XX. The Accumulation of Capital

CHAPTER XX THE ACCUMULATION OF CAPITAL THE significance of the conditions of stationary equi librium of capitalistic production which we have discussed in the last chapter will become clearer only as we apply the same considerations to the study of the effects of different types of changes. In the present and the suc ceeding chapters we shall therefore discuss the readjust ments that will be made necessary if, after various types of changes in the data, equilibrium is to be re established. It will be convenient to divide this task into two parts. In this and the next chapter we shall discuss those changes which can properly be said to originate on the side of Types of changes to capital, that is, where what is usually called be discussed a change in the supply of capital is the originating and active cause. Under this head we shall have to deal with the accumulation and de cumulation of capital (or with saving and dissaving), in a competi tive society with otherwise stationary conditions. In Chapters XXII-XXVII we shall go on to discuss the effects of other changes, and particularly of unforeseen changes in the data. These will include changes in the supply of factors and in technological knowledge. They will also include such changes in tastes as reflect them selves in the relative values of different commodities at a moment of time but not in the time shape of the desired income stream. In this second group of changes capital plays essentially a passive role, although, as we shall see, these changes too will in most cases give rise to new processes of saving and dissaving.

This attempt to treat certain dynamic problems will 268 OH.XX The Accumulation of Oapital 269 not, however, allow us to deal with all the complications which arise under similar circumstances in the actual world. In particular it is necessary to remember that the analysis still proceeds in real terms, that is, under what amounts essentially to barter as-AbHnoootunusedre sumptions, and that this makes it impos-sources assumed sible to deal with the existence of unused resources due to the rigidity of money price-so The following analysis of the effects of the accumulation of capital must therefore assume that there are no unused resources in existence in the technical sense of the term. In other words, at the ruling prices no more resources are obtainable than are actually employed. On these assumptions every increase of capital will mean an increase of capital relatively to the quantity of other factors of production, or what Bohm-Bawerk called an increase in the amount of capital DlsolIS5lon oooOood Per head This is an important point since 10 changes In capital • , relatively io pur. loan increase of capital in proportion to the pui other factors employed will have different effects from an increase which is merely proportional to a simultaneous increase in the amount of pure input that is employed.

It is only in the former case when the quantity of capital increases ceteris paribus that it will lead to the peculiar consequences usually connected with an increase of capital, such as a fall in the rate of interest, a change in the technique of production, or the adoption of longer or more roundabout methods of production. This first case must be sharply distinguished from the second where the increase in the total quantity of capital merely means that a proportionally larger quantity of labour and other pure input is being equipped with the same kind of instruments, etc., as those formerly employed. In past discussions these two cases have sometimes been described as a growth of capital in its "time dimension" and in its " labour dimension" respectively, and Knut Wicksell referred to it as a distinction between 270 Capitalistic Production under Competition PT. III the growth of capital in " height" and in " breadth ".1 More recently Mr. Hawtrey has for the same distinction suggested the terms "deepening" and "widening" of capital. 2 The point which is relevant here is that on our assumptions, and for the economic system as a whole, the growth of capital can take the form of a growth in "height", or of "deepening", only. Although this assumption is of course somewhat unrealistic, it has the advantage of stressing an effect of the accumulation of capital which will always be present to some extent, and which, as has been amply proved by recent discussions, is liable to be overlooked. Yet it is the changes which are connected with this" heightening" or " deepening"

of capital in which the special characteristics of a growth of capital are best seen, and for this reason this assump tion actually helps to bring out an important point. In the discussion of stationary conditions in the last chapter it was generally assumed that everybody endeavours to maintain his income stream constant. The effects of plan-What this implies when there are changes ning for an in~reasing in the other data we shall consider in some or a decreaSing income stream detail in the next chapter. The question which we have to discuss at present is what happens if some members of the community aim either at an increas ing or at a decreasing income stream, and consequently invest either more or less than would be required to keep their income constant. It will be remembered that all resources except those small marginal quantities (or fractions of the value of some resources) which owe their value to the last incre ment of investment, give their owners a potential com mand over consumers' goods. In this sense, then, all the resources in the possession of ~n individual (except 1 cr. Wicksell, Lectures, vol. i, pp. 163 and 266, and also ibid. p. 164, where he speaks of changes in the" vertical" and the" horizontal"

dimension of capital. 2 Hawtrey, 1937, p. 36.

