “Economics and Knowledge”, Economica, N.S., vol. iv, no. 13 (February 1937), and, in a rather unsatisfactory form, much earlier, in an article on “Das intertemporale Gleichgewichts-system der Preise und die Bewegungen des Geldwertes”; Weltwirtschaftliches Archiv, vol. 28 (1928).
It must be admitted, however, that there is partial justification for this in the fact that there is no reason to believe that any general equilibrium could ever be fully realised except after all changes in data had ceased (that is as a stationary state was reached), and that in consequence there is no obvious need for the explanation of the pose economic process as it proceeds in time to make use of such a hypothetical construction. It may be thought that this is more than we require or can expect from the equilibrium method: and that all we need do is to explain how temporary equilibria are formed on particular markets. This would involve explaining how, once the more mobile elements have been adjusted, a temporary state of rest is arrived at which will last until the slower changes in the more permanent part of the productive equipment are effected. We could then describe the conditions that will prevail when all these changes have been completed (that is the hypothetical state which would ultimately be reached where all the data would remain unchanged). After all, decisions about what and how to produce are being made and revised periodically at fairly short intervals, and it may seem that period analysis which makes use of the concept of partial short-term equilibrium at each stage takes account of this essential fact and will come as near a realistic explanation of events as we can reasonably hope for from this type of approach.
There arises serious doubt, however, whether the concept of short-period equilibrium, if applied to an economic system as a whole,  has any definite meaning. The question is whether there is any such interval of comparative rest between the moment when the more mobile factors have been adjusted and the time when the more rigid elements of the structure can be effectively adjusted. 
 I.e. as distinct from a particular industry in which special conditions make it possible to mark off a particular period as being short compared with another.
 Without some such assumption the use of the term equilibrium has no justification whatsoever. It becomes a completely empty concept, saying no more than that at any moment some factors have had time to adjust themselves and others have not had time, and this would be true of any position. The distinction between short- and long-period equilibrium does of course make sense where, as in all the examples used by Marshall, it is applied to a particular industry, because in many cases the changes inside that industry will take place in two stages separated by an interval of time. But to make the later of these changes (i.e. the changes in the durable equipment) possible, changes must be going on during the interval in some other industry. And while we may be justified in disregarding these changes elsewhere so long as we are only concerned with the situation in the first industry, this becomes clearly illegitimate when we speak about the system as a whole. The use of the concept of a general short-term equilibrium in recent monetary analysis seems to me highly questionable.
This presupposes that with respect to the time it takes to adapt them to new circumstances, the existing means of production can be divided into two distinct groups. It assumes that the times it takes to alter different items of the stock of existing resources by using them up and producing new ones (which will depend on the durability of the individual resources and the time it takes to produce them) are not dispersed over a fairly continuous range but are definitely clustered about two most frequent points with a more or less empty interval between them. It seems highly doubtful whether this assumption is in any way justified by the facts, and for this investigation at any rate I prefer to adopt what seems to me the more plausible assumption that these periods are spread fairly continuously and without any marked break (though not necessarily evenly) over the whole range of periods in question. 
 The distinction between the “short” and the “long” period equilibrium is the most general case of a distinction which arises in several interconnected fields. The distinctions between “prime” and “supplementary” cost, between “circulating” and “fixed” capital, and between “current” production and (gross) “investment”, all belong to the same category and raise the same difficulties. They ought all to be treated, and win be so treated here, as limiting cases of a continuous range of variations, and not as representative of a particularly characteristic or most frequent type. No attempt will be made here to draw any arbitrary line of division in place of a frank recognition that these forms of the phenomena in question shade imperceptibly into each other.
Yet, quite apart from this particular point, it is apparent that this use of the equilibrium concept fails to take advantage of some of the most valuable aids that are to be derived from this powerful intellectual tool. So long as the pretence is kept up that the idea of equilibrium must refer to something which we can observe in the real world, or which at least can be shown to arise spontaneously under certain conditions, there is probably no other way of dealing with these problems. But I am inclined to helieve that these attempts to give the equilibrium concept a realistic interpretation (the legitimacy of which remains in any case somewhat doubtful) have deprived us of an at least equally important use, which the concept will serve if we frankly recognise its purely fictitious character. It has often been emphasised that the concept of a state of equilibrium is independent of any possibility of showing how such a state will ever come about. The reason why this assertion has had so little effect on the use which is actually made of the equilibrium concept is probably that those who made it did not properly show how such a fictitious construction could help to explain real events. In fact when it came to any concrete use of the concept, either it was defined as timeless,  or else resort was had to the stationary state.
 In which case, as I have tried to show in the article already referred to, it is meaningless.
In the sphere of capital theory, as we have seen, the construction of a stationary state is particularly useless because the main problem, that of investment, arises just because people intend to do in the future something different from what they analysis are doing in the present. The investment itself they may intend continuously to repeat as the instruments created need replacement. But the results of investment, whether they be direct services for consumption or (as in the majority of cases) an aid to further production, will necessarily alter the things that need to be done and can be done in the future. To postulate a self-repeating stationary state is to abstract from the very phenomena that we want to study. Nevertheless there is a very significant sense in which the concept of equilibrium can be of great use if it is made to include plans for action varying at successive moments of time. The essential problem remains that of whether the plans of different individuals will tally and will accordingly all stand a chance of being successful, or whether the present situation carries the seed of inevitable disappointment to some, which will make it necessary for them to change their plans. We must not lose sight of the reason why we are interested in the analysis of a particular economic system at a given moment of time: our purpose is to be able to proceed from a diagnosis of the existing state of affairs to a prognosis of what is likely to happen in the future. Now, if we want to predict at all, it must be on the basis of the plans which entrepreneurs are likely to make in the light of their present knowledge, and of an analysis of the factors which in the course of time will determine whether they will be able to carry out these plans or whether they will have to alter them. It seems natural to begin by constructing, as an intellectual tool, a fictitious state under which these plans are in complete correspondence without, however, asking whether this state will ever, or can ever, come about. For it is only by contrast with this imaginary state, which serves as a kind of foil, that we are able to predict what will happen if entrepreneurs attempt to carry out any given set of plans. The description of the equilibrium position in this sense is at the same time a description of the mutual interdependence of the decisions of different entrepreneurs.
The direction in which an entrepreneur will have to revise his plans will depend on the direction in which events prove to differ from his expectations. The statement of the conditions under which individual plans will be compatible is therefore implicitly a statement of what will happen if they are not compatible. 
 This is strictly true only if we are thinking of a single deviation of a particular element in a situation which is otherwise in equilibrium, that is on the assumption that all other expectations are confirmed. If more than one element turns out to be different from what was expected, the relation is no longer so simple.
It will be seen that this extension of the equilibrium concept provides the bridge from equilibrium analysis to the explanation in terms of causal sequences, since it is designed to elucidate the factors which will compel entrepreneurs to change their plans and to help us to understand the way in which their plans will have to be changed. In fact this use of the equilibrium concept is not fundamentally different from the comparison between the prospective and retrospective (or ex ante and ex post) views of a particular situation, as used by the younger Swedish economists,  since the ex post situation can be derived from the ex ante only by reference to the degree of correspondence or non-correspondence between individual intentions. The state of equilibrium as here understood is a state of complete compatibility of ex ante plans, where in consequence (unless changes occur in the external data about which economic theory cannot say anything in any case) the ex post situation is identical with the ex ante. It serves as a kind of standard case by reference to which we are able to judge what to expect in any concrete situation.
 Cf. G. Myrdal, Monetary Equilibrium (London, 1939). p. 46.
The significance of these abstract considerations will be clearer if we illustrate them by reference to the problems of investment. The problems of capital or of investment, as here defined, are problems connected with the activity of making provision in the present for the more or less distant future. The relevant future with which we are concerned is, however, somewhat more extensive than the periods for which the individual consciously invests at a particular date. His plans at any moment will be based on the expectations of a certain future state of the market which will allow him to dispose of his products at a certain price; and beyond this his interest will not extend. But the objective “state of the market” on which he counts is largely the result of the present decisions of other people. In order that he may succeed in disposing of his products as he expected, it will be necessary for others to have made preparations which will enable them to use just those products at the prices at which he expected to sell them. In other words, the state of the market at the time for which he plans will largely depend on what others have decided at the same time as he made his plans. This is so not only, or even mainly, because the incomes which these other people will have to spend will depend on what they have produced, but also because what instruments and materials they will need will depend on what plans for production they have embarked upon. This means that although every individual will be guided only by (more or less well-founded) expectations of particular prices, he will actually be performing part of a larger process of the rest of which he knows little; and his success or failure will depend on whether what he does fits in with the other parts of that larger process which are undertaken or contemplated at the same time by other people. What he performs will in the majority of cases be no more than a single step in a long chain of successive operations. His action may be removed from ultimate consumption by many stages, and its success will be dependent at each stage, not so much on the final demand as on the presence or absence of complementary instruments in proportionate quantities, and on there being people willing to use them in subsequent stages of production. All these successive operations have to be viewed as parts of one integral process, each of them having chances of success only by reason of its position in the whole.
In any system with extensive division of labour (particularly where it is of the “vertical” type and many successive operations by different entrepreneurs are dependent chainwise upon one another) every decision to produce one thing rather than another will be dependent for its success on other things being produced in appropriate quantities. Thus we have definite quantitative relationships between the required output of different kinds of goods, which (owing to the technological character of the process) will usually be of a more rigid character in the case of producers’ goods than in the case of consumers’ goods. Almost any quantitative combination of different kinds of consumers’ goods will be capable of use in some way or other. But the limits within which the proportions between the quantities of the different kinds of producers’ goods may vary are much narrower. There are definite proportionalities, quantitative relations, between the different parts of the structure of production, which must be preserved if some of these parts are not to become completely useless.
It is clearly possible to study the quantitative relations between the different parts of the real structure of production that will result from current plans, independently of the question of the forces which will secure, or fail to secure, the actual bringing about of such a correspondence. In any given situation there will be one (and in most instances only one) way in which the plans of the various entrepreneurs can be made to harmonise with one another and with the preferences of the consumers. The use of the equilibrium method here then means constructing an imaginary state in which the plans of the different people (entrepreneurs and consumers generally) are so adjusted to one another that each individual will be able to sell or buy exactly those quantities of commodities which he has been planning to sell or buy. What will exist will of course still be only the separate plans of different individuals which are connected only by the fact that the quantities of goods which are expected to pass at different dates out of and into the possession of the various individuals exactly match. Any particular person need know neither who will take his products nor who will provide him with what he expects to get – he will only have expectations about what the anonymous  group called the market will provide and take; nor need he know much about the way in which the goods which pass into his hands have been produced, or about the way in which the goods he has produced will be used.
 Cf. F. Machlup, “Why Bother with Methodology ?” Economica, N.S., vol. iii, no. 9 (February 1936), pp. 43 et seq.
Nevertheless coincident expectations about the quantities and qualities of goods which will pass from one person’s possession into another’s will in effect co-ordinate all these different plans into one single plan, although this “plan” will not exist in anyone mind. It can only be constructed, and it is in fact often convenient to adopt the practice, which has been followed by many economists, of proceeding for a time on the assumption that the actions of the different individuals are directed by somebody in accordance with a single plan.  In the nature of the case this fictitious assumption can be only provisional, and must later be abandoned in favour of the assumption of separate but perfectly matched plans of the different individuals – that is: of competitive equilibrium in the sense outlined above.
 This device was used most systematically by F. Wieser, first in his Natural Value and later in his Social Economics, where he prefixed his theory of the social economy with an elaborate theory of what he called a “simple economy”, i.e. a centrally directed economy. More recently Professor Pigou (in his Economics of Stationary States, 1935) has once again made use of Robinson Crusoe for the same purpose. It is interesting to note that Marshall, when he comes to discuss investment, finds it also convenient first to discuss it “by watching the action of a person who neither buys what he wants nor sells what he makes, but works on his own behalf” (Principles, 7th ed. Book V, chap. iv/1).
It is inevitable that opinion will be divided about the usefulness of such an admittedly fictitious construction as the concept of equilibrium here employed. And there is no way of demonstrating its usefulness other than by applying it to a particular problem. It is however, important that no misunderstanding should arise about the justification that is claimed for it. Its justification is not that it allows us to explain why real conditions should ever in any degree approximate towards a state of equilibrium, but that observation shows that they do to some extent  so approximate, and that the functioning of the existing economic system will depend on the degree to which it approaches such a condition. The explanation of why things ever should, and under what conditions and to what extent they ever can, be expected to approximate to it, requires a different technique, that of the causal explanation of events proceeding in time. But the fact that it is probably impossible to formulate any conditions under which such a state would ever be fully realised does not destroy its value as an intellectual tool. On the contrary it seems to be a weakness of the traditional use of the concept of equilibrium that it has been confined to cases where some specious “reality” could be claimed for it. In order to derive full advantage from this technique we must abandon every pretence that it possesses reality, in the sense that we can state the conditions under which a particular state of equilibrium would come about. Its function is simply to serve as a guide to the analysis of concrete situations, showing what their relations would be under “ideal” conditions, and so helping us to discover causes of impending changes not yet contemplated by any of the individuals concerned.
