APPENDIX II. The "Conversion of Circulating Capital into Fixed Captial"

APPENDIX II THE "CONVERSION OF CIRCULATING CAPITAL INTO FIXED CAPITAL" THE idea of a "conversion of circulating into fixed capital" has played a considerable role in the discussions of the dynamics of capital formation and of industrial fluctuations in particular from the times of Ricardo down to Knut Wicksell. A certain confusion about its exact meaning, and in particular between two different situations which the concept might describe, has, however, deprived it of much of the fertility it might have had. A short note may therefore be devoted to the task of clearing up the confusions involved. The idea, if not the actual terms, appears to have ,been intro duced into economic discussion by Ricardo, when in the new chapter" On machinery" in the third edition of his Principles he admitted in reply to Barton that the sudden discovery and extensive use of improved machinery may have the effect of "diverting capital from its actual employment" as circulating capital in order to increase the amount of fixed capital l and in consequence decrease the gross produce and consequently the fund for the employment of labour. As J. S. Mill put it later, the capital" has been converted from circulating into fixed capital, and has ceased to have any influence on wages or profits". 2 It was not, however, in connection with the doctrine of the effect of technical progress on wages and profits that this idea became most influential. The expressions used here seemed to provide a perfect description for the phenomena which were observed during the major booms, when a period of extensive construction of fixed capital was followed by an intense scarcity of capital, and the idea was consequently turned into 1 Work8 (ed. McCulloch), p. 24l.

2 Principle8 of Political Economy (ed. Ashley), p. 734. A more systematic treatment of the subject, in fact the most complete to be found anywhere in the classical literature, is given earlier in Book I, chap. vi, of the same work. 424 Oirculating and Fixed Oapital 425 an explanation of commercial crises which for a long period was very widely held. One of the first authors to explain the crises by a scarcity of circulating capital caused by an excessive conversion of circulating capital into fixed capital appears to have been the American Condy Raguet. 1 But it was largely through the considerable elaboration which this idea received in James Wilson's Oapital, Ourrency and Banking (1847) that it became widely accepted and for the next thirty years remained almost the dominating explanation of crises. J. S. Loyd, T. H. Williams, O. Michaelis, R. Torrens, J. G. Courcelle-Seneuil, V. Bonnet, J. Garnier, W. S. Jevons, J. Mills, H. v. Mangoldt, Leone Levi, Bonamy Price, and Yves Guyot, to mention only the more important represent atives in chronological order,2 all made use to greater or less extent of this idea in their theories of crises.

In all these different versions of the theory the crucial point is that, towards the end of a boom, a scarcity of circula ting capital and a consequent rise in the rates of interest make it impossible either to complete the large projects for invest ment in fixed capital or profitably to use the additional plant so created. It would lead us too far afield to discuss here the rela tions which are supposed by the different writers to exist be tween these phenomena and credit expansion. Nor is it possible 1 A Treatise on Currency and Banking (London, 1839), pp. 62 et seq. 2 Cf. Lord Overstone (J. S. Loyd), evidence given in 1848, Tracts on MetaUic and Paper Currency (1857), pp. 489, 590; T. H. Williams, Observation8 on Money, Credit and Panics (1857); O. Michaelis, Die Handelskrisis von 1857, reprinted in Volkswirtschaftliche Schriften (1873), vol. i; R. Torrens, Principles and Practical Operations of Peel's Act (3rd ed., 1858), p. 95; J. G. Courcelle-Seneuil, Traite d'eco nomie politique (1858-9), vol. i, pp. 361-363; V. Bonnet, Que8tions economiques etfinancieres a prop08 des crises (1859), pp. 1-11; J. Garnier, art_ "Crises commerciales" in the Dictionnaire universel theorique et pratique du commerce, vol. i, p. 925; W. S. Jevons, A Serious FaU in the Value of Gold (1863), p. 10 (reprinted in Investigations upon Currency and Finance, p. 28); J. Mills, "On Credit Cycles and the Origin of Commercial Panics", Transactions of the Manchester Statistical Society (SessIon 1867-68), 1868, pp_ 9-40; H. v. Mangoldt, Grundriss der Volks wirthschaftslehre (1863), p. 68; Leone Levi, Banker's Magazine (New York, 1878), vol. xxxiii, pp. 40-45, 118-126; Bonamy Price, Chapters on Practical Political Economy (1878), pp. llO~124; Yves Guyot, La Science economique (1881). On these authors see E. v. Bergmann, Geschichte der Nationa16konomischen Krisentheorieen (1895), and T. S. Ashton, Economic and Social Investigations in Manchester (1934).