ell. xx The Accumulation of Capital 271 these marginal quantities) are" free capital" to him and give him an option whether to consume a corresponding quantity of consumers' goods immediately or to invest this" free capital". Stationary conditions would of course exist if every individual divided this potential command over ready consumers' goods between current consumption and pro vision for future consumption in such a Net changes only will way as to ensure him a constant income be considered stream. But this, as is well known, is not likely ever to be the case, nor is it necessary in order to create stationary conditions. The mere following on of successive genera tions, if nothing else, will regularly have the effect that at anyone time some people will be saving and others dissaving. But it is a familiar axiom that, if the popula tion as a whole remains stationary, these different actions may just cancel out and the supply of capital in the aggregate remain constant, although perhaps no single individual aims at a constant income stream in perpetuity.

We are here interested, however, merely in the net effects of the decisions of all individuals in the aggregate. And in order to avoid the complications arising out of the existence of a multitude of divergent decisions, of which only the net result is of interest to us, we shall assume for the present that the majority of people do aim at a constant income stream in perpetuity, but that there is just a small group of people all of whom aim at an increasing (or a decreasing) income stream. So that the only net change we have to take into account is due to the decisions of those comparatively few people. When we use here the terms saving and dissaving, this will always mean, by definition, that people provide for an increasing or decreasing income stream respectively. In the earlier part of this chapter we shall .. Saving" and" dis concentrate mainly on the case where saving" people aim at an increasing income stream, i.e. on the ease of positive saving and investing. Most of the geneml 272 Capitali8tic Production under Competition PT. III conclusions arrived at in this connection apply also to the reverse case of dissaving. But certain peculiarities of this case will be discussed explicitly in Chapter XXVI.

If we assume that in a society in which up to now stationary conditions have prevailed some persons sud denly decide to consume less and to demand a greater Foreseen and unfore-quantity of capital goods, this cannot affect seen saving the relative quantities of the two kinds of goods available until such time as it takes for production to be adjusted to the changed relative demand for con sumers' goods and capital goods. And an actual reduction of consumption will in consequence only be necessary when, as a result of the diversion of resources from the production of consumers' goods to the production of capital goods, fewer consumers' goods become availabie. 1 If, however, as we may assume to start with, the reduction of consumption occurs suddenly or unexpectedly, this must, as can easily be shown, lead in the first instance to an at least temporary accumulation of stocks of consumers' goods. Since there is a great deal of confusion in all the discussions of this barter mechanism of saving and invest ment, it is necessary to go into the matter in some detail.

It has often been argued, and is implicitly assumed in much of the classical doctrine on the subject, that the consumers' goods saved by one class of people are merely The producers of new transferred to and consumed by another caphal goods are not class of people at the same time as they supported out of the consumers' goods would have been consumed by the first saved class. 2 We shall see later that this is to some extent true in the case where the amounts saved are used to employ formerly unemployed workers.3 But this 1 For this and the following three sections, cf. Bresciani.Turroni, 1936, and R. v. Strigl, 1934a. 2 Cf. Adam Smith, Wealth of Nations, Book II, chap. iii (ed. Cannan, vol. i, p. 320): "What is annually saved is as regularly consumed as what is annually spent, and nearly at the same time too; but it is consumed by a different group of people". 3 cr. below, pp. 370 et seq., and Hayek, 1939, pp. 44 et seq.

elI. xx The Accumulation of Capital case is here exCluded by our assumptions, and it must at least appear doubtful whether it was to this case, and to this case alone, that the argument was thought to apply. The idea which underlies the argument and which has been given wide currency by the exposition of J. S. MilV appears to be that the amounts saved are offered t,o the resources which are to be diverted to the production of capital goods in order to attract them away from the current production of consumers' goods. This suggests in particular that the increased remuneration, which in consequence of the new accumulation of capital has to be paid for the services of labour and other per manent resources, is paid, or "advanced", out of the savings. It can, however, be shown that it will be neither necessary nor profitable for the savers to use their savings to offer increased remuneration to the services of the input newly invested. 2 In fact, if they did The use of savings SO they would find that the transaction to pay Increased remuneration to factors would end in a loss, and that they would neither necessary n()r not even be able to complete the intended profitable investment unless they were prepared to do further saving.