 It should be remembered that nearly the whole of economic science is based on the empirical observation that prices “tend” to correspond to costs of production, and that it was this observation which led to the construction of a hypothetical state in which this “tendency” was fully realised. A good deal of confusion has been caused in this connection by the vagueness of the term tendency. A given phenomenon may tend to (approximate towards) a certain magnitude if in a great number of cases it may be expected to be fairly near that magnitude, even if there is no reason to expect that it will ever actually reach it, however long the time allowed for the adjustment. In this sense “tendency” does not mean, as it is usually understood to mean, a movement towards a certain magnitude but merely the probability that the variable under consideration will be near this magnitude. The ideal state in which all the variables would be at the magnitude to which they tend to approximate in this sense is a state which one could not expect ever to be reached.
THE SIGNIFICANCE OF ANALYSIS IN REAL TERMS
The analysis of the relations between the production plans of different entrepreneurs must necessarily proceed in what is known as “real terms”. If we assume – as we must if we are to investigate the compatibility of the different plans – that the entrepreneurs make definite and detailed plans for fairly long periods, there is indeed little room for money in the picture at all, except as “mere counters” which stand for definite quantities of particular commodities. In fact, so long as we assume entrepreneurs to decide every detail in advance in the certain expectation that they will be able to adhere to all their plans, the need for holding money almost vanishes. For in the actual world money is largely held because the decision as to when to buy or to pay for something is deliberately postponed; and this is contrary to our assumptions. But even to the extent to which money would still be held under these conditions (because of the discontinuity of transactions and the cost or inconvenience of investing it for the short periods until it was needed) it would cease to playa significant role. For money would enter into the plans, not in the quasi-independent character of command over things in general (that is as something which confers on its holders the chance of taking advantage of unforeseen opportunities), but only as a transitory item representing the definite quantities of commodities for the purchase of which the particular amounts of money are held.
The existence of such a condition in which all that would be relevant to the plans made by the public would be the concrete quantities of goods which they expected to get in exchange for money, but not the quantities of money itself, is often silently assumed, usually illegitimately. On our assumptions such a condition would actually exist. We should dum analysis would therefore gain nothing if we were to introduce quantities of money as separate magnitudes into this type of analysis in place of the quantities of commodities for which the money would stand. Such a procedure would merely entail a very considerable and unnecessary complication of the argument. Particular money prices stand in a determinate relationship to quantities of goods which will be produced or sold at these prices only on the assumption that all other prices are given. In principle any particular money price for a commodity may correspond to the production or sale of any quantity of that commodity, according as the prices of other commodities vary. There are no such definite relationships between prices in money terms and quantity of goods, as there are between the real ratios of exchange and such quantities. The introduction of money at this stage would therefore merely have the effect of introducing an additional variable which is irrelevant for our purpose and would make it more difficult to see the relationships, between quantities of commodities and real ratios of exchange, in which we are here interested.
Economists have often felt the need for some such analysis in real terms, and in fact a considerable part of classical economics, explicitly or implicitly, makes use of this idea. Its exact meaning and significance have, however, scarcely ever been made clear. Recently the concept of “neutral money”  has been widely used in this connection. While this has at least the advantage of drawing attention to the existence of a problem, it is in itself, of course, nothing more than a new name for an old problem and does not provide us with a solution. It makes it clear that we cannot, as has often been done, treat money as non-existent so long as its value remains stable, and that it is erroneous to assume that if its value remains stable it exerts no influence on the formation of prices. Neither do the special constructions which certain economists have used to meet this difficulty really solve the problem. The best known of these is Walras’ “numeraire”. According to definition the “numeraire”, which may be any of the commodities, serves merely as a unit of account; but it is not actually used as a medium of exchange and consequently there will be no additional demand for it to hold it as money. All that the introduction of this concept does is to solve the difficulty of the mathematical economist in expressing all the different ratios of exchange in one common unit. It contributes nothing to the explanation of how the triangular and multi-angular exchange transactions, which are necessary to bring about equilibrium, can be effected without the use of one or more media of exchange which are demanded and held merely for the purpose of exchanging them against other commodities.
 The present author must plead guilty of some responsibility for the popularity of this concept and even for the incautious way in which attempts have occasionally been made to use it as a practical ideal of monetary policy. But while for this second purpose it is clearly not of much help, it still appears to me as a useful concept to describe a real theoretical problem: the conditions under which it would be conceivable that in a monetary economy prices would behave as they are supposed to behave in equilibrium analysis.
The crux of the matter is that where analysis aims directly at a causal explanation of the economic process as it proceeds in time, the use of the conception of a money-less exchange economy is misplaced. It is self-contradictory to discuss a process which admittedly could not take place without money, and at the same time to assume that money is absent or has no effect. In the case of our ideal position of equilibrium, which we construct as a guide to interpretation, and in which all parts are assumed to be perfectly matched, the case is different. Here analysis in real terms is not only in place, but is almost essential. Since at each point money is in the strictest sense only an intermediary between definite quantities of certain goods, all the essential relations in this system are relations between goods (rates of substitution between certain quantities of goods determined by the total quantities of these goods). Or, in other words, it will be true of this system – what has sometimes been asserted to be true in the real world – that the total supply of goods and the total demand for goods must be identical. (This so-called “Law of Markets” of J. B. Say is indeed one of the first formulations of the modern concept of equilibrium.)
It would, however, be a mistake to believe that, since these relationships will exist only in a purely fictitious state of equilibrium, it is mere waste of time to work it out. The fact that in the real world relations between money prices, and not real ratios of exchange, directly determine human action, does not make these real ratios uninteresting. Relations between money prices in themselves tell us little, unless we know what prices are appropriate to the existing real structure of productive equipment, or what price relationships are required to enable people to go on with the plans they have made. Nor is it sufficient, as is sometimes supposed, to know whether the prices of finished products exceed or fall short of a given money cost of production as represented by the prices of a particular combination of productive resources. Whether this or some other combination of resources will be used in the manufacture of the product will itself depend on prices. The costs of production of a particular good do not therefore move in exact conformity with prices of any particular collection of resources, but are also affected by changes in the technique of production made profitable by changes in the relative prices of the different resources.
In the real world production is so obviously dependent in the first instance on concrete money prices that the suggestion that it “ultimately”· depends on some real relationships which lie behind these money prices, is undoubtedly, as the whole history of economics shows, in sharp contrast with the conclusions that are first suggested by experience. It is therefore necessary to justify our procedure somewhat more fully than by merely repeating the mostly metaphorical phrases which are commonly used in its defence. That there are “underlying real forces which tend to reassert themselves, although they may be temporarily hidden by the monetary surface”, or that the real relationships which “ultimately” determine the relations between prices show a certain resiliency and are more permanent than the temporary distortion caused by money, or that the real determinants are more fundamental or basic in the sense that they win be restored when the monetary disturbances have disappeared, is all approximately true; but it hardly proves or explains the significance of these real factors.
It is undeniably true that in the absence of continuous progressive monetary changes, and with given tastes and a given distribution of incomes, the relations between the prices of different commodities will be uniquely determined by the quantities of these goods in existence. But this is not the whole story, because these quantities can themselves be changed by monetary influences. The decisive fact, however, is that the effect on prices of these changes in quantities brought about by monetary influences will be in exactly the opposite direction from the direct effect on prices of these same monetary changes. We may suppose, for instance, that, at the point where a net addition to the total money stream makes its first impact on the commodity markets, there will result an increase first of the prices and then the output of the commodities affected.
The effect of this increase in output will be that, as soon as the additions to the money stream cease, the prices of these commodities will fall relatively to the prices of all other commodities and will reach a lower level than prevailed before the monetary change. Monetary changes have this effect in common with all merely temporary changes which are not recognised as such. But they have it in a particularly high degree. This is so not only because by their very nature they cannot continue indefinitely, but more especially because a change in the volume of the money stream which takes place at one point of the economic system works round and is bound to cause further changes in all other prices. Monetary changes are therefore in a peculiar sense self-reversing and the position created by them is inherently unstable. For sooner or later any deviation from the equilibrium position – as determined by the real quantities – will cause a swing of the pendulum in the opposite direction. 
 Cf. in this connection the discussion in my Monetary Nationalism and International Stability, pp. 31 et seq.
Unfortunately the significance of these real factors cannot be fully demonstrated without a systematic analysis of the operation of the monetary factors which we propose largely to disregard in this study. But an illustration may be given by referring briefly to the main problem in connection with which this question is continually cropping up. This problem relates to the possible differences between the prospective profitability of a given investment according to whether the investor has to use real resources which he owns or borrows, or whether he can obtain those resources by borrowing money for the purpose. There can be no doubt that under certain circumstances the possibility of borrowing money will make investments profitable which would never appear attractive if the investors could only use such resources as they owned or could borrow in natura. The reason for this, now very familiar, is that the amounts of money offered on the loan market are capable of changing quite independently of the supply of real resources available for investment purposes. 
 Much confusion has been caused in this connection by the assump. tion sometimes made that there could be a real capital market without money on which there would be some determinate in natura rate of interest. In fact there would not and could not be one rate of interest without money, and the effect of the limitation placed on the possible amount of waiting by the scarcity of the stock of non-permanent resources would make itself felt exclusively via the changes in relative prices of the different kinds of commodities.
In point of fact, monetary changes facilitate investments and cause resources to be put to uses which are not in accordance with a state of equilibrium between the demand for and the supply of real resources. This does not, of course, mean to say that monetary factors may not change the composition of the real quantities in existence. On the contrary. By affecting the uses to which the available resources are put, they will inevitably bring about a change in the real structure of production. But the point is that this new, changed, material structure of production will require for its maintenance a new set of price-relationships, namely those which the initial monetary change temporarily created or led people to expect, but which this monetary change cannot perpetuate. Most additions to, or deductions from, the money stream will not stay where they have first appeared; they have the inherent tendency to reverse  the changes in price-relationships which they have caused. But the significance of the further changes in relative prices which will be brought about by the monetary change will have to be judged in relation to the price structure appropriate to the changed organisation of production.
 Of course this does not mean that the position which would have existed without the monetary disturbance will – or even can – ever be fully restored. The losses and redistributions of income caused by the misdirection of production will naturally have a permanent effect – but an effect in a direction opposite to the impact effect of the monetary change.
We cannot judge the effect of any change in money prices without a knowledge of the system of prices which is appropriate to the existing structure of production. There is thus a task which is logically prior to the study of the monetary mechanism : the task of analysing the principle on which particular systems of quantities of goods and particular systems of prices (or real ratios of exchange) are coordinated. This is what the so-called analysis in real terms attempts. Like equilibrium analysis in general its aim is not to give a direct explanation of any real phenomena, but to analyse in isolation a set of relationships which are relevant for the explanation of actual events. In other words: there are conditions of stability of the economic system which not only can be described more simply if we neglect the monetary factor, but which, although they can be changed by monetary influences, exist independently of them. These conditions are at any moment determined by the technical structure of the material equipment in existence and by the tastes of the people.
In the particular case we have to study the amount of abstraction involved in disregarding money is especially great. We are setting out to investigate problems of capital and at the same time the possibility of lending and borrowing money. Now this is of course a phenomenon with which the problems of capital and interest are so closely connected in real life that it may appear futile to talk about capital at all without taking money-lending into account. But that this appears so only goes to show that in our minds the terms capital and interest are so closely connected with monetary phenomena that it would perhaps have been better if they had never been used by economists in connection with the real phenomena which, though somehow connected with the monetary phenomena, would exist even in a money-less capitalist society. It has, however, become so firmly established a usage to apply the same terms to the underlying real phenomena as were first applied to their monetary manifestations that it would be difficult, at this stage, to introduce new terms for them.
In one respect, indeed, this tradition has recently been seriously challenged. In his last work  Mr. Keynes has placed very strong emphasis on the desirability of confining the term “rate of interest” to the rate at which money can be borrowed. And quite apart from the fact that his use of the term would be more in conformity with its meaning in ordinary life, there can be no doubt that it is only in this form that interest appears as a price actually quoted in the market and directly entering into the calculations of entrepreneurs. The real or commodity rates of interest, which have played such a prominent role in traditional economic theory, are in comparison merely secondary or constructed magnitudes which, besides, vary according to the commodity in terms of which we compute them. These considerations probably make it advisable, in all investigations dealing with monetary phenomena, to restrict the term interest, as Mr. Keynes suggests, to the money rate, and to introduce some other term for the “real rates”. This objection, however, does not apply, or at least not as strongly, so long as we confine ourselves to the real aspects of the problem. Here the danger of confusion does not arise, and it has seemed on the whole expedient to use the term interest here in the sense in which it has become customary to use it in pure economics, that is as referring to real percentage rates of return.
 Cf. Keynes, 1936.
In what sense and to what extent it is justified under the assumptions made here to speak of one uniform rate of return can be shown only as the investigation proceeds. But in order that the term rate of interest which we propose to use in this connection should not mislead, it is necessary at this stage to explain at least a little more fully what will be designated by this term. It has already been mentioned that the rate of interest in these conditions is not a price of any particular thing. It is an element in the relations between the various prices of different commodities, a ratio between the prices of the factors of production and the expected prices of their products, which stands in a certain relationship to the time interval between the purchase of the factors and the sale of the product. The problem of the rate of interest in the sense in which it will be discussed in this book is therefore the problem why there is such a difference between the prices of the factors and the prices of the products and what determines the size of this difference. It would perhaps be more correct if we referred to this difference between cost and prices as profits rather than interest. But as it has become customary – particularly since Böhm-Bawerk, to whom this particular statement of the problem of interest is due – to refer to this difference in equilibrium analysis as the rate of interest, and as the term rate of profit is now generally reserved for such “abnormal” differences as will arise only under dynamic conditions, it will probably cause less confusion if in equilibrium analysis we retain this established although somewhat unfortunate term.
