materials, goes from hand to hand, migrates from industry to industry and from country to country in a manner that suggests that he was thinking of balances and not of goods. It may be replied, of course, that sums of money may be thought of as representing goods and also that, especially within the theoretical organon of that time, monetary processes may be reduced to ‘real’ processes in the last analysis. But such reductions are dangerous short cuts at best and involve neglecting monetary mechanisms that give rise to many essential problems. Thus, even if the fundamentals of the economic process could in fact be satisfactorily described in real terms, we should still have to observe that, in capital theory as elsewhere, the ‘classic’ attempt to carry out fundamental analysis in real terms only was seriously at fault. over, like our initial Stock, it was a stock of goods. But, unlike our initial stock, it did not include all goods existing at an instant of time. From these the ‘classics’ separated out their capital by excluding from it, first, natural agents (though not ‘improvements’ such as drains, fences, and the like) and, second, all consumers’ goods other than the means of subsistence of productive labor. Let us stop for a moment in order to make this clear. First of all, it should be understood that the division of the stock of wealth existing at an instant of time into a mass of things that are capital and another mass of things that are not is a device for describing what we have called above the structure of, or the structural relations within, the universe of goods. The second thing to observe is that the effect of excluding natural agents was to set up, in addition to labor, another ‘original’ factor of production, though many, especially the Ricardians, failed to realize the implications of it. This, then, would have left us with the stock of produced goods. But, third, the structure of the mass of these produced goods was further analyzed by means of a distinction that warred between two of an indefinite number of possibilities. On the one hand, if we want to separate out that part which is a requisite of production in the technological sense, we arrive at the concept of produced means of production or, as it was to be called by the Böhm-Bawerkians, of intermediate products. It is, however, one of the salient characteristics of the theoretical schema of the English ‘classics’ and their continental followers that they understood the term ‘requisite of production’ in a wider sense which included the consumers’ goods that support labor during the process of production. There is no logical reason for not including also the consumers’ goods that support landowners—Senior even included those that support capitalists—during that process but, actually, these were mostly excluded by the Ricardians because their schema prevented them from considering rent as an element of cost. On the other hand, if we want to separate out that part of the mass of wealth existing at an instant of time which is in the business process or serves business purposes—or, as A.Smith put it, from which ‘profit is ex-pected’—we are led to include besides plant, equipment, raw materials, and the ‘means of subsistence of productive labor,’ also other items, in particular the following two. One is another mass of consumers’ goods—which in part overlaps with the one included among requisites of production—namely, the mass of consumers’ goods that is still in the hands of manufacturers, wholesalers, and retailers irrespective of who (laborer or capitalist) is going to buy them. The other is cash in hand. The implications of this, though not uninteresting, cannot be considered here. All that can be said is that this distinction is not more right or wrong than is the other. Both serve relevant analytic purposes; that is, both are useful in describing relevant aspects of reality. But we shall keep to the first one (‘requisite of production’ in the wider sense) because it is more germane to what I have called above the dominant analytic purpose of History of economic analysis 602 the epoch, especially of the work that was summed up by J.S.Mill. Marx would have approved of our choice. He was all for the first distinction. The second he considered to be incapable of serving any purpose other than that of copying the surface of reality as it appears to the capitalist. What has been said above, barring details, reproduces the actual way in which A.Smith ‘structured’ what he called the ‘general stock of any country or society’ by separating out capital (Wealth, Book II, ch. 1) and by enumerating its chief components. It matters little that he (and Malthus) did not specifically include wage goods or the means of subsistence of labor. For he always argued as if he had included them. 21 Also, the capital concept described represents fairly well the wording of most of the leaders. Thus, Ricardo defined the concept: ‘Capital is that part of the wealth [my italics] of a country which is employed in production, and consists of food, clothing, tools, raw materials, machinery, etc., necessary to give effect to labour’ (Principles, ch. 5). This does not differ in essence from Senior’s definition: ‘an article of wealth, the result of human exertion [meaning, as he explained a few lines further on, ‘of labour, abstinence, and the agency of nature,’ or simply, a produced article of wealth] employed in the production or distribution of wealth.’ Nor does it differ from J.S.Mill’s influential passage: ‘What capital does for production, is to afford the shelter, protection, tools, and materials which the work requires, and to feed and otherwise maintain the labourers during the process… Whatever things are destined for this use…are Capital’ (Book I, ch. 4, 1). 22 Marx added nothing to this except that, in obedience to his principle of amalgamating economics and sociology, he confined the term capital to those things of this class that are owned by capitalists—the same things in the hand of the workman who uses them are not capital. ‘What capital does for production,’ however, means two very different things, and the distinction between wage capital and the rest—we shall call it technological capital— readily suggests itself and so does a coefficient, descriptive of the quantitative relation between the two, which must obviously constitute one of the most important characteristics of the structure of capital. Nevertheless, it was left to Marx to point this out in so many words and introduce such a coefficient explicitly. Denoting by the term constant capital (c) what has just been called technological capital, and by the term variable capital (ν) 23 what has just been called wage capital, he chose for structural coefficient the ratio: which he called the Organic Composition of Capital. 24 The merit there is in introducing such a concept explicitly must not be underrated. But, of course, the writers from A.Smith to J.S. Mill had not failed to recognize the peculiar role of wage capital within total capital. This is abundantly indicated by the fact that wage capital is identical not only with Marx’s variable capital, but also with the ‘classic’ wage fund. Moreover, both Ricardo and Mill sometimes used the Marxist concept inadvertently: they sometimes meant variable capital when they actually wrote Circulating Capital. 25 21 See, e.g., Wealth, p. 316 [Modern Library ed.]. No additional comments are needed on A.Smith’s inclusion of ‘the acquired and useful abilities of all the…members of the society,’ a precedent that General economics 603 was widely followed: Roscher even included ‘virtue.’ For this remained entirely inoperative. Note, however, the parallelism with ‘the improvements of land’ that may have suggested the Marshallian concept of quasi-rent. 22 The merely verbal concession to the labor theory of value (pointed out above) that is contained in this formulation could be easily rectified. 23 He expressed both in terms of labor embodied. But inasmuch as his coefficient always refers to a given point of time, expression in money values does just as well. It should be observed that both measures carry full meaning and in particular allow of intertemporal comparison only for states of perfect equilibrium. The reason why Marx chose those terms (and why we cannot accept them here) is that, in his theory, technological capital transmits to the product only its own value—or that, in the productive process, its value in terms of labor embodied remains constant—whereas wage capital, as it were, swells in the process by virtue of the labor hours that workers add to the labor embodied in it. 24 Thus, since the labor-hour dimension is present in both numerator and denominator and therefore cancels out, this coefficient is a pure number. Still it is worth while to remember that the components of the coefficient are values and not physical quantities. 25 Thus, Ricardo wrote in 4 of ch. 1 of the Principles: ‘In one trade very little capital may be employed as circulating capital, that is to say [my italics], in the support of labour…’ Other instances are found in ch. 31. Not less obvious is the necessity of analyzing the internal structure of technological capital. It was quite clear to the physiocrats, whose various avances A.Smith replaced by the distinction between Fixed and Circulating Capital. The former he defined as capital from which the owner derives profit by keeping (using) it, such as factory buildings and machines; the latter he defined as capital from which the owner derives profit by ‘parting with it’ (turning it over), such as raw materials. Ricardo saw that there was something of deeper significance behind A.Smith’s common-sense and commonplace distinction, which accordingly he brushed aside. 26 Let us make an attempt to reconstruct his thought. Evidently, Ricardo’s attention was drawn to the problems of fixed capital by the fact that its presence causes exchange values of products to diverge from the labor-quantity law unless, of course, all the branches of industry employ ‘the same proportion of fixed and circulating capital.’ He also perceived, apparently without difficulty, the further fact that, in order not to disturb this labor-quantity law, this fixed capital would in addition have to be of the same durability everywhere. Finally, however, he perceived something else, namely, the analogy that exists between different durabilities of the fixed capital 26 See Principles, ch. 1, sec. 4, first footnote. This footnote—saying as it does that the distinction between fixed and circulating capital is not essential—reads strangely in view of the fact that the whole argument of Section 4 turns upon problems of fixed capital. But our reconstruction of his thought will make it clear that he meant not more than that A.Smith’s distinction failed to bring out the essential point involved. History of economic analysis 604 used in different lines of production and the different rates of turnover of different kinds of circulating capital, such as the farmer’s seed and the baker’s