And, as we shall also see, in this case the amounts saved in the first instance would not in any way help to bridge the gap which would occur later in the income stream in consequence of the diversion of current input to invest ment. The actual source of the additional amounts which unquestionably will have to be paid to labour and other resources is to be sought elsewhere than in the savings. 3 1 Cf. his "third fundamental proposition concerning Capital", that, although saved, and the result of saving, it is nevertheless con sumed. (Principle8 of Political Economy, Book I, chap. v, § 5.) 2 Contrast, however, the statement in Bresciani-Turroni, 1936. 3 If unemployed resources exist, and savings make it profitable to employ these resources, then the savings will be used to remunerate these additional resources. In this case current consumption will not be reduced but only redistributed in consequence of the new savings, and here, and only here, the Smith-Mill doctrine about the savings which are also consumed, but by another group of people, does apply.

If) 274 Capitalistic Prod'uction unde1' Competition PT. III We shall begin by considering what is in some respects the simplest, although by no means the most realistic case, where the unexpected saving occurs just once and ElIects 01 a single un-is not expected to be repeated. We shall loreseen act 01 saving assume that the savers decide at a certain moment that they do not want at that time to consume a certain quantity of consumers' goods which they either actually possess or which they would obtain in the normal (stationary) course of events in exchange for some other commodities in their possession. And we shall further assume that the most profitable form of new investment is of a kind which can be completed within a comparatively short period, so that it can be made to serve consumption at the end of that period without requiring any further new investment. An instance which would fit our present simplified model would be the production of some semi-finished product which, once the required amount of it has been produced instead of consumers' goods, will be continually reproduced and can then be used to produce consumers' goods in much the same way as, but much more effectively than, the input from which it was made.

In order to obtain these new capital goods the savers will have to induce resources, which would otherwise have been devoted to the production of consumers' goods, to produce the required capital goods. The use 01 the savings and the redirection of And by so doing they will reduce the outinvestment put of consumers' goods at some future date. (Exactly when this consequent deficiency in the output of consumers' goods will occur will be discussed presently.) But this input would have been remunerated in any case for producing consumers' goods, so that if the savers now want additional capital goods instead of consumers' goods, all they need to do is to offer what remuneration they would have had to offer to obtain con sumers' goods at this date to obtain the capital goods instead. And since in consequence they will obtain less OR. XX The Accumulation of Capital 275 consumers' goods, what they will do is to keep the con sumers' goods they have saved in order to consume them when the current flow of consumers' goods falls off. (We need hardly consider here the exceptional case where input that is to be diverted would have produced con sumers' goods instantaneously. In any event this would certainly apply only to a very small fraction of the input diverted.) There will therefore be no increase in the consumption of those who do not save, and no additional transfer of consumers' goods from those who command them to others. All that happens is that at Wh' d I at IS save s not the earlier date the savers consurne less consumed at the time h h b · f d' it Is saved t an t ey 0 tam rom current pro uctIOn, and at the later date (when current production of con sumers' goods has decreased and additional capital goods are turned out instead) they are able to consume more consumers' goods than they get from current production. 1 We shall explain later why it is that, despite the con stancy of the quantity of consumers' goods which is left for the rest of the communi~y, the share which goes to some of the resources, and particularly the permanent resources, will increase. For the moment it is sufficient to have shown that this is not essential in order to induce input which has previously been producing consumers' goods to produce capital goods instead; for this purpose it is sufficient if the remuneration, which would in any event have been offered for them to produce consumers' goods, is now offered to that input to produce capital goods. It will also be clear that if the savers actually do use their savings to offer additional remuneration to the input which is now to be invested for longer periods, they will find, in the first place, that later on, when the current 1 Whether they actually keep the same batch of goods from the first date to the second, or whether they merely keep a constant amount of carry·over from moment to inoment which is continually renewed, is quite immaterial from the point of view of the general problem.