426 Appendix II here to trace the important influence which these views have had on the theory of crises of Karl Marx, through him on M. v. Tougan-Baranowski, and through the latter on such contemporary authors as G. Cassel, A. Spiethoff, and D. H. Robertson. K. Wicksell, on the other hand, who repeatedly makes use of these concepts, is probably more directly indebted to the earlier writers.1 In this note, however, we are not so much concerned with these elaborations of the theory. We want merely to disen tangle the different meanings attached to the concept of the conversion of circulating capital into fixed capital. It will be shown that the original Ricardian contention about the effect of such a conversion on the size of the "gross produce" rested on a confusion between the stock of circulating capital proper and the stream of output available for current consumption - a confusion which also is responsible for the cruder forms of the wage fund analysis; but that in the way in which the proposition was used by later writers as an explanation of crises, that is as referring to a temporary phenomenon during periods of transition, it described a real phenomenon, and that the" reduction of the fund destined for the support of labour" describes the same phenomenon which later became generally known under the name of forced saving.

Our discussion may be conveniently divided into three parts. We shall first try to show that the proposition which the classical economists used has really little to do with the particular distinction between circulating and fixed capital as defined by them, but is connected with changes in the time dimension of capital in general (or the substitution of a growth of capital in height for a growth in width) irrespective of whether this is in connection with a relative increase of fixed capital or not. Secondly, we shall show that so long as we compare alternative positions of equilibrium, one with rela tively more and the other with relatively less fixed capital, 1 Cf. Lecture8, vol. i, p. 164: "That the transformation of circulating capital into fixed capital, i.e. the change from short-term to long-term capital investment, may frequently injure labour, is beyond doubt ". Ibid. p. 185: " ... during booms, when large quantities of circulating capital are converted into fixed capital and it is not possible to replace the former quickly enough. In the subsequent depression the con· ditions are usually reversed; there is plenty of circulating capital, but it is no longer projitaJJle to convert it into fixed capital."

Circulating and Fixed Capital 427 this difference cannot affect the size of the gross produce in the sense of the classical writers, and that consequently their argument about the effect of such a conversion on wages was mistaken. And thirdly, we shall try to explain how, under dynamic conditions and during periods of transition from one equilibrium position to another, the effect in question may actually lead to a temporary reduction of gross produce, and thus, if savings are not increased sufficiently, give rise to that scarcity of consumers' goods which is the real equivalent of the phenomenon described by the classical writers as a scarcity of capital. (1) The argument rests in the first instance on the simple idea that while with a given amount of capital, if it assumes such a form that the whole of it is turned over once a year, say in the form of a stock of raw materials, the product derived from it in the course of a year will be equal to the total value of this capital, yet if the same amount of capital is invested in such forms that only one-tenth of it will be turned over in the course of one year, the annual product due to it will be only one-tenth of its former value. From this it follows that if "circulating capital" (in the sense of goods in process) is converted into" fixed capital" (in the sense of durable goods), the annual product due to that amount of capital will be decreased. But it will be seen without difficulty that this is quite independent of whether the lengthening of the invest ment periods involved is due to a substitution of durable goods for goods in process or not, and that exactly the same con sequences will follow if a given amount of circulating capital is used to finance a process of longer duration instead of one of shorter duration. If a manufacturer who cannot increase the amount of capital at his disposal changes from one kind , of process of production where the" period of production"

in the narrower sense of the term is shorter to one where that period is longer, he will clearly now be able to employ only fewer men than before, and his annual output also (at least measured in factor terms) will be smaller than before (although presumably his profits will be larger). On the other hand, it is at least conceivable that a change which involves a substitu tion of durable goods for goods in process may not have that effect, because the periods for which the input remains invested in the durable good may be actually shorter than the investment 428 Appendix II periods involved in some very time-consuming process, such as some kinds of tanning. We must therefore conclude that the proposition that a conversion of circulating capital into fixed capital will bring about a reduction in the rate of output due to that capital is not strictly correct if we define fixed capital as durable goods and circulating capital as goods in process, but becomes true if we define the two kinds of capital, as has been suggested above (Chapter XXIV), according to their final distance from t ~ --------------,Q consumption. If, for in stance, we decide to define all parts of the existing stock of capital which will be transformed into consumables within a year as Tl circulating capital (includ ing in this therefore those parts of durable goods which can be used up dur ing the next year) and all other capital as fixed capi tal, it is clear that any change in the composition of a given quantity of o 1" capital so that less of it is FIG. 32 now circulating capital in this sense and more of it is fixed, must mean that the rate of final output from that capital must decrease.

(2) This proposition applies, however, only to the output which is due to a particular quantity of capital, and it is a mistake to generalise this argument so as to apply to the output of society in general. This can be most conveniently demonstrated by adapting one of our earlier diagrams repre senting input function in its simplest form. In Fig. 32 the fully drawn curve OR represents the input curve in its inverted form, and the area enclosed by this curve and the two co ordinates measures the quantity of capital (this, since we are using the input function and are disregarding interest, is measured in factor terms). If we decide to call circulating capital that part of the total capital stock which will mature Oirculating and Fixed Oapital 429 within a period of the length T 1T 2, the stock of circulating capital will be represented by the area T 1T 2RP, while the area of the remaining part of the curvilinear triangle, OT1P, would represent the stock of fixed capital.