That these differences between costs and prices which pervade – and are expressed in – the whole system of relative prices will in equilibrium stand in a definite relationship to each other which can be expressed, in some sense, as a uniform time rate is, strictly speaking, a fact which should not be assumed at the beginning of this investigation but forms one of its results. But as in this respect we are only going over ground which has often been covered in a similar manner, there can be no harm in anticipating this result, with which every reader will be familiar, and in occasionally speaking of a rate of interest in this sense before we have shown why there should be a tendency to adjust all the various price differences to a common standard.
If this methodological discussion is not to grow to disproportionate length, we must leave it with this rather cursory discussion of the relation between analysis in real terms and analysis in monetary terms. A more systematic and exhaustive treatment would be impossible without explicit consideration of the role money does actually play; and this is just what we want to avoid here. What has been said is merely an attempt to indicate certain consequences which follow from the treatment of the problem of capital as part of general equilibrium analysis.
There is only one more point which should be stressed in conclusion of this discussion. The fact that almost this entire volume is devoted to the equilibrium or “real” aspects of our problem must not be taken to mean that we attach excessive importance to these aspects. The idea is rather to emphasise the width of the gulf which separates this exercise in economic logic from any attempt directly to explain the processes of the real world. It would have been easy enough to expand this exposition with occasional disquisitions about the significance of the considerations advanced here in a scheme of causal explanation of the real economic process. The author has on the whole tried to resist this temptation as far as possible and to keep strictly within the limits explained in this and the preceding chapter. The application of the results of equilibrium analysis to the real world means a transition to an altogether different plane of argument and requires a very careful re-statement of the assumptions on which it proceeds. It is impossible to do this by occasional remarks without running the risk of illegitimately turning analytic propositions into assertions about causation. It seems much better frankly to recognise the limits of what can be achieved with the method here employed, and to reserve the task of applying the results to causal explanation for separate investigation. Some suggestions concerning the treatment of these further problems which will arise in a money economy will be found in Part IV of the present study.
THE RELATION OF THIS STUDY TO THE CURRENT THEORIES OF CAPITAL
As already remarked above, the explanation of interest will not be the sole or central purpose of the present study, as was the case with most of the similar investigations in the past. The explanation of interest will be an incident although necessary result of an attempt to analyse the forces which determine the use made of the productive resources. Our main task is not to explain a particular form of income, or the price of a particular factor of production, but to display the connection between the supply of the various kinds of productive resources, the demand for real income at different dates, and the technique of production that will be chosen. Most of the analytical tools which we shall have to use were, however, created in the past in the search for the explanation of interest. And it is natural that the theories of interest which have contributed most to the elucidation of the problems which we are going to study should be those which stressed the “productivity of capital” and were in consequence based on an analysis of the material structure of production.
What follows is in some respects no more than an attempt towards a systematic development and elaboration of the fundamental ideas underlying the theory of interest of W. S. Jevons, E. v. Böhm-Bawerk, and Knut Wicksell. If, in the course of its reformulation, parts of their theory are changed beyond recognition, this does not alter the fact that their work contains, though perhaps in a somewhat crude and excessively simplified form, nearly all the basic ideas on which the following exposition builds. Jevons’ work, although he was not given time to formulate it in a way in which it was readily intelligible, contained the essential elements of the more fully developed theory.  Böhm-Bawerk in many respects simply developed the ideas propounded by Jevons and made them intelligible to wider circles by elaborating them: but at the same time he gave the impetus to a movement away from what seems to me to be the more fruitful approach on Jevonian lines.  His effective, although I think mistaken, critique of the earlier productivity theories of interest had the effect of causing later development to centre increasingly round the “psychological” or “time-preference” element in his theory rather than the productivity element.
 Apart from the relevant chapters of the Theory of Political Economy (1st ed. 1871, 4th ed. 1911, particularly chap. vii), his unfinished Principles of Economics (1905) and the additional chapter to this work, printed as Appendix II to the fourth edition of the Theory, should be consulted.
 Cf. Kapital und Kapitalzins, published in two parts (1886 and 1889) and translated under the titles Capital and Interest (1890) and The Positive Theory of Capital (1891), which is still by far the most elaborate and comprehensive discussion of the problems of capital. The third and fourth German editions contain a good deal of important additional material in the form of further elucidations, replies to criticisms, and discussions of later theories. This material has not so far been available in English, although a new complete translation by Mr. Hugh Gaitskell is in preparation. This additional material is particularly important for its treatment of durable goods, which were unduly neglected in the first edition available in English – a fact which has given rise to much misunderstanding of Böhm-Bawerk’s doctrines among English-speaking economists. Some remarks on this subject will be found in Böhm-Bawerk’s Recent Literature on Interest (1903). A number of smaller essays in German dealing with particular problems in this field were collected after Böhm-Bawerk’s death by Professor F. X. Weiss under the title Kleinere Abhandlungen über Kapital und Zins (1926).
In the first instance Professor Irving Fisher, without in any way denying the importance of the productivity element, has, in a number of earlier works,  stressed the psychological factor so much more than the productivity factor that he was at least understood to attach more importance to the former.
 Particularly The Nature of Capital and Income (1906) and The Rate of Interest (1907), which in spite of the new exposition of the same set of problems which the author has given us since, will still be found useful for their more detailed treatment of particular problems.
More recently he has, however, given us, in the most systematic work on the subject which we possess, a formally unimpugnable exposition of the theory approach of interest.  It is a work with which every student of the subject must be familiar. But because of a different distribution of emphasis, and in particular his concentration on interest rather than on the methods of production, Professor Fisher’s work hardly touches on a good deal of what is treated as important in the present study.
 The Theory of Interest (1930).
The time-preference element has, however, been stressed much more exclusively by another author who has developed this side of the Böhm-Bawerkian analysis, namely Professor F. A. Fetter. His writings on the subject, which, apart from the relevant sections of his two textbooks,  include numerous articles in various periodicals, will be found very suggestive, and in spite of certain obvious differences, have a close affinity to some of the leading ideas of the investigation that follows. This is particularly true of the idea of the rate of interest as an element pervading the whole price structure.
 Principles of Economics (1907) and Economic Principles (1915).
In addition to this branch there is a second branch which also springs from the Jevons-Böhm-Bawerk stem. This is represented almost exclusively by K. Wicksell  …
 Wicksell first treated these problems in extenso in 1893 in his Wert, Kapital und Rente (now reprinted as no. 15 of the Series of Reprints of Scarce Tracts in Economics and Political Science, 1933). He later incorporated the main argument, with some improvements, in his Vorlesungen (vol. 1, 1913, and earlier in Swedish), now available in English under the title Lectures on Political Economy (vol. i, 1934). Certain important points are also contained in his Finanztheoretische Untersuchungen (1896) and Geldzins und Güterpreise (1898; English edition, Interest and Prices, 1936).
… and his pupils (particularly Professor G. Akerman  and Professor E. Lindahl ), who, with the help of certain ideas derived from L. Walras,  have systematically developed the productivity approach. It is in the shape into which this type of theory has been fashioned by Wicksell that it provides the most useful basis for the present study. Wicksell has seen nearly all the important problems left open by Böhm-Bawerk; and in fact, after one has oneself found the solution of a difficulty arising when one abandons Böhm-Bawerk’s simplifications, one frequently finds it implied or even explicitly stated in some inconspicuous remark in Wicksell’s work. It must, however, be admitted that Wicksell did not give an adequate answer to Böhm-Bawerk’s objections to an explanation of interest which was based mainly on the marginal productivity principle, and it will be one of the tasks of the present investigation to show why the factors affecting the supply of new capital ought to be relegated to a secondary place, at least in an analysis which is not primarily concerned with the conditions of long-term stationary equilibrium.
 G. Akerman, Realkapital und Kapitalzins, 2 Parts (1923 and 1924).
 Most of Professor Lindahl’s contributions are now available in English in a volume Studies in the Theory of Money and Capital (1939). See, however, also the Bibliography at the end of the present volume.
 Elements d’economie politique pure (1847-77, 4th ed. 1900), section 5. Probably Walras deserves more than this mention in passing, although his direct influence in this field was not very considerable, and even Wicksell, who in most other respects had absorbed so much of Walras’ teaching, fully comprehended his theory of interest only at a late stage. See his Lectures, vol. i, p. 226, particularly the footnote – which incidentally is also interesting for the distinction between what is now known as the ex ante and ex post rate of interest (or the anticipated and the actual rate of interest, as Wicksell calls them).
Besides the three authors who were responsible for the main steps in the development of the marginal productivity analysis of interest, there are several others who should be mentioned as having helped to shape those doctrines. In the first place there are the ingenious predecessors of this school, H. von Thünen and, more especially, John Rae.  The latter’s New Principles on the Subject of Political Economy (1834)  contains some acute analyses of points of detail still not to be found elsewhere, and has had considerable effect through its influence on J. S. Mill. With regard to more recent contributions this study owes much to Professor F. W. Taussig’s Wages and Capital (1897),  especially for the more felicitous terminology which he has introduced in certain connections. Among the great mass of other pre-war monographs Professor A. Landry’s L’Interet du capital (1904) deserves special mention. And finally, L. von Mises, although his published work deals mainly with the more complex problems that only arise beyond the point at which this study ends, has suggested some of the angles from which the more abstract problem is approached in this book.
 Perhaps Ricardo should also be mentioned here, even if he could scarcely have been aware of all the implications of his theory which Dr. Victor Edelberg has so ingeniously worked out (1933). There can be no doubt, however, that Wicksell was to a large extent inspired by Ricardo.
 Republished in a rearranged form with an Introduction by Professor C. W. Mixter under the title The Sociological Theory of Capital (1905).
 Reprinted as no. 13 of the Series of Reprints of Scarce Tracts in Economics and Political Science (London, 1932). Cf. also Professor Taussig’s articles: “Capital, Interest and Diminishing Returns”, Quarterly Journal of Economics, vol. xxii/3 (1908), and “Outlines of a Theory of Wages”, American Economic Association Quarterly, Third Series, vol. xi, 1910.
For reasons already explained in the preface, this general acknowledgement of the main obligations will have to stand in place of more detailed references throughout the text. Particularly in the case of Jevons, Böhm-Bawerk, and Wicksell, the constant references which an adequate acknowledgement of the real indebtedness would require have been omitted. But the same applies to most other authors, and the comparatively few references that are given are intended not so much as an acknowledgement of an obligation as an illustration, by similarity or contrast, of the point under discussion. 
 While references in the text to contemporary discussions of these problems have been kept to a minimum, a fairly full list of contributions in this field during the past ten or twenty years which have come to the knowledge of the author has been added as an appendix to this volume.
The general line of thought which this investigation follows has of late often been described as the “Austrian” theory of capital. In view of the varied nationality of the founders of this theory, and in view of the fact that the men who are commonly regarded as the leaders of the “Austrian School” of economics are by no means in agreement on it,  it is questionable whether this designation is appropriate.
 Neither C. Menger nor F. von Wieser, nor – to mention only one name from the later generation – Professor Schumpeter accepted Böhm-Bawerk’s views.
But, in spite of Jevons and the other English and American adherents, it cannot be denied that these views have in recent times intruded into Anglo-American discussions as a sort of alien element. And perhaps it will assist the reader if an attempt is made to sketch the main points on which the approach followed here differs from the traditional Anglo-American treatment of the same problems, and particularly, it seems, from the views of those authors who were mainly influenced by the teachings of Alfred Marshall. This may be conveniently done by setting out the differences point for point in tabular form. In order to make them quite clear one may also be permitted to state the points that are emphasised by the two lines of thought in a rather trenchant and even exaggerated form. It is of course not claimed that the description of either of these approaches in its extreme form does justice to the real position. Indeed one of the tasks of the following pages will be to amalgamate the two lines of thought into a coherent whole. All that is claimed is that in the “Anglo-American” treatment the aspects stressed by the second or “Austrian” approach have in more recent times  been unduly neglected.
 It may perhaps be mentioned here that the classical English economists since Ricardo, and particularly J. S. Mill (the latter probably partly under the influence of J. Rae), were in this sense much more “Austrian” than their successors.
In the following list of propositions the first of each pair is. intended to represent the traditional or “Anglo-American” point of view, while the second gives the contrasting “Austrian” view on the same problem:
1A. Stress is laid exclusively on the role of fixed capital as if capital consisted only of very durable goods.
1B. Stress is laid on the role of circulating capital which arises out of the duration of the process of production, because this brings out particularly clearly some of the characteristics of all capital. 
 Cf. Wicksell, Lectures, vol. i, p. 186: “Strictly speaking only short-period capital (in other words circulating capital) can be regarded as capital proper”.
2A. The term capital goods is reserved to durable goods which are treated as needing replacement only discontinuously or periodically. 