276 Capitalistic Production under Competition PT. III output of consunlption goods is reduced in consequence of the past investment, they (or perhaps somebody else) will be compelled to repeat the saving. In the second place, they will find that even if they are willing to repeat the saving, there will be no means out of which they can recover the extra expenses that have gone in paying these factors as additional remuneration the whole of what they have saved. The same effect would also occur if the additional unforeseen saving were made by the consumers merely spending less money on buying consumers' goods. Although this might to some extent lead to an accumula tion of unused stocks in the hands of the dealers, it would probably lead to some reduction of price and through it to some increase of consum ption on the part of the people who did not save, and it would again happen that, by the time current output of consumers' goods was reduced, no or insufficient reserves would be left to cover the deficiency and the saving would have to be repeated.

It will be evident by now that savings are actually required only at the time when, in consequence of a past diversion of input from the production of consumers' Savings are usually goods to the production of capital goods, required only some the current output of consumers' goods is time after new Investments have been falling off. And the success of any investstarted ment undertaken at a particular moment of time will generally depend, not on the amount of saving at that time, but on the rate of saving and the consequent price relationship between consumers' goods and capital goods at some date in the future. Saving which occurs unforeseen merely serves as a signal which creates the expectation that the demand for consumers' goods will be less in the future than it was in the past. And it would be quite sufficient if, without any present saving, it became known sufficiently in advance that saving would take place in the future. All this makes it necessary to OH.XX The Accumulation of Capital 277 investigate in somewhat greater detail what wiU be the effect of additional investment on the flow of income.

Oversimplified conceptions of the" period of produc tion" have led to illegitimate conclusions in this sphere as in many others. If it is assumed that the average period of production is a uniform period 1 MIsleadIng elIecls 0 for all the input, and if it is further the Idea 01 a uniform d h 1 h · . d perIod of production assume t at on y t e mvestment peno s of the pure input can be changed, while all the non permanent resources are completely specific (and both of these assumptions have often made their way explicitly or implicitly into the analysis of problems of this kind), then the answer to our question becomes indeed very simple.! If, e.g., the uniform investment period of all the pure input has previously been one year and is now suddenly extended to thirteen months, the effect will evidently be that a year after the change has been made the stream of consumers' goods will entirely cease for one month and will not begin to flow again until the date when the products of the input invested for thirteen months come on the market. If the investment period were lengthened for only part of the total input, we should have a proportionate reduction of the output stream during the corresponding time interval. And if we assume that the investment period, instead of being lengthened with a sudden jerk, is extended gradually and continuously, the effect on the size of the output stream can best be explained by the adjoining Fig. 24, which is a modification of Fig. 5 (see p. 115).

In this diagram the slope of the line OP represents the rate at which pure input is being invested, and the vertical distance between this line and the line TQ the period for which this input is invested. A continuous lengthening of the investment period will then be shown 1 For an analysis based on these assumptions see Bresciani.Turrolli, 1936.

278 Capitalistic Production under Competition PT. III by the Hne TQ becoming steeper than the line OP, as between the points U and V in the diagram, and a con tinuous shortening of the investment period by the line TQ being less steep than OP, as between Wand X. Under stationary conditions, it will be remembered, the output during any given period will always be the product of a quantity of input equal to that which is invested during the same period. While the investment t Q x T4~--------------~~--------~~ T3-----------~~---r------~ v w ~~------*----+----4--X Tl~---~--+_---r--~ u T o p period is being lengt,hened, however, as during the interval TIT 2, it will be seen that the output rnaturing \vill be the product of a quantity of input R 1R 2 , while during the same interval the larger quantity of input R4R6 is being invested. And while the investment period is being shortened, as during the interval T3'114' the output maturing during that interval will be the product of a quantity of input R3R5' which is consider ably larger than the quantity of input R7RS which is being invested during the same interval.