The dotted curve OR represents an alternative input function, that is, an alternative method of production which requires the same total amount of capital but is composed of a greater amount of fixed capital and a smaller amount of circulating capital. The amount of circulating capital in this case is represented by the area of T 1T 2RP and is by an amount represented by the area enclosed between the fully drawn and the dotted curves RP smaller than the amount of circulating capital used in the first case, while the amount of fixed capital is correspondingly larger. The classical economists deduced from this that the amount that will be available for the payment of wages (and other incomes) during the unit period will be correspondingly reduced. But this is clearly wrong and due to a confusion between the stock of circulating capital and the flow of income derived from it, as can easily be shown. The rate at which income matures under stationary conditions is measured in our diagram by the line T 2R, and the amount of income maturing during the unit period of time will, under stationary conditions, always be represented by the area of the rectangle T 2T 3QR, whatever the composition of the capital. The amount of income, if we measure it in terms of its own, will of course vary with changes in the methods of production, but measured in factor terms, as done for our present purposes and as the classical economists did, income (or output) will under stationary conditions always he equal to current input. We have seen before that changes in the structure of production due to technological changes may be injurious to labour by changing its marginal productivity.

But although this may be accompanied by a conversion of circulating into fixed capital, it is not a direct or necessary consequence of it, as the classical economists believed.! 1 Cf. Wicksell, Lectures, vol. i, p. 164: "That the transformation of circulating into fixed capital, i.e. the change from short· term to long. term capital investment, may frequently injure labour, is beyond doubt. But Ricardo was mistaken in his belief that this consequence was due to the fact that the gross product is simultaneously reduced. This, as may easily be proved, is theoretically inconceivable. The 430 Appendix II (3) The situation is, however, different when, instead of considering two alternative positions of stationary equilibrium, we ask what happens during the period of transition from the one state to the other, particularly when the relative increase of circulating capital is merely a prelude to a change which required an increase in the aggregate quantity of capital. In the case where, in spite of the increase in the relative amount of fixed capital, the total amount of capital remained constant, this result was obtained because the lengthening of the invest ment periods of part of the input was compensated for by a shortening of the investment periods of other parts of the input. (In Fig. 32 this was shown by the new input curve in its left part lying below, and in its right part lying above the old one.) By this double change it was made possible for output to continue to mature throughout at a constant rate in spite of the change in individual investment periods.

But it is of course conceivable that fixed capital may be increased at the expense of circulating capital by lengthening the investment periods of some input without a compensating shortening of others. And at first, and for a period corre sponding to the original period of investment of the input which is now invested for a longer period than before, this will be possible merely at the expense of circulating capital, that is, without increasing the total quantity of capital. But after a while the effect of such a net lengthening of investment periods must be that for a time the current output will be reduced below the product of current input. And if the new investment structure is to be completed, it will be necessary that for a time people consume less, and by their saving make it possible to create the additional capital which the new structure requires. n this case the increase of fixed capital, which at first took place at the expense of circulating capital, will require a later net increase of capital by corre sponding additions of capital in the lower stages, and of circulating capital in particular. In this case the conversion of circulating capital into fixed capital has created an incomplete capital structure which needs completing by further net addi tions to capital and corresponding saving.

gross product under free competition (where such is at all possible) always tends in the main towards the maximum which it is physically possible to obtain with the existing means of production."

Circulating and Fixed Capital 431 This case may again be illustrated by a simple diagram. In Fig. 33 the fully drawn input curve OR represents again the old process before the conversion of the circulating capital into fixed, and the dotted curve OQS the situation after some of the existing circulating capital has been converted into fixed. In this case, however, the second curve represents not a complete structure but merely the position at a point during the process of transition when the new processes have been started but not yet concluded. The complete t new process would be dp~ scribed by the dotted curve OQR, and in order that it T2t-------S-'i----~R can be completed it will be necessary to add during the period TIT 2 to the stock T,I-------T'--4: of capital an additional amount corresponding to QRS which can only be pro vided if during that period people reduce their con sumption from T2R to T~S.

If we could assume that at this stage people voluntarily and spontaneously will re duce their consumption to o I , , FIG. 33 r this required extent, no problem arises. But if they do not and continue to spend on consumption goods as much as before, the amount of capital required for the completion of the process will not be forthcoming; that is, there will arise that " scarcity of capital" discussed in classical theory, which of course means a scarcity of consumers' goods and a rise in their price and in profit margins generally, which will make investment in long processes of this kind unprofitable. It hardly needs pointing out that to discuss this whole phenomenon in terms of changes between " circulating" and " fixed" capital is somewhat misleading. Quite apart from the fact that we have found it necessary, in order to make the argu-. ment consistent, to substitute a definition of what we mean by fixed and circulating capital other than the usual one, it is clear that even on our definition the argument does not 432 Appendix II apply only to shifts between the two parts of capital which we have arbitrarily divided on the basis of some standard period, but equally to any other change in investment periods, as for instance to an increase of what we would have to call relatively more fixed at the expense of relatively less fixed capital. It appears that here as elsewhere any attempt at a sharp division of capital into two groups, although sometimes illustrative, is dangerous and misleading, and has to give place, in more precise analysis, to a treatment which takes account of the essential continuity of the range of periods for which input is invested.