 A consequence of this concept of capital which we cannot discuss here further is the concept of gross investment as referring to the aggregate production of durable goods, and the belief that this magnitude is of special significance. It is, of course, closely connected with the distinction between the short and the long period, which, as was shown before, has little meaning for the economic system as a whole.
2B. Non-permanence is regarded as the characteristic attribute of all capital goods, and the emphasis is accordingly laid on the need for continuous reproduction of all capital. 
 Cf. J. S. Mill, Principles, I/v/7, ed. Ashley, p. 74: “Capital is kept in existence from age to age not by preservation but by perpetual reproduction; every part of it is used and destroyed, but those who destroy it are employed meanwhile in producing more”; and Wicksell, Lectures, vol. i, p. 203: “The accumulation of capital is itself, even under stationary conditions, a necessary element in the problem of production and exchange”.
3A. The supply of capital goods is assumed to be given for the comparatively short run.
3B. It is assumed that the stock of capital goods is being constantly used up and reproduced.
4A. The relevant time factor which we need to consider in order to be able to understand the effect of changes in the rate of interest on the value of a particular capital good is assumed to be its individual durability.
4B. It is not the individual durability of a particular good but the time that will elapse before the final services to which it contributes will mature that is regarded as the decisive factor. That is, it is not the attributes of the individual good but its position in the whole time structure of production that is regarded as relevant.
5A. The technique employed in production is supposed to be unalterably determined by the given state of technological knowledge.
5B. Which of the many known technological methods of production will be employed is assumed to be determined by the supply of capital available at each moment.
6A. The need for more capital is assumed to arise mainly out of a lateral expansion of production, i.e. a mere duplication of equipment of the kind already in existence.
6B. Additional capital is assumed to be needed for making changes possible in the technique of production (i.e. in the way in which individual resources are used), and to lead to longitudinal changes in the structure of production.
7A. The change that will initiate additions to the stock of capital is sought in an increase in absolute demand, i.e. in the total money expenditure on consumers’ goods.
7B. Changes in the stock of capital are supposed to be determined by changes in the relative demand for consumers’ and producers’ goods respectively.
8A. In order to make a lateral expansion of production appear possible, the existence of unemployed resources of all kinds is postulated.
8B. In order to stress the changes in productive technique connected with an increase of capital, the existence of full employment is usually postulated.
9A. The demand for capital goods is assumed to vary in the same direction as the demand for consumers’ goods but in an exaggerated degree.
9B. The demand for capital goods is assumed to vary in the opposite direction from the demand for consumers’ goods.
10A. The analysis is carried out in monetary terms, and a change in demand is assumed to mean a corresponding change in the size of the total money stream.
10B. The analysis is carried out in “real” terms, and an increase in demand somewhere must therefore necessarily mean a corresponding decrease in demand somewhere else. The last four propositions relate to problems which are already outside the pure theory of capital which forms the subject of this book: they belong more properly to the main theory of monetary problems to which the present study is merely preparatory. But their inclusion in the list may help the reader to see the practical significance of these different ways of approach.
THE NATURE OF THE CAPITAL PROBLEM 
 An earlier version of this chapter has appeared in German as an article in the Zeitschrift für Nationalokonomie, 1937.
In the first stage of economic analysis it is usually assumed that all productive resources are given in an unalterable form. They are regarded as sources of services which will continue permanently to be available independently of any deliberate action to provide them. This is nearly enough true of were permanent free  human labour (which is not deliberately created from economic considerations) and perhaps also of the so-called “indestructible powers of the soil”. And the shorter the period of time which we regard as relevant, the wider will be the circle of resources which, for that period, can be regarded as definitively given.
 “Free” as opposed to slave labour.
This procedure is convenient as a first approach, because it allows us to analyse a number of important relationships without the complications which arise as soon as we take account of the fact that many of the existing resources may be of only limited durability. It is one of the devices which enables us to treat the economic process as “stationary” and to disregard all changes which occur in time. It will be assumed here that this part of economic theory has been fully worked out. 
 This part of pure economic theory is sometimes referred to as the theory of kapitallose Wirtschaft.
There can be no doubt that the picture obtained in this way corresponds very little with reality. If we look at the productive resources of any society at a given moment, we find that only a very small part of them (even apart from the human beings themselves) will continue indefinitely to render useful services without any deliberate provision for their upkeep or replacement; they cannot therefore be regarded as “permanent” or “self-perpetuating”.  This is not only true of practically all those bearers of useful services which have been created by man in the past, and can rarely, if ever, be expected to last indefinitely or to remain permanently useful. It applies  to all the capacities acquired by human beings through education and training, and also to the greater part of the natural resources. Some of the latter, such as the fertility of the soil, can only be expected to endure permanently if we take care to preserve them. Others, such as mineral deposits, are inevitably exhausted by their use and cannot possibly render the same services for ever.
 Cf. Wicksell, Lectures, vol. i, p. 150.
 At least from a social point of view: the individual can hardly use up his knowledge and training before he ceases to be interested in its usefulness (except possibly by overwork). It comes to the same thing when Professor Knight (1936, p. 641) makes the capital quality of human capacities dependent on their “presenting any possibility of deliberate over- or under-maintenance”.
This distinction between permanent and non-permanent (or “consumable” ) resources is of fundamental importance for the approach to the capital problem that will be followed in this study. It is not, however, a distinction which is in all cases unambiguous. The main point to be kept in mind is that what matters is not permanency in any absolute sense, but the opinion of the economic subject as to whether particular resources at his command will last throughout the period in which he is interested (be it his lifetime or a longer period), or whether they will be exhausted or used up earlier than this.
 It is unfortunate that in English the term “consumable” refers so definitely to final consumption that it can hardly be used, without danger of misunderstanding, to include things used up in production. If this were not the case it might be preferable to “non-permanent” in this connection, and Walras (Elements, edition definitive, 1926, p. 246) uses “consommables” as one of the essential characteristics of the “capitaux proprement dits”, although he adds that they must also be “produits”.
In this sense his own person may be regarded as a permanent resource if he is not interested in what will happen after his death. It could of course be argued with some plausibility that, strictly speaking, all resources are non-permanent. But this would only mean that our distinction is merely a distinction of degree, and would by no means deprive it of its significance. What may be regarded as an even more fundamental basis for the distinction is the fact that the future services of some resources cannot be anticipated, as they will continue to give the same services in the future no matter how they are used in the present, while the present use of the services of other resources decreases the amount of such services which will be available in the future. This is not affected by the objection that no rigid line can be drawn between permanent and non-permanent resources. The underlying fact, and in a sense the most general aspect of the phenomenon under consideration, is the irreversability of time which puts the future services of certain resources beyond our reach in the present and so makes it impossible to anticipate their use, whereas the present services of those resources can as a rule be postponed. 
 An alternate concept which is probably better but clumsier than the concept of permanent resources is the concept of non-anticipatable returns, i.e. those final services which would still be available at any future date even if up to that date their consumption had been kept at the maximum level attainable at every moment without regard to the future. In order that a resource may be permanent in this sense it is not essential for it to be indestructible in a physical sense. All that is necessary is that it should be expected to be useful, not in consequence of being kept in that state at a sacrifice, but because no present advantage would arise from destroying its future usefulness.
It is evident that certain resources may have to be treated as permanent in this sense, because their expected future services cannot be sacrificed in order to increase satisfaction in an earlier period, but will have to be treated as non-permanent in the other sense because their services, once they become available, are non-recurrent. In cases like these the significance of what we have called the irreversability of time, i.e. the fact that we can postpone but not anticipate the use of certain resources, becomes particularly clear. The range of time during which any force of nature can be turned to useful purposes has, as it were, always a definite beginning but frequently no necessary end.
It should be clear from these considerations that it would be equally misleading to gloss over this distinction by treating all resources as non-permanent, as it is to treat all resources as permanent. It is the existence of differences between the resources in this respect, and not the existence of extreme types, that is relevant.
Whether there are no really permanent resources in this sense, as is sometimes suggested, is open to doubt. When we remember that the relevant fact is not indestructibility in a physical sense but the lack of any inducement to destroy, there can be little doubt that, apart from the human beings themselves, not only a number of forces such as water-power but also quite a considerable part of the productive power of the soil must be regarded as permanent. A great deal of land (pastures) retains its fertility, not because provision is made to keep it fertile, nor despite its being used for current needs, but just because it is being used each year in such a way as to give the greatest possible service in that year. If anyone wanted seriously to deny (as I understand Mr. Kaldor does) that there are such things as permanent resources, he would have to assert that it is conceivable that by raising the rate of consumption (or rate of output) to the highest level obtainable in the near future, the productive capacity of the more distant future could be reduced to zero.
The non-permanent nature of all “wasting assets” creates a problem which is not dealt with in the theory of timeless production. These assets cannot be directly used to contribute to the output of the time when they have ceased to exist. In so far as their existence does help to maintain output permanently above the level at which it could be kept with the help of the permanent resources alone, it must do so in an indirect manner. lf the fact that we have command over resources which remain useful only for a limited period of time did not help us to use the services of the permanent resources more effectively, it would be quite impossible to keep our income permanently above the level where it would stay if these non-permanent resources had never been available. We might stretch their use over a longer period of time, but ultimately we should inevitably exhaust them and should then have to be content with what services the permanent resources could render by themselves. It is this problem of why the existence of a stock of non-permanent resources enables us to maintain production permanently at a higher level than would be possible without them, which is the peculiar problem connected with what we call capital. 
 Cf. F. von Wieser, Natural Value (1893), p. 124: “On the other hand, it is a matter for wonder to find that the perishable powers of the soil, and all the movable means of production, raw materials, auxiliary materials, implements, tools, machinery, buildings, and other productive apparatus and plant, which are consumed, quickly or slowly, in the service of production, are sources of permanent returns, – returns which are constantly renewed, although the first factors of their production may have been long before used up. This brings us face to face with one of the most important and difficult problems of economic theory; with the question, namely, how we are to explain the fact that capital yields a net return.” In a footnote to this passage Wieser adds that “in what follows I understand by the term capital the perishable or (with the extended meaning explained in the text) the movable means of production”. A similar passage occurs in K. Wicksell, Wert, Kapital und Rente (1893), p. 73: “Dass nun aber die verbrauchbaren Güter, d. h. Güter, die in einer begrenzten Reihe von Verbrauchsakten ihren ganzen Nutzgehalt zu erschöpfen scheinen, dennoch ‘kapitalistisch’ angewandt werden können, so dass ihr ganzer Wert dem Eigentümer aufbewahrt bleibt und sie ihm dennoch Einkommen schaffen, diese scheinbar paradoxe Erscheinung, dieses perpetuum mobile des Volkswirtschaftsmechanismus bildet, wie früher gesagt, den eigentlichen Kern der Kapitaltheorie”.
The term capital itself, in so far as it is required to describe a particular part of the productive resources, will accordingly be used here to designate the aggregate of those non-permanent resources which can be used only in this indirect manner to contribute to the permanent maintenance of the income at a particular level. 
 This definition of capital is far less revolutionary than may at first appear. The first move in this direction was made by Ricardo when he decided to include “permanent improvements” with the “original and indestructible powers of the soil” because, “when once made, the return obtained will ever after be wholly of the nature of rent and will be subject to all the variations of rent”, differing in this from the “perishable improvements” which “require to be constantly renewed and therefore do not obtain for the landlord any permanent addition to his rent” (Principles, chap. xviii; Works, ed. McCulloch, p. 158, note); and when for similar reasons he excluded from rent proper “the compensation given for the mine or quarry” (Principles, chap. ii ; Works, p. 35). Cf. also J. S. Mill, Principles of Political Economy, ed. Ashley, Part I, chap. vi, p. 93: “But as the capital … cannot be withdrawn, its productivity is thenceforth indissolubly blended with that arising from the original qualities of the soil, and the remuneration for the use of it thenceforth depends, not upon the laws which govern the roturns to labour and capital, but on those which govern the recompense for natural agents”.
The definition adopted here is essentially the same as Wicksell’s, and Wicksell in turn seems to be indebted to Wieser for it. Cf. Wicksell, Wert, Kapital und Rente (1892), pp. 72-73: “Der wichtigste volks-wirtschaftliche Unterschied zwischen dem Grund und Boden und den produzierten Sachgütern scheint nämlich darin zu liegen, dass ersterer seine Nutzleistungen nur successive in einer vorher bestimmten und unveränderlichen zeitlichen Reihenfolge, dafür aber auch in einer unendlichen Reihenfolge abgiebt, wogegen die produzierten Güter nur eine endliche Summe von Nutzleistungen, diese aber beinahe in beliebiger Reihenfolge abgeben kennen … ; man kann sagen, dass die Produktions-werkzeuge um so mehr einen kapitalistischen Charakter (im engeren Sinne) bewahren, als sie nach Belieben verwendet werden kennen, z. B. die Maschinen, welche in schnelleren oder langsameren Lauf versetzt werden oder auch still stehen kennen, dabei aber keine Abnützung erfahren. Andere Vorrichtungen im Gegenteil, z. B. gewisse Bodenmeliorationen, sind, einmal angestellt, so ganz und gar mit dem Grund und Boden verwachsen, dass sie den erwähnten Charakter verlieren, d. h. nunmehr eigentlich Rentengüter, nicht mehr Kapitalgüter im engeren Sinne sind”; and on p. 79 of the same book: “Für die folgenden Untersuchungen scheint es mir jedoch am zweckmässigsten, die verschiedenen Kapitalien einfach nach ihrer Dauerbarkeit zu rangieren. Die eminent dauerbaren Güter, seien sie selbst Produkte oder, wie der fungfräuliche Boden, reine Naturalgüter, und mögen sie ihre Nutzleistungen spontan oder nur unter Zusetzung von menschlicher Arbeit abgeben, nenne ich im folgenden Rentengüter. Die verbrauchbaren oder schnell abgenützten Produktions- oder Konsumtionsgüter, solange letztere sich noch nicht in den Händen der Konsumenten befinden, nenne ieh Kapitalgüter oder Kapitalien im engeren Sinne.” (Italics not in the original.) See also ibid. pp. 93-94, and the same author’s Finanztheoretische Untersuchungen (1896), pp. 28 et seq., Geldzins und Güterpreise, pp. 117-118, and Lectures, vol. i, pp. 186-187; and the passage from Wieser’s Natural Value, quoted in the previous footnote, to which Wicksell refers in this connection.