This sort of analysis, which is rarely given even in this much detail, suffers, as has already. been remarked, OR. XX The ACC1lmulation of Capital 279 from two defects in the underlying assumptions. In the first place, what we have to deal with in actual fact is not a single uniform investment period but a range of different investment periods. In the second Defeets of analysIs place, it is always possible to alter the based on this idea investment periods not only of the pure input but also of a great part at least of the" intermediate products" and other nonpermanent resources (i.e. of all the" mixed input "). The consequence of the first of these two facts is twofold. First, since even the total length of the different processes of production is different, the effect of lengthen ing them all at the same moment will not be DiIT i f IT t r us on 0 e ec S 0 felt at one and the same later date, but Investment on the ·11 b d 1 . d f t· output stream WI e sprea over a ong perlO 0 Ime.

Secondly, since every single process requires successive investments spread over a period of time, and since the investments in the later stages of the new' process can obviously be made only after the corresponding invest ments in the earlier stages have already been made, the whole process of lengthening the investment structure will be diffused over a period of time. When the new invest ment first begins to take place, only the input applied at the beginning of the various processes will be invested for longer periods; but people will do this with the intention of changing (and in the expectation that it will be possible tq change) in succeeding periods the investment periods of units of input in the later stages of the same process. The transition from one sort of investment structure to another will therefore make it necessary during a con siderable period of time for input to be transferred from one" stage" of the process to another. In addition, the effect of each of these successive changes will be spread over varying ranges of the output stream according as the new investment periods of the individual units of input are longer or shorter than the older periods.

This diffusion of the effect on the size of the output 280 Capitalistic Production under Competition PT. IH stream caused by changing the investment periods is enhanced by the second fundamental fact, namely, that not only "pure input" but also intermediate products and other nonpermanent resources can be diverted from shorter to longer investment periods (and vice versa). For this reason it is not possible, as one might suppose, to deduce from the comparative shapes of the input functions of the old and the new structure the exact effects on the output stream of any lengthening of the investment structure. If the intermediate products were completely specific and only the investment periods of the "pure input" could be changed, this would indeed be possible, provided that both the aggregate input functions and also the input functions for the separate processes were given. It would then be possible to deduce from these functions what quantities of input would have to be changed over to other investment periods at every stage of the process, and by what amount their investment periods would have to be changed, and this would uniquely determine the effect on the shape of the output stream. Where, however, the intermediate products also can be shifted to other uses, this clearly becomes impossible and we must be satisfied with the following two general conclusions.

First,-the temporary decrease of the output stream con sequent upon any new investment will be spread in some way over the whole period between the moment when the new investment begins and the moment when the new structure is completed. Second, the completion of any new investment structure which is begun at anyone moment of time will usually require further new invest ment throughout the whole period until the new structure is completed. It will be useful now to consider the effects of saving from a different angle. Let us consider what kinds of investments will appear most advantageous to entre preneurs if they expect that in the neal' future there will for some time be a shift of demand away from conCR. XX The Accumulatl:on of Capital 281 sumers' goods and towards capital goods generally. We shall assume that the entrepreneurs regard a present increase of saving as an indication that, for a consider able time to come, saving and investing The elleets of foreseen will continue at the same rate as at savings on the plans present. I am deliberately speaking here of entrepreneurs of entrepreneurs in general, and not merely of savers or of those people who have the newly saved funds to invest, because, as we shall see, the decisions of all entrepreneurs will be affected by the expectation of fut_ure savings.