flour. This, then, makes three apparently different facts that have, at first sight, nothing in common except that they all interfere with the operation of the labor-quantity law of value. Now, by what almost amounted to a flash of genius, he saw that all three of them did so for the same reason or, to put it differently, he saw the same fundamental element in all of them, namely, the time distance between investment and the emergence of the corresponding consumers’ good. 27 This was very easy to see in the case of differences in periods of turnover: wheat that is being used as seed and wheat that is being worked up into flour directly differ (from Ricardo’s standpoint) by the time distances between each of them and the emergence of flour and by nothing else. But it was not so easy to see that the difference which the presence of fixed capital goods and of fixed capital goods of different durability makes to the process of production, and hence to values, is of the same kind in that it also may be looked upon as a matter of differences in these time distances or rates of turnover. Consider, for example, a machine that, in the Ricardian manner, has been produced by labor alone, say, in a single day. Suppose it is to last ten years. During these ten years, the machine or the labor embodied in it matures into consumers’ goods exactly as will raw materials or semifinished products. Each of the days of service ‘contained’ in it— becoming available in a definite sequence—behaves like the seed in the ground until its turn comes. This definite sequence is a restriction upon economic decision or action, analogous to the restriction placed upon the farmer’s decision by the circumstance that he must wait until his seed matures into a crop. There is thus in fact, at least on the most abstract level, no essential difference and no clear line of demarcation between fixed and circulating capital, as Ricardo pointed out in the footnote referred to above. Both are nothing but immature (elements of) consumers’ goods—intermediate products or ‘inchoate wealth’ as Taussig was to call them about eighty years later. Or both may be ‘resolved’ into hoarded labor—James Mill’s term, which expresses Ricardo’s meaning very well and was to be used again by Wicksell, also about eighty years later 28 —though we must not forget that the various agglomerations of hoarded labor embodied in the various goods carry different indices of time distance or indices of places in the time sequences to which they belong. 27 I call this a flash of genius without necessarily committing myself to the theory of capital that developed from this flash. In order to clarify a very important piece of doctrinal history, I refrain entirely from criticism at this point. Professor Knight himself, who rejects the theory in question, could therefore accept my exposition just as could have F.W.Taussig, who made it his own. The essential thing, for the moment, is to see the relation between Ricardo’s analysis and Böhm- Bawerk’s, which has been emphasized by both Knight and Taussig. 28 More precisely Wicksell said saved-up services of labor and land, and should have added the services of previous accumulations. General economics 605 Thus, Ricardo’s rudimentary capital analysis issues in a time concept of technological capital, 29 time being the element that unifies all its specific forms. Those who entertain sympathies for the labor-quantity theory of value might claim with some justice that he thereby saved the latter by making it valid (to some extent at least) for labor quantities that carry different time indices. Those who accept Böhm-Bawerk’s theory of capital might claim, also with some justice, that Ricardo worked up a bad theory of value into a good theory of capital. In any case, Ricardo was clearly Böhm-Bawerk’s forerunner so far as this set of problems is concerned. This is not to say that Ricardo’s theory of capital was complete or that he saw all the implications of his flash of genius. In particular, he neglected all its short-run implications. 30 Also, though he did study cases of conversion of circulating into fixed capital—the most important case is to be found in his chapter On Machinery—and occasionally brushed against the manifold relations of substitution that exist within the universe of technological capital—the ‘Ricardo effect’ if thought through would afford an instance—he was, as were most of the ‘classics,’ too much given to accepting time sequences as technological data and to neglecting the fact that durabilities, and, in general, the relations between the quantities of capital goods of different types, also the relation between wage capital and non-wage capital, are economic variables that depend, and in turn react, upon wage rates, efficiency of labor, rate of interest, and other factors. But this is only another way of restating the fact that his theory was no more than a preliminary sketch—a fact that must always be borne in mind, just as much when we are criticizing as when we are defending his performance. (c) Senior’s Contributions. Two most curious facts now call for our attention. On the one hand, Senior realized that Ricardo used the terms fixed and circulating capital in a sense that differed from A.Smith’s use (Outline, pp. 62–3). But the true meaning of Ricardo’s capital analysis escaped him so completely that he saw nothing in that difference except a reprehensibly unusual use of terms. On the other hand, in spite of this failure to understand the Ricardian analysis, he actually carried it further in two directions—an excellent example of the way in which we blunder along. First, there is Senior’s third postulate or elementary proposition, which reads: ‘That the powers of Labour, and of the other instruments which produce wealth, may be indefinitely increased by using their Products as the means of further Production.’ This proposition, which might have been derived from Rae, improves Ricardo’s theory by adding the powers ‘of the other instruments which produce wealth’ to those of labor. But it adds also something else that lies entirely beyond Ricardo’s analysis. With Ricardo, the time element comes in to cause deviations of values from the labor-quantity law by putting a brake on the supply of the products of the capitals that turn over more slowly than others: the man whose products take relatively long spans of time to reach their markets simply ‘must’ be compensated for this disadvantage. According to Senior, 29 If the line of thought, for which he offered but fragments, were carried on to its logical consequences, we might even say that his analysis comprised technological plus wage capital, resolving all physical capital into wage capital or rather into a general subsistence fund. This idea is still more clearly indicated in James Mill’s Elements. Nevertheless, its explicit recognition belongs History of economic analysis 606 to Jevons and Böhm-Bawerk, though implicitly it is also present in one of the various aspects of the wage-fund theory. 30 Such as that fixed capital, in the short run, behaves like ‘land.’ however, there is more to this higher value of such products than the mere fact, if it is a fact, that a profit of £100 accruing every second year is not economically equivalent to a profit of £50 accruing every year. The profit from the two-year investment will be more than twice the profit from two successive one-year investments of the same quantity of (say) labor, because the productive ‘power’ of this labor, hence its product, is increased if the product of the first year be used ‘as the means of further production’ in the second year. Now, Ricardo’s real value of a product cannot increase simply because the same quantity of labor produces a greater quantity of product in the two-year process than it would in two successive one-year processes. But Senior’s value may. 31 This puts an entirely new face upon the matter and points straight ahead toward Böhm-Bawerk—who, by the way, did not understand Senior any better than Senior did Ricardo, but who nevertheless carried on Senior’s capital analysis exactly as Senior carried on Ricardo’s. The relation that exists on this point between Senior and Böhm-Bawerk will stand out particularly if we observe that using a product as the means of further production might well be called using it in a ‘roundabout’ way. The only difference is that Senior confined himself to stating that the productive power of labor is increased ‘indefinitely’ by using it in that way; whereas Böhm-Bawerk added the hypothesis that the rate of this increase decreases as the ‘length’ of the productive process increases. Second, there is Senior’s abstinence theory of capital. Though his name is chiefly remembered because of this contribution, it (abstinence) is actually much less important, as an analytic performance, than is the contribution just discussed (using a product as the means of further production). It will be as well to distinguish two different aspects of Senior’s abstinence. On the one hand, if for reasons good or bad, we choose to analyze the structure of technological capital in terms of what we have called the time indices of its com-ponent elements, the fact we wish to emphasize is that these component elements (that is, the various capital goods) have different rates of turnover or that their products become available or ‘mature’ after lapses of time of different lengths, which somehow or other intrude into the list of costs of production. Whenever we mean this, we had better use the term Waiting that was to be suggested later on by McVane and to be adopted by Marshall. On the other hand, if we adopt the theory that technological capital is the result of a ‘conversion of revenue’ into something that is expected to yield revenue in the future, but, in order to do so, has to be withdrawn from the sphere of revenue for good, then we had better use the term Abstinence. In this case we reserve the term for the psychic cost of saving or, if we keep saving sufficiently close to investment, of the capital goods in which past saving has been invested. This element of psychic cost then becomes 31 No difficulty arises from the fact that the value of the greater quantity of product that results from the two-year process need not have a greater value than the smaller quantity that results from two successive one-year processes. For in this case the two-year process will not be used. General economics 607 analogous to the ‘psychic cost’ of labor, later on called Disutility. Going a step further, we can constitute abstinence itself—instead of the saving or the capital goods that result from it—as a factor of production. 32 This is the sense in which Senior’s abstinence has been commonly understood and in which the phrase will be used in this book, although his own definition shows that he meant to comprise in his concept also what above has been called Waiting. 33 Recognition of what the term abstinence—in the strict sense—is intended to denote is, of course, as old as is the recognition of the role of saving. A. Smith’s