It should be specially noted, however, that the important point is not whether it is expedient to use the term capital for this purpose; on this reasonable people may differ, although they will scarcely find it worth while to argue about it. The essential point is that the existence of this kind of resources creates an important and peculiar problem which is different from and, we believe, of much greater significance than the problem of the kind of time-discount which would exist even in a society where all resources are permanent. It is this problem to which the present study is mainly devoted and it is for this purpose that we find it most useful to employ the term capital in the sense indicated. And at a later stage we shall attempt to show that it was this problem which originally gave rise to the conception of capital in ordinary business usage. 
 See below, Chapter VII, p. 89.
It is, however, not necessary that this definition should be interpreted and applied with too rigid adherence to the literal sense of the terms “permanent” and” non-permanent”. In fact, what in a particular situation will have to be regarded as capital will to some extent depend on the context in which we use the concept. Perhaps it would even be better to attempt a general definition of capital only in the negative form of saying that the only things which never will have to be regarded as capital are the really permanent resources in the strictest sense of the term. In the case of all resources which are not strictly permanent the decision whether we have to treat them as capital or not will depend on whether or not they can be used up (or used up more quickly) during the period of time relevant for the problem in question.  Certain kinds of goods, particularly those which are commonly referred to as circulating capital, can be used up even during very short periods and will therefore have to be treated as capital in practically all contexts (this is the reason why circulating capital shows the peculiar characteristics of capital in a particularly high degree, while with respect to others the problem of the gradual exhaustion and the need for replacement will arise only when the period relevant to the problem in hand is much longer.)
 Cf. Wicksell, Lectures, vol. i, p. 186: “If, therefore, our analysis is only applicable within a fairly short period, then, strictly speaking, only short-period capital (in other words, circulating capital) can be regarded as capital proper”.
The final justification of any particular definition of the term capital can of course come only from its use as a tool of analysis. At this stage we do not propose to dwell much longer on the definition; we want only specially to emphasise its comprehensive character. Included under this term are not only the man-made productive equipment in so far as it is not expected to remain useful for ever but also natural resources in so far as they are “wasting assets”, and all consumers’ goods existing at the moment in so far as they are non-permanent sources of final income. But although this concept is related to the familiar concept of the “produced means of production”, it is not identical with it. It does not necessarily include all the produced means of production, since it is at least conceivable, although not very probable, that some of the produced means of production may be expected, once they have been created, to remain useful for ever;  and in this case they would not be capital in our definition of the term. 
 It may sound curious that we reckon, e.g., houses as capital only if and in so far as they are non-permanent. There can of course be no doubt that we would be better off if houses, once they are built, lasted for ever, and the fact that they need replacement is clearly a disadvantage. Yet it is this fact, that we have to replace them by something if we want to keep our income stream at a given level, and that we can use the amortisation quotas earned on houses in the same way as the amortisation quotas earned on any other capital good to replace these capital goods by whatever form of new investment appears most advantageous at the moment, which gives all the capital goods a common attribute, that of being the source of the “fund” which makes current investment possible – and necessary.
 Cf. Wicksell, Lectures, vol. i, p. 186: “Such improvements to the land often leave a permanent residual benefit. This happens, for example, in the case of major blasting operations to secure water in mountain regions, the building of roads, protective afforestation, etc. Thus new qualities which, once acquired, the land retains for all posterity, cannot be distinguished either physically or economically from the original powers of the soil; in the future they are to be regarded not as capital, but as land. . . . It may be further pointed out that nearly all the long-term capital investments, nearly all so-called fixed capital (houses, buildings, durable machinery, etc.) are, economically speaking, on the border-line between capital in the strict sense, and land.”
Nor is it identical with the wider concept which identifies capital with the total stock of wealth,  for it excludes the sources of really permanent services. 
 This latter definition of the concept of capital, which has been given wide currency through the writings of Professor Irving Fisher (and has also been used by Walras, although he singles out the consumable and produced capital as “capitaux proprement dits”), has the advantage of great logical clearness and of avoiding any distinction based on mere differences of degree. Its disadvantage for our purposes, however, is not only that it uses the term capital for a magnitude for which there are other terms (particularly “wealth”) available and in general use, but, what is more important, that it severs all connection with the special problems which have given rise to the concept of capital and which, since they need close study, are most conveniently treated under that general heading. Where the definition of capital is entirely subservient to an explanation of interest, as has been the case with most of the traditional discussions, this definition may be usefully adopted. But where the peculiar problems arising out of production in time are the main subject of discussion, it would be a pity to have to invent a new term so long as the term capital is available for the purpose.
 It may be useful to summarise the relation between the various capital concepts and the categories of resources which they include, with the help of a schematic table:
Our definition includes the categories b and d, while the more familiar definition of capital as produced means of production (or as “augmentable resources” in the revised form in which Mr. Kaldor has recently revived this definition – see 1937, p. 219) would include c and d. Professor Fisher’s definition (and the wider capital concept used by Walras) would of course include all four groups, while Walras’ narrower concept (the “capitaux proprement dits”) would include d only. We shall later see (see Chapter VII, p. 90, below) that while of the things which exist at any moment only those belonging to the groups band d serve as capital, their existence enables us to “invest” by creating things belonging either to the group d or to the group c, the latter ceasing thereby, however, to be capital in our sense.
The first question to which we have to turn, then, is how the existence of a stock of non-permanent resources enables us to maintain our income at a raised level for an indefinite period. The answer to this question is especially significant because it also explains why a great part of those resources, namely those which are “produced means of production”, ever came into existence. The answer is of course that the non-permanent resources provide an income stream for a limited period; and that in consequence we are in a position to postpone the return from some of the current services of the permanent resources without reducing our consumption below the level at which it can be permanently kept. We are thus able to take advantage of the celebrated productivity of the “round-about methods” of production  which have been the cause of so much misunderstanding. In other words, the existence of non-permanent resources makes it at the same time possible and necessary to “invest” some of the current productive services, that is to use them in such a way that they will not yield consumable services until a later date than they might otherwise have done, but will then yield a larger amount of such services than they would have done at the earlier date. It is only because of this that the provision of an additional amount of services for a limited period in the future puts us in a position to raise for all time the return which we may hope to obtain from the meagre supply of really permanent resources.
 This was not, of course, a new discovery of Böhm-Bawerk’s, although he invented the term “round-about methods of production” and brought out the role of time in production much more clearly than anyone before him. The essential point was understood fairly well by most of the classical writers and it was particularly well formulated by N. W. Senior in the third of his “Four Elementary Propositions of the Science of Political Econonmy”, viz. “That the powers of Labour, and the other instruments which produce wealth, may be indefinitely increased by using their Products as a means of further Production” (Political Economy, 1836, 8vo edition, p. 26).
But why should the more time-consuming methods of production yield a greater return? Ever since the time when it was first put forward this proposition has been the source of unending confusion, and it has given rise to so many misunderstandings that, however much space one were prepared to give to the subject, it would scarcely be possible to deal with all of them. Nor is it certain that there is any single explanation that will necessarily fit all cases. There is, however, one general fact which makes it appear probable that it will always be possible to increase the amount of final services which can be obtained from given resources if more time is, allowed to elapse between the time when the resources are applied and the time when their final product emerges. And this is of course all that is required.
This general fact is, briefly, that there will almost always exist potential but unused resources which could be made to yield a useful return, but only after some time and not immediately; and that the exploitation of such resources will usually require that other resources, which could yield a return immediately or in the near future, have to be used in order to make these other resources yield any return at all. This simple fact fully suffices to explain why there will nearly always be possibilities of increasing the output obtained from the available resources by investing some of them for longer periods.
It has never been asserted that every investment for a longer period will necessarily yield a larger product, although the critics have sometimes attacked the theory on these grounds. All that is important is that, so long as there are possibilities of increasing the period by investing for a longer period, only such prolongations of investment periods will be chosen as will actually give a greater product. The rather obvious reasons for this we shall consider later.
The explanation of the greater productivity of some time-consuming methods of production is closely connected with another question, namely: In what sense can it be said that the services of the permanent resources are given in a definite quantity? It is necessary to begin by considering the meaning of this assumption in detail. Of all the potential sources of satisfaction of human needs only comparatively few can be used directly. There are always an infinite number of natural forces which are capable of being turned to some human use, and which are in this sense potential or latent resources. And of those actually used only a part will be scarce, and will therefore be counted as valuable assets on the use of which the satisfaction of human needs depends. What part of the total of the potential resources will actually be used, and what part will be scarce, will always depend on concrete circumstances and will vary with these circumstances. When we speak here of a constant stream of services from the permanent resources being available, what we have in mind is always the totality of such potential resources, irrespective of what part of them is actually being used at any particular moment or what part of them has become scarce.  It is only in this sense that they can be regarded as an extra-economic “datum”. What part of them will be used and what part will be scarce and therefore have value, will depend on human decisions which it is the task of economists to explain. In general, the reason why resources which are capable of being turned to some useful purpose are not actually so used is that they would have to be combined with other resources which are more urgently needed elsewhere. So long as these other, complementary, resources cannot be spared because the total quantity of them available is required for purposes where they will yield a greater return than they would yield in co-operation with the potential resource in question, this resource will remain unused, or latent, and will not become scarce.
 The fact that with changing circumstances the amount of the services accruing from permanent resources that are scarce will vary instead of remaining constant, creates a serious difficulty which it will hardly be possible to take into account at all stages of the exposition without making it unduly complicated. During the earlier part of the analysis, which is devoted to a mere description of the technological interrelations, we shall in any case have to take it as a given fact that only a certain part of the potential resources is scarce and must therefore be taken into account, the free resources being neglected.
This general phenomenon of complementarity between different productive resources becomes significant for our particular problem if the potential resources, which might be used to produce useful services, will not yield these services until some time after they have been combined with other resources which can be used to produce such services immediately. It is of course by no means a priori necessary that the product which will be obtained in this time-consuming way shall be greater than (i.e. that it will be preferable to) that which would have been obtained from the direct use of the complementary resources. All that we can say in general is that men will take the trouble to use the services of additional resources only if, as a result, the product not only becomes different but is also preferable to what it would otherwise have been.  But that it is technically possible does not mean that it will always be done. So long as the other resources, such as human labour, which are required to utilise the potential resources, but which can also be used for the satisfaction of immediate current wants, cannot be spared, these potential resources will remain unused. And so long as they remain unused we can hardly regard them as separate resources since they naturally remain free goods. Even when it has become possible to divert some of the other resources from the service of current needs to utilise the latent resources, it will be some time before the latter grow scarce. And although the increased product will be due to the fact that use is now being made of resources which it was impossible to use previously, we need pay no attention to this fact so long as these additional resources are free and not “economic” or scarce factors.
 It is, perhaps, not unnecessary today to stress the fact that the goal of economic activity is not to use the greatest possible quantity of resources, but to produce the maximum of satisfaction, and that these two things are not identical.
From among the different latent resources some can be made to give a return after a short interval and some only after a long one. Under otherwise equal circumstances, those which yield a return sooner will be taken up first.  But among the other circumstances which must be equal in order for this to be true is the size of the return which may be obtained by applying to them a certain amount of resources withdrawn from use for current consumption. If in a particular instance the return obtainable from investing resources for a certain period is considerably greater than twice the return obtainable from investing the same resources for half that period, the longer investment will evidently be taken up first. But the detailed consideration of this question must be reserved for a later chapter.
 Suppose that there are two latent resources of which one can be made, by the application of labour, to yield a return after one year, and the other, with the same application, will yield a return of equal size only after two years. Then if we assume that both of these investments will be repeated continuously, the return obtained during any stated period will always be greater from the first resource than from the second.
For the moment another point is more important. At first what count as “investments” are only the services of those resources which might also have given an immediate return. So long as they are the only resources which can be used at all, or fully used, it is only their investment which leads to a reduction of current output. But as, in consequence of these earlier investments, the formerly latent resources, with which the other resources are combined, become first effective and then gradually scarce, these too will begin to count, because some consumption will be dependent on the particular use to which they are put. If they are then combined with some other potential resources which will not yield a consumable product until still later, this will represent an additional postponement of consumption. This fact that, as investment proceeds, more and more of those natural forces which before were only potential resources are utilised and gradually drawn into the circle of scarce goods, and have in turn themselves to be counted as investments, is of great importance for the understanding of the whole process. 