Since it is assumed that the shift of demand is foreseen and that there is in consequence time to adjust supplies, there is no reason to conclude that there will immediately be a corresponding change in relative prices. And it will be less question-begging if we describe the general situa tion by saying that the expected demand curves for consumers' goods and capital goods in general are being shifted, the former to the left and the latter to the right. The main question which arises is the following: Towards what kind of capital goods will the new demand be directed, and how will the relative prices of the different resources, and the quantities produced, be The mechanism of affected by this change in demand? In the redirection of inabstract terms the situation can be de-vestment scribed by saying that a greater potential command over consumers' goods available in the comparatively near future is offered in exchange for means to obtain con sumers' goods in the more distant future. And it is reasonable to expect this to induce an expansion of pro duction for that more distant future. Our task is to show the mechanism by which such a redirection of resources, if it actually does take place, is brought about.

The fact that people in general anticipate a shift to the right in the future demand curve for capital goods means that they expect to be able to sell at that future date, at given prices, a greater quantity of intermediate products than would have been possible under the old 282 Capitalistic Production under Competition PT. III conditions. But a greater quantity of intermediate pro ducts can be provided for that date only by spreading present investments over a wider range of future maturity dates. This clearly means that investment all round will be pushed on to a " contour line " of lower marginal returns over costs, i.e. all individual investments, whether they Equallsatlon 01 all are expected to mature before or after the relurns from lovesl-date at which the change in relative demand monl allho new lower rate is expected, will be extended to a point where the "rate of increase of the produce divided by the produce" is again uniform for all investments but lower than it was before.! And since even before the date at which the change in demand is expected, the returns on some investments will be lowered, competition will so adjust the prices of all the resources invested as to make the return on them the same in all uses.

It will be observed that all this happens before the new savings actually become available. It takes place solely as a result of the expected shift in demand. The capital which is used in the first instance to finance these investment changes, and which has its returns reduced, is old capital which was previously used to produce con sumers' goods and which is now used to produce capital goods for the date concerned. The new capital (the savings) will not be required until, in consequence of these changes, less consumers' goods come on to the market. We shall presently discuss certain exceptions to this general rule which are due to the fact that in many cases the investment structures cannot be varied continuously, and that, in consequence of changes in the productive tech nique used and the composition of the stock of capital employed in a particular line of industry, they will change 1 So far as concerns the production of consumers' goods for the date at which the shift in demand is expected to take place, the reduc tion of the output of consumers' goods for that date will cause the return on investment to be actually larger than it would have been if the change in demand had occurred without any change in output.

CR. XX The Accumulation of Capital 283 only discontinuously. For present purposes it is sufficient to point out that what can be reinvested in the new and different form will be only the current pure input and the more versatile nonpermanent re-Etl 1 I • ect 0 nvestmen. sources. Such capital, on the other hand, on value 01 speclDc as was irrevocably sunk, before the new resonree. saving was foreseen, in very durable and highly specnic equipment, cannot, of course, be promptly or wholly shifted to a different use. The return obtained from it will be reduced by the rise in the prices of the services of the resources which are needed to co-operate with it, and in consequence it will not be profitable to reproduce these instruments; and in some cases it will be found that even in the course of time only part of the capital origin ally invested can be recovered and reinvested III a different form and that the rest has been lost.

Any additional investment (even while it is only expected) will thus cause a shrinkage in the return on all capital. This happens not because the capital has an unchanged absolute return which is dis-The source of the In counted at a lower rate of interest thus creased re~UJleration , 01 the servICes 01 the reducing the yield per cent, but because the permanent resources magnitude of the share of the product going to the old capital is teduced. This fact also provides the answer to the question of the source out of which the increased remuneration of the services of the permanent resources is paid. It is not out of the new savings but out of the share of the product which used to go to the owners of the old capital, that this additional remuneration is provided. It has been shown that, as soon as the marginal returns which can be obtained from investment fall in response to the changed relative demand, competition between capitalists will tend to drive up the price of labour and the other services of the permanent resources, at the expense of the capitalists' own income. And this rise in the prices of these services also explains why an (expected) increase in the demand for capital goods (in terms of 284 Capitalistic Production under Competition PT. III potential consumers' goods), and the consequent fall in the rate of interest, has the effect not only of making methods of production profitable which were previously unprofit able, but also of making methods of production un profitable which were previously profitable.