 Cf. C. Menger, Grundsätze der Volkswirtschaftslehre (1st ed. 1871; reprinted London, 1934), pp. 129-130.
We shall have to return to this question of the general function of capital in the process of production in the next chapter but one. But before we can do so it is necessary to consider somewhat more concretely the various forms in which time enters into the process of production. This will have to be the task of the next chapter.
THE DURATION OF THE PROCESS OF PRODUCTION AND THE DURABILITY OF GOODS: SOME DEFINITIONS
We must now begin to consider the process of investment in its more concrete manifestations. There are two main ways in which the productivity of investment shows itself, and although the distinction is not fundamental, it is well to keep them clearly apart. The difference is due to the fact that the investment of any group of services of the permanent resources can be combined into one single “process” of production in two ways. In the one case it is the duration of the actual process of production where the time factor enters, and in the other case it is the durability of the product (or of the non-permanent resources used in production). In the first case the essential point is that resources will have to be applied some time before, and frequently throughout a considerable period before, any consumable services are produced. In the second case the essential point is that it will not be worth while to make the investment unless it results in a stream of useful services that will continue to accrue for some time to come.
The “flow of services from the permanent resources” which becomes available during any given period of time (or, in the limiting case, at a moment of time) we shall henceforth describe, following a suggestion of Dr. Hawtrey’s,  as the amount of pure input of that period (or moment). It should be specially noted, however, that the term input will not be confined to that part of the services of the permanent resources which is “invested” in the ordinary sense of the word but will comprise all the pure input that is used at all, including that part which is used to produce current output. The term investment (or “an investment”) will correspondingly describe the act of applying a unit of input in any process of production. Since, as we shall see later, it is not only pure input in the strict sense of services of the permanent resources which can be invested for varying periods, but also the services of non-permanent resources, we shall also adopt Dr. Hawtrey’ s expression mixed input when we want to emphasise that the latter kind of input is also included.
 Hawtrey, 1937, p. 15: “We may apply to the operation of the original factors of production in any interval of time the convenient term ‘input’. When we treat capital as a factor of production, we shall call the operations of the original factors in conjunction with capital ‘mixed input’, and if we want to distinguish input in the fonner sense from mixed input we shall call it ‘pure input’.” It will be noticed that the definition of pure input used in the text, in conformity with the terminology used throughout this book, substitutes “permanent resources” for Dr. Hawtrey’s “original factors”. The term “input” itself had been devised earlier, probably by Professor R. Frisch in connection with the distinction to which the next footnote refers.
The term output will be used to describe the stream of final services to the consumer.
The distinction mentioned in the opening paragraph of this chapter amounts to a difference in the way in which aggregates of input are connected with aggregates of output. In the real world the two cases are of course never completely separate. But it is useful to construct ideal limiting cases which show their peculiarities in the purest form. The first case is best represented if we conceive of a continuous application of input through a period of time, leading to an output all of which matures at a moment of time at the end of the period. This has been described as the “continuous input – point output” case.  The second case is ideally represented if we imagine a durable good which is produced at a moment of time and then renders services continuously over a period of time. This case has correspondingly been described as the “point input – continuous output” case. As we proceed we shall have to devote a good deal of attention to the relationships ideally described by those extreme cases.
 These terms were, I believe, first suggested by Professor Ragnar Frisch.
Before going on to discuss the way in which the productivity of investment manifests itself in each of these two cases, we must mention another way of describing them which seems to bring out the relevant peculiarities even more clearly, and to have the further advantage of stating them in a way which is more in accordance with the concepts used in other branches of economic theory. The first case, where the actual process of production takes time, may be regarded as a case where the final services wanted at a particular moment of time give rise to a joint demand for factors to be applied at different moments of time. The second case, where it is a question of the durability of the good, may be regarded as a case where the investment made at a moment of time gives rise to a joint supply of services over a period of time. 
 Cf. Wicksell, Lectures, vol. i, p. 260: “The annual uses [of a capital good] successively following one another constitute a kind of joint supply (to use Marshall’s terminology)”.
As has already been observed, it is almost impossible in real life to find cases where time elapses between the application of the factors and the enjoyment of the results in only one of these ways. It is only under comparatively primitive conditions that we can conceive of cases which will correspond perfectly to either of the two extreme types. If we could assume that fireworks were made without the use of any durable tools or machinery, the work of our hands over a period of time would lead to a display lasting little more than a moment; and this would therefore correspond fairly closely to the “continuous input – point output” case. If on the other hand we cut a straight branch from a tree and used it for years as a walking-stick, we should have a fairly good instance of the “point input – continuous output” case. But as a rule the two sets of relationships are so completely intertwined that it is extremely difficult to disentangle them. This means that in practice we shall almost always have to deal with cases of a stream of consumption services accruing at successive moments of time which are the joint product of a process which also involves a joint demand for factors applied at successive moments of time. But this only makes it so much the more necessary to try to isolate conceptually the way in which each of these two kinds of investment increases the output from given resources.
For purposes of theoretical analysis it is necessary to isolate the connection between individual units of input and individual units of output, and at the same time we have to recognise that in real life production is as a rule continuous. Experience has shown that it is sometimes difficult to keep the right balance between these two aspects. It is important always to remember that the continuity of the actual process of production is due not so much to the fact that the same sort of process is continuously being repeated as to the fact that most of the investments which form part of the continuous process are made with a view to obtaining a stream of returns over a period of time, and that almost all returns are due not to a particular investment but to a range of investments over a period of time. 
 Even Professor Knight, who in general is so thoroughly unsympathetic towards the whole investment period analysis, appears to admit the necessity of distinguishing between the investment periods of various units of input – although on other occasions he appears to deny the possibility of such a distinction. This, at all events, is the only meaning I can make of the following passage from one of his more recent articles (1936, p. 447): “Because the process of investment must be spread over time and because, in general, there is more or less disinvestment in connection with the yield of any particular capital good, it is necessary to recognise the separate periods of investment, from zero to infinity, of each infinitesimal increment of capital invested in any source or capital good. Only in this way can different investments be made at the same rate and the maximum yield obtained on the whole capital, which is possible under the given economic conditions.”
It requires a high degree of abstraction to arrive at the idea of separate individual processes which consist of separate and clearly distinguishable  inputs and outputs and which will yield a continuous stream of output only if they are continually repeated in an unchanged manner. But it is only by means of such abstraction that it is possible to isolate the relevant relationships between the different parts of the continuous process.
 That is, distinguishable on technological grounds.
The fact that a series of successive investments is usually combined in order to produce any kind of commodity, and that the same productive operation results in a stream of final services extending over a period of time, is the source of another serious difficulty when we come to consider the changes in investment periods involved in changes in technique. The fundamental fact with which we are concerned is the change in the periods for which particular units of input are invested, that is, in the interval between the application of a unit of input and the maturing of the quantity of output due to that input. This interval of time we shall describe as the investment period of that unit of input.
If the variation in the technique of production used always either affected the investment period of only one unit of the input or else affected the investment periods of all units in the same direction, there would be no problem, and we should be able to speak of changes in the “period of production”, or the “length of the process as a whole”, as a short way of referring to changes in the investment periods of the various factors used. In fact, however, most of the changes in productive technique are likely to involve changes in the investment periods of different units of input to a different degree and perhaps in different directions. This raises all kinds of difficulties which we shall have to consider later. In particular it makes it impossible to use the terms “changes in investment periods” and “changes in the length of the process” or “changes in the period of production” synonymously. It must indeed appear doubtful whether the second and third of these concepts, which necessarily refer to aggregates of investment periods, have any clear meaning. It is rather unfortunate that the time aspect of production should have been first introduced into theoretical analysis in this form, for it has led to much unnecessary confusion. But since the use of the expression “changes in the length of the process” is a convenient way of describing the type of change in a whole process where the changes in the investment periods are predominantly in one direction, there is probably something to be said for retaining it, provided it is used cautiously, until we are ready to give a fuller explanation of what is meant by one process as a whole involving more waiting than another.
It is necessary, however, to define somewhat more exactly than has been done so far what is meant by one process of production. In general the expression refers to the series of operations which lead up to the production of a particular kind of good. But this still leaves some ambiguity of meaning. The term process may be used to describe the whole series of operations which lead up to the production of a definite quantity of the product at a particular moment of time. The work of a potter who makes a clay vessel under primitive conditions, and bakes the clay on a fire made for the occasion, would represent a single process in this sense. The term may, however, also refer to the whole chain of continuously repeated operations which lead to a continuous output of pottery. If the term process is used without a qualifying adjective it will here refer to the first concept, and the “continuous process” will be described either as such or as a “line of production”. It will soon be seen that the concept of a process in the first sense involves a considerable amount of abstraction and in most instances does not refer to anything which can be clearly isolated in the real world.
The advantages of time-consuming processes of production are closely connected with the advantages of the division of labour. If the process which leads up to a certain final service to the consumer is broken up into a number of separate operations, it becomes possible to use certain capacities, materials, and tools which could not have been used if all the labour had to be applied in the way that would give the final result by the shortest possible route. But we must not be deceived here by a further ambiguity of the term “one process”, i.e. the reference to a particular technique of production which it occasionally implies. If the advantages of the division of labour consisted solely in the fact that the same series of operations as were previously performed by one man were divided between a number of men, and in consequence each of them became more efficient at his special task, the effect would probably be to shorten the duration of the process instead of lengthening it. But in many cases the division of the process which leads up to the satisfaction of any particular need will be a division among a greater number of co-operating factors, including some that before were not used at all. It will mean a change in the method of production, and in the materials, tools, and human capacities used. And the resulting product may be technically a very different one from what it was previously. If the needs which it serves, however, are the same, these needs will now be provided for by a different and longer process. 
 Cf., however, F. X. Weiss, 1921.
In fact the greater productivity of this longer process will frequently express itself in the circumstance that the new product serves the same ends more effectively, or perhaps serves other ends at the same time.
In all cases the greater output derived from the inputs which are now invested for longer periods will be due to their combination with forces which could not be put to any use during the shorter period. It may be that, as with natural processes of growth or fermentation, the natural forces will exercise their effect to the desired extent only if they are left to operate for a considerable period of time. Or the materials, tools, or accessories which it is advantageous to use may themselves be obtainable only as the result of a process which takes time. In short, the increase in output will always be due to a change in the method of production used, a technical improvement.
The term improvement, however, although it is quite appropriate and is frequently used in this connection, is yet another source of possible misunderstanding which should be guarded against from the outset. The term has often been understood, when used in this connection, to refer to inventions or discoveries. The argument would then seem to imply that technical progress in this sense, the advancement of knowledge, tends necessarily to increase the duration of the productive process. Against this it has been rightly argued that the discovery of new, hitherto unknown, ways of producing a thing will be just as likely – or perhaps even more likely – to shorten the duration of the process as to lengthen it. The considerations advanced above – and it is important to remember this throughout the discussion – have nothing to do with technological progress in this sense. On the contrary, they refer to changes under conditions where knowledge is stationary. All that is assumed is that at any moment there are known possible ways of using the available resources which would yield a greater return than those actually adopted, but would not yield this return until a later date, and for this reason are not actually used.
Among the wide range of possible methods of production known at anyone time there will be some which will yield their product after shorter periods of time and some which will not yield it until after longer periods. From among each group of methods involving the same “amount of waiting” – if we may make provisional use of this vague term – the one that will be chosen will be the one which yields the greatest return from a given investment of factors. But so long as there is any limitation on the “amount of waiting” for which people are prepared, processes that take more time will evidently not be adopted unless they yield a greater return than those that take less time.
We must return now to the relation of these changes in productive technique to the division of labour. The important thing about the transition to processes which on the whole involve investments for longer periods is that it is always undertaken in order to make use of additional forces of nature, and that in consequence it will as a rule involve a greater number of successive applications of distinct factors of production. Once this is understood it is easy to see how the transition to these processes will tend to give rise to the phenomenon which has been described as the vertical or successive division of labour as distinguished from the horizontal or simultaneous division of labour (i.e. the type of division which is due to the fact that people specialise in the production of different final products). The vertical or successive division of labour means that the process leading up to anyone product is broken up into distinct “stages”, or, that is, into a number of separate operations which in the modern organisation of society will be performed by different firms. But although the number of these separate “stages” in any one line of production will often tend to increase. with the lengthening of the process, we must not expect any strict proportionality between the time a process will take and the number of stages into which it will be divided.
In this sense, as the part of a complete process which is under the control of a particular firm, the concept of a “stage” of production is of little theoretical interest. The term, however, can be conveniently used for a grouping of the various kinds of capital goods according to their remoteness from ultimate consumption. In this sense it serves simply as a means of a further and very necessary subdivision beyond the usual rough division of goods into consumers’ goods and capital goods. It has the advantage that it takes better account of the fact that we have to deal with a continuous range of various kinds of goods, and that wherever we draw the line between consumers’ goods and capital goods by far the greater proportion of the goods existing at any moment will always fall into the latter category. Here the concept of stages and the distinction between earlier and later stages provides the distinction which will prove very necessary later on. When in the further course of this discussion the term stage is used, it will always be in this abstract sense and will not imply any reference to a division of the process between different firms or persons. 
 For a characteristic misunderstanding of the sense in which the concept of “stages of production” is used in theoretical analysis, cf. Ellis, 1935.
It would be a mistake, however, to concentrate too much attention on this vertical division of labour. It would be quite wrong to suppose that the lengthening of the investment periods will always mean an increase in the number of separate operations which follow each other in linear succession. What is no less important is that, in the course of the lengthening of the process, the stream of operations leading up to a given product will as a rule be split up into many branches and sub-branches. And it may be that, long before the first move is made to produce the actual material from which the product is to be made, work is being taken in hand to provide some auxiliary material or tool which will be needed later to convert the raw material into the final product. At each stage of the process from the raw material to the finished product the main stream will be joined by tributaries which in some cases may already have run through a much longer course than the main stream itself. But all these activities, many of which may be carried on at the same time at different places, have to be regarded as part of the same process, and have to be taken into account when we talk about its length. The series of operations which are required in order to provide the fuel or lubricant, and the tools or machines which are needed for turning the raw material into the finished product, are just as much part of the process of production of the good as the operations performed on the raw material.
There is some difficulty about introducing tools, and still more machinery, into the picture at this stage, because they raise the problem of durability which still awaits discussion. This is in fact one of the main instances of the way in which the problems of the duration of the process of production and the durability of goods are so inextricably mixed up. Now, although tools are usually durable, they are not always so. The moulds needed in many kinds of casting processes, or the dynamite used for blasting, probably have to be regarded as tools although they can only be used once. They are examples of how extensive preparations, resulting in elaborate tools or auxiliary materials, may be necessary in order to make use of certain laws of nature in the transformation of any raw material into a useful form. The significance of the circumstance that tools and machines are as a rule durable will be discussed presently along with the general problem connected with the durability of goods.
Before we can pass on to this problem it is necessary to return for a moment to the difficulty of talking about changes in the length of the process of production. It will probably be fairly obvious by now that as the complete processes of production with which we have to deal become indefinite meaning creasingly complex it becomes more and more difficult, and may in some cases be impossible, to say in any general way which of several alternative processes under consideration is as a whole the shortest or the longest. The total length of time which elapses between the very beginning of the process and the completion of the product may be shorter in one process than in another, and yet by far the greater part of the input used may be applied very early in the first process and very late in the second process. Which of these two processes is to be regarded as the longer ? It is impossible to answer this question at the present stage, and there is in fact no general answer to it. It is only mentioned at this point in order to warn the reader against any attempt to provide himself with an answer by introducing some concept of an “average period” of production. Such a concept, as we shall see, is not only unnecessary but is also highly misleading.
For our present purposes we do not need to know whether a whole process as such is longer or shorter than another. The only points that are relevant here are, first, the periods for which units of input are invested, and, secondly, the fact that they will not be invested for longer periods unless the return due to them will be greater in consequence.  The reader will save himself a good deal of trouble if he accustoms himself from the start to the habit of regarding the periods for which particular units of input are invested as the primary factor and of regarding the length of the whole process of production from which a particular product results as only a secondary phenomenon.
 An exception to this rule, which may be disregarded at this stage, occurs when in the course of a thorough change in the technique of production employed which leads to an increase of the total output, the product due to particular kinds of input may possibly decrease.
The distinction between the periods for which we have to wait for the product of a unit of input and the period for which we have to wait for a unit of output will occupy us yet a good deal. At this point, however, one particular misunderstanding of the theorem that round-about processes of production are more productive may be mentioned, as it is due to a confusion between these two concepts. It has sometimes been argued that an increase of capital is more likely to shorten than to lengthen the time during which we have to wait for the product. And this is quite true when we speak of the time interval which will elapse before a given quantity of output will emerge. But this is quite compatible with a simultaneous increase in the periods for which we have to wait for the product of particular units of input. The use of elaborate machinery may not only very much shorten the time it takes to turn the raw material into a finished product but even make the time between the moment when the first input is invested in the machinery and the moment when the first output emerges shorter than the period during which we had before to wait for the product. Yet this has been made possible only by investing some of the input used in producing the machinery for a much longer period than any had been invested before. 
 Cf. A. A. Young, 1929, p. 796: “The use of capital saves time, in the sense that a larger product can be had with a given amount of labour. But it increases the average interval of time which elapses before the products of a given day’s labour reach their final form and pass into the hands of consumers.”
This way of looking at the concept of the period of investment also prevents us from falling into another common error. It is frequently supposed that all increases in the quantity of capital per head (at least when they do not involve changes in the quantities of durable goods) must, mean that some commodities will now be produced by longer processes than before. But so long as the processes used in different industries are of different lengths, this is by no means a necessary consequence of a change in the investment periods of particular units of input. If input is transferred from industries using shorter processes to industries using longer processes, there will be no change in the length of the period of production in any industry, nor any change in the methods of production of any particular commodity, but merely an increase in the periods for which particular units of input are invested. The significance of these changes in the investment periods of particular units of input will, however, be exactly the same as it would be if they were the consequence of a change in the length of particular processes of production. In referring to tools and machinery we have already had occasion to mention one of the most important groups of durable goods. These instances also show the sense in which the word durable is used in the present context. It may be not altogether unnecessary to point out that the word is not used here merely to indicate that a good will not soon perish, like meat or fruit, by the mere lapse of time. 
 The use of “durable” in the sense of merely storable is fairly widespread in the English literature on the subject (cf. for instance J. M. Keynes, 1936, p. 222) and is probably one of the reasons why the term “capital good” is so widely used by English writers to describe durable goods in the sense in which this term is used in the text. It seems, however, preferable, and more in conformity with the usage by the classical economists, to distinguish between perishable, non-perishable, durable, and permanent goods and to reserve the term capital goods as a description of all the first three types, that is all non-permanent goods.
It is used here to describe goods that are not destroyed in a moment by a single act of use but can be used repeatedly or continuously over a period of time. Durability in the first, more restricted sense may also give rise to investment, as when it is possible to produce a commodity at a season when it is cheapest to produce, and then to store it. But this would not be durability of the kind with which we are here concerned. Otherwise indistinguishable commodities which are available at different times of the year (e.g. ice in January and ice in July) must, for the purpose of economic analysis, be regarded as different commodities, and storage (transformation in time) as part of one possible technique of production. This case is therefore more properly to be regarded as an instance where a longer process of production gives a larger product.
Changes in the use made of durable goods may affect the quantity of capital used in two different ways. Either the quantity of durable goods used in a given process may change, or the durability of the goods may change. We shall first consider the effect of changes in the durability of the goods used.
Even within the category of durable goods in the narrower sense in which the term is here used some further distinction has to be made with regard to the way in which the durability of a particular good is determined. It is possible to conceive of a durable good which will last for a predetermined period of time irrespective of the amount of use to which it is subjected. During that given period it may be used more or less intensively and will accordingly give a greater or smaller amount of services. But once it is made its durability is finally determined, and we can speak of variability in its durability only in so far as we have the choice of producing otherwise similar goods of greater or lesser durability. Most buildings probably belong to this class. The other extreme will be represented by a good which embodies a definite quantity of services which can be used up at will either over a shorter or over a longer interval of time. In this case the time the good will last is determined not by the way in which it is made but by the way in which it is used. The obvious example here is machinery of most kinds, e.g. a motor-car. Neither of these types of durable goods is ever likely to be encountered in its pure form. In actual life we have to deal with various combinations of the two elements, though it will sometimes be useful to group them according as they approach more closely to the one or the other of the two extreme types.
More important, however, than this distinction is a similar one which is connected with the reasons why durable goods are used. It may happen that a particular instrument can only be made in a form in which it will last, and that the costs of making it are such that it is only profitable to produce it provided it will be used repeatedly or for a certain period of time. Or it may be that although a particular instrument could be made so as to serve for a single time only, its services are provided more cheaply if it is made in a durable form.
The first case is the rule where an instrument does not give off part of its substance or energy in the process of being used, but serves merely “as a tool”. A hammer or a derrick which would break while it was being used would be no good at all, and there are an almost infinite number of cases where a tool, in order to be useful, has to be made so strong that it will last – and in most cases remain useful – for a considerable period of time. But once it is known to last, it will be produced in such quantities, and the value of its services will fall to such a level, that if each tool were used only a single time it would not repay its cost : i.e. input will deliberately be invested in order to obtain a stream of services spread over a period of time. The question of what share of the input invested in such a durable good has to be regarded as invested for particular periods of time presents difficulties with which we cannot attempt to deal until later. For the present all that we need to consider is the fact that in many instances advantage cannot be taken of the help that instruments will render unless input which will repay its costs only over a considerable period of time is actually applied before the results begin to mature.
But this case where a particular instrument can be made only of a certain durability or not at all is but one of a much wider group of cases where the use of durable equipment enables us to obtain a greater return from given quantities of input. The more frequent case is associated with the fact that as a rule the additional expenditure involved in making equipment more durable so that it will give a greater amount of service, is much less than proportional to this added service. Even if it were possible to build a house which would last only for a month but would serve its purpose properly during this time, it would hardly be worth while to do so because such a house would probably cost little less than one which lasts for years. And although there will generally be some additional expenditure involved in making a good more durable, this additional expenditure will usually be very much smaller than the expenditure that is required to obtain the same amount of service during the early life of the good.
The effect of this can be best seen if we assume for a moment that a fixed amount of input is set aside in a society for the provision of particular durable goods. This input may be used to produce these durable goods in their cheapest and least durable form. In this case a large number of the durable goods will soon be available, periods and for some time the total amount of services obtained from them will continue to increase. But fairly soon a maximum will be reached: as soon as the goods which were produced first begin to wear out, all the available input will be needed for replacing them and no further increase in services will take place.
Compare this with the case where the input set aside for making durable goods is used to make fewer goods but more durable ones. At first the amount of services available will be smaller than in the first case. But since the durability of the individual good will increase more than in proportion to the increase in the expenditure on it, the total number of such goods which will have been created before the ones that were produced first begin to wear out (i.e. before it becomes necessary to devote the available resources to mere replacement purposes) will be greater than in the first case. The effect of using a given quantity of input for making more durable goods will thus be to provide a smaller quantity of services for some time at the beginning, but ultimately and permanently to provide a larger quantity.
As has already been suggested, what is called a transition to the use of “more durable goods” may mean two very different things, which, although the effect is very much the same for our purpose, must be clearly distinguished. First, it may mean a change towards the use of goods of greater durability; that is, it may refer to the phenomenon we have just discussed. But although this is probably the case most commonly referred to under this heading, it is not the only case and probably not even the more important case. “More durable goods” may also mean that a greater quantity of goods of a given durability (or of durable goods in general) will be used, compared with the amount of pure input which is invested in circulating capital. In this way the amount of capital may be increased by the use of “more durable goods”, although every individual durable good used now may be actually less durable than was the case before, because the quantity of such durable goods used in proportion to the pure input employed has increased more than the durability of the individual durable goods has decreased.
In practice a change to the use of more goods of a given durability (as distinguished from a change to the use of goods which are more durable) will as a rule mean that a different, more expensive, and more labour-saving type of equipment will be used in a given process of production. Professor G. Akerman, who has devoted a special study to this phenomenon, has proposed to describe this difference between different instruments of the same durability which are designed to co-operate with a proportionately larger or smaller amount of “labour” (pure input) as the degree of automatism of these different kinds of capital goods. 
 See G. Akennan, 1923, chap. iii/4 and chap. v/1, pp. 27, 39 et seq., and 1924, p. 284; also Wicksell, Lectures, vol. i, appendix, and Lindahl, 1925, p. 81.
They can also be described as more or less labour-saving types. In certain contexts these terms are convenient and we shall occasionally use them. But it is important to remember that the use of more “automatic” or more “labour-saving” machinery is only a special instance of a change towards the use of more durable instruments and that, whether a greater quantity of durable goods (in the sense of durable goods possessing a greater value and probably of a different kind) or durable goods of greater durability are being used, comes for our purposes to very much the same thing. In both cases part of the total pure input used in the process of production will now be invested for longer periods. The only important difference is that in the first case input which has already been invested in durable goods will now (together with some additional quantities of pure input) be invested in still more durable goods, while in the second case input invested in durable goods is substituted for input invested in goods in process. Nor should the question why it should become profitable to use a greater quantity of durable goods require any further explanation beyond that already given for why durable goods should be used at all. The fact is simply that of the various tools, etc., that can be used some will be more efficient than others; that among the tools of equal costs and equal durability the most efficient will always be chosen; that consequently the adoption of still more efficient instruments will require more capital because these more efficient instruments will be either more durable or more costly for some other reason; and that, therefore, if capital is supplied more amply and cheaply, this will lead either to the use of tools of greater durability, or to the use of more costly tools of given durability, or both. 
 It is also possible that the increase in the supply of capital may bring about one of these effects to which an extent would more than offset the consequences of the other effect going in the opposite direction.
CAPITAL AND THE “SUBSISTENCE FUND”
We must return now to the considerations from which we started in Chapter V. The significance of the various forms of the productivity of investment which we discussed in the last chapter is, as we have seen, that the existence of a stock of resources of limited durability enables us to ment keep income permanently above the level that could be secured by the direct use of current pure input. It does so by providing income during the time that we have to wait for the return of the input that is being currently invested. But it is important, even at this early stage of the exposition, that we should not unduly simplify the relationship between the stock of non-permanent resources and the possible range of investment periods. The very expressive term “subsistence fund” which has been used to describe this function of the stock of non-permanent resources is apt to be misleading in a number of ways. It is of course not a stock of actual means of subsistence, but only a stock of resources which can be turned into means of subsistence, i.e. into consumers’ goods.  The process of transforming these resources into consumers’ goods requires the co-operation of current input; and the amount of consumers’ goods that they will yield and the time when those consumers’ goods will accrue, will depend on the way in which the services of the permanent and the non-permanent resources are combined. The latter, therefore, do not represent a fixed quantity of consumers’ goods or a stream of them of fixed time shape. It is all a question of which combination of the different resources is most advantageous. The fact that it is the existence of non-permanent resources which enables us to invest the services of the permanent resources does not mean that the fruits of the former will always be consumed at the earliest date when they can be made available, and that only the latter will be invested. Whenever a greater output is to be expected from further postponing the return from non-permanent resources, and in the meantime using the services of the permanent resources, this arrangement will be the one to be adopted. It cannot be too strongly emphasised that the services from non-permanent resources, no less than those from permanent resources, are objects of investment.
 N. W. Senior was again the first person until comparatively recent times who saw this connection at all clearly. “Nor is it absolutely necessary”, he writes (Political Economy, 1836, pp. 78-79), “though if Adam Smith’s words were taken literally, such a necessity might be inferred, that, before a man dedicates himself to a peculiar brand of production, a stock of goods should be stored up to supply him with his subsistence and materials and tools, till his own product has been completed and sold. That he must be kept supplied with these articles is true; but they need not have been stored up before he first sets to work, they may have been produced while his work is in progress. . . . That fund must comprise in specie some of the things wanted. The painter must have his canvas, the weaver his loom, and materials, not enough, perhaps, to complete his web, but to commence it. As to those commodities, however, which the workman subsequently requires, it is enough if the fund on which he relies is a productive fund, keeping pace with his wants, and virtually set apart to answer them.”
As has already been remarked, most non-permanent resources owe their existence to the fact that they make investment possible without a temporary reduction in the income stream. All or nearly all the man-made equipment is non-permanent, and the greater part, although by no means all, of the non-permanent resources existing at any moment consists of man-made production equipment. This brings us back to the relation of my definition of capital as the “stock of non-permanent resources” to the traditional one of the “produced means of production”.
Indeed these two definitions merely emphasise different aspects of the same process, which, as completely stationary conditions were approached, would tend to become identical. As time goes on the non-permanent resources have to be replaced by deliberately produced equipment, and ultimately, if technological progress stopped, all non-permanent resources which were the gift of nature would be exhausted. And there would of course be no additions to strictly permanent resources in a stationary state. Consequently, under perfectly stationary conditions, where everything continually repeated itself in an unchanging way, the genetic definition of capital as the produced means of production would also define the non-permanent resources and the way in which they would be currently reproduced.
But the fact that a definition would be adequate under purely stationary conditions means less in the theory of capital than almost anywhere else in economics (with the exception of the theory of money). The theory of capital is largely concerned with the significance of those wasting resources which, in Wicksell’s words, “cannot, strictly speaking, be included in the scheme of a stationary economy”.  Thus it is only by an extreme and almost numbing effort of abstraction that the theory of capital can be made to satisfy the requirements of stationary analysis. In actual life the existing stook of capital goods is always the result of an accidental historical process, consisting of a succession of unforeseen changes, and they will never be reproduced in exactly the same form. They were only produced in this particular form because certain kinds of equipment happened to be available as the result of past history.
 Cf. Lectures, vol. i, p. 151.
The essential characteristic of capital, and the one which affects the use of current input, is that it needs replacement and in consequence leads to investment.
This in turn leads to the creation of new capital, but once this new capital exists the historical aspect becomes irrelevant. The important thing is not that the capital has been produced, but that it (or some equivalent) has to be reproduced. Under stationary conditions the two aspects will of course coincide. But under dynamic conditions this will not be so. If income is to be maintained permanently at the higher level which the wasting natural resources make possible, these resources will, as they are exhausted, have to be replaced by produced means of production.  And occasionally it may be found advantageous to replace non-permanent resources by some change of the surface of the earth, which, like a tunnel, may be expected to remain permanently useful.
 It may be pointed out here, although it does not strictly belong to our present subject, that a treatment of the problem of the conservation of exhaustible resources under the aspect of their representing part of the national capital, would at last put the somewhat confused discussion of these problems on a sounder basis. There is, of course, no reason why, e.g., forests should be maintained at the particular size at which they happen to be at any given historical moment – although there may be other considerations that have to be taken into account than merely the direct profitability of the forests in question (effects on climate, soil erosion, aesthetic considerations, etc.). But it should always be kept in mind that any exhaustible resource represents just one item of the national capital which may be more useful in some other form into which it can be converted.
The last case is particularly interesting because it points to a significant difference between the two definitions. On the classical definition the tunnel, and any other piece of man-made equipment that was regarded as permanent, would be counted as capital merely because of its historical origin. Actually, however, once it existed it would have none of those effects on the use of current input which are peculiar to the non-permanent resources. It would no longer represent a supply of capital which, if conditions changed, could be transformed into a more desirable shape. Nor would it (either from the social or from the private viewpoint) represent a reserve on which one could draw in order to obtain a temporary increase in current income. Although it would be a “produced means of production”, it would have none of the characteristics which create the special problems relating to “capital”. 
 Cf. the passage from Wicksell, quoted above, p. 57, footnote.
It should be observed at this point that the concept of capital arose out of the need for distinguishing the “substance” of an asset (which has to be replaced) from its yield: that is, for dividing gross returns between amortisation and interest. This is of itself sufficient reason for reserving the term for non-permanent assets. Where there is no “turnover” of stock, but only a permanent stream of services, no problems of capital arise.
But there is another, even more fundamental, reason why the definition of capital as the produced means of production should be definitely abandoned. And this is, that it is a remnant of the cost of production theories of value, of the old views which sought the explanation of the economic attributes of a thing in the forces embodied in it.  But, except as a source of knowledge, the actual history of a particular thing, i.e. the way in which it has acquired its qualities, is entirely irrelevant. It has nothing whatever to do with the decision as to how the thing shall be used henceforth. Bygones are bygones in the theory of capital no less than elsewhere in economics. And the use of concepts which see the significance of a good in past expenditure on it can only be misleading.
 Cf. C. Menger, Zur Theorie des Capitals, passim, and the passage from Professor F. H. Knight’s discussion of these problems already quoted (see above, p. 10, footnote).
All this does not mean that the relation of capital to investment, and the creation of capital by investment, is not of the utmost importance. In fact it is so important that the greater part of what follows will consist of an attempt to clarify this relationship. What it does mean is that the important thing is not the relation of existing capital to past investment, but its influence on current investment, its influence on the creation of the capital goods of the future. For all problems connected with the demand for capital, the possibility of producing new equipment is fundamental. And all the time concepts used in the theory of capital, particularly those of the various investment periods, refer to prospective periods, and are always “forward-looking” and never “backward-looking”. But for determining what resources are functioning as capital at any given moment, the essential point is not that particular resources have been produced; it is that they are not permanent, but of limited durability, and therefore must be replaced by some new resources if the income stream is not to decline.
If in our definition of which of the present resources are to be considered as capital and which are not, we single out the non-permanent resources as capital, we must not overlook in doing so that in a sense the capital problem is double-faced and that its two aspects make it necessary to recognise two categories of goods which are not necessarily identical. From the first point of view we are concerned with what it is that enables us to wait, and from the second point of view with what it is that enables us to draw advantage from this possibility of waiting. The investment which the waiting makes possible can of course take the form only of things which can be produced. And when we are thinking of the capital to be produced this can of course consist only of such things as can be produced. But what we mean when we say that the existence of particular present resources makes waiting possible is that these resources provide a temporary income stream during the period while we wait for some other income to mature. There is no other concrete meaning which we can attach to the vague but much used concept of the “amount of waiting” available than the amount of services available in the near future as compared with those which will only be available in the more distant future. And there can be no doubt that this is the sense in which the term capital is used in everyday speech. A person who in the present year cannot lay hands upon a greater amount of resources than will recurrently become available to him in every successive year is not commonly regarded as commanding any capital in addition to his income, however large that income may be. A big landowner, for instance, may well be short of capital and, unless he is able to “raise capital” on the security of his land, be unable to make any investments to intensify the cultivation of his land.  And a country may be very rich in land and lack the capital, which may be supplied by another country whose total wealth may be much smaller. A person or a country, on the other hand, who in addition to a secured (and non-anticipatable) stream of permanent income commands an amount of resources which can be used up in the near future, is generally held to own capital.
 It is true that because of this possibility to borrow on the security of land (or any other permanent resource) the distinction between capital and land tends to become blurred and individuals more and more tend to treat land as a part of their “capital”. But the distinction between capital and land is surely not an invention of the economists and I find it difficult to believe that the efforts of a number of modern economists (particularly Professor Fisher, Professor Fetter, and Edwin Cannan) to expand the concept of capital so as to include all wealth are to be recommended.
It has recently been suggested  that instead of the traditional concept of produced means of production or of the concept of non-permanent resources used bere, the reproducible or augmentable resources ought to be considered as representing “capital”.
 N. Kaldor, 1937, p. 219.
But quite apart from the fact that the criterion of augmentability is either exceedingly vague or, if taken literally, would narrow down the range of augmentable resources in such a way as to make the concept useless for our purpose,  the point that is relevant for our problem is not that certain existing resources can be replaced by others which are in some technological sense similar to them, but that they have to be replaced by something, whether similar or not, if the income stream is not to decline. A deposit of metal ore is no less capital because it cannot be reproduced, and vice versa the water-power of an existing stream would not become capital in the relevant sense because of the possibility of creating a new stream by collecting rain-water which is now allowed to evaporate.
 Mr. Kaldor’s definition of “producible” or “augmentable” resources appears to be that they have perfect (or at least very close) substitutes which it is economically possible (i.e. profitable) to produce. If perfect substitutability (in the sense of the goods by which the existing resources are replaced having an infinite elasticity of substitution with the latter) were required, nothing except goods which it is profitable to reproduce in exactly identical form would have to be counted as capital. But if only the existence of close substitutes is required, where is one to draw the line? On the other hand, if only resources which it is economically possible to reproduce are to be counted as capital, all obsolete machinery would cease to be capital as soon as better machines became available; while if the technological possibility of producing substitutes were to decide, even land could not be excluded, since it is possible to grow practically any fruit in artificial compounds (“tray farming”).
What determines the special common characteristics of capital goods is not that they can be reproduced, but how they can be used: namely, that they can be made to yield all their services in the comparatively near future. And it is this fact and no other which in a sense makes them one common “fund”; that they are capable of producing income for the same period of time, the comparatively near future. Even very different consumers’ goods that are available at the same time are substitutes to a much higher degree than even otherwise identical goods which, however, are available at dates very distant from each other. And so long as we can increase our income by investing, that is, postponing the date when some resources will yield consumable services, everything which can be used to give an income during the interval, and so will enable us to make use of the opportunities of investment, possesses the common attribute of being a condition of making investment possible. In other words, so long as there is a special inducement to postpone the date when some resources will yield their final services, the common attribute of the things capable of rendering services in the interval is a scarce “factor” on which certain additions to our future income depend. But if the opportunities of adding to output by investment should cease, although part of our resources would still have the exclusive attribute of being capable of rendering services in the near future, yet the existence of any particular unit of such factors would cease to be a condition for the possibility of investment and their common quality of being available in the near future would cease to be scarce relatively to demand.
The recognition that the constituents of the stock of capital possess in this sense a common quality has often led economists to speak of it as if it could be treated as a homogeneous “fund”, an “amount of waiting”, or as a given quantity of “capital disposal” or of “pure capital” in the abstract. If these terms were used occasionally to express no more than has been explained in the preceding paragraphs, little objection would be raised against them. Unfortunately, however, much more far-reaching assertions have been made about the real existence of such a fund which I cannot but regard as pure mysticism. The best known representatives of this view are of course J. B. Clark and Professor Gustav Cassel, and the views of the former have recently been revived by no less an authority than Professor F. H. Knight. In his opinion, the “basic issue” which at present divides economists “is the old and familiar one between two conceptions of capital. In one view it consists of ‘things’ of limited life which are periodically worn out and reproduced; in the other it is a ‘fund’ which is maintained intact though the things in which it is invested may go and come to any extent. In the second view, which is of course the one advocated here,  the capital ‘fund’ may be thought of as either a value or a ‘capacity’ to produce a perpetual flow of income.”  I am afraid, with all due respect to Professor Knight, I cannot take this view seriously because I cannot attach any meaning to this mystical “fund” and I shall not treat this view as a serious rival of the one here adopted. What I have to say about the former I have said in another place,  and here I shall not discuss it again beyond pointing out certain errors which are due to its influence. 
 I.e. by Professor Knight.
 F. H. Knight, 1935c, p. 57.
 Hayek, 1936a.
 See below, Chapters XXIII-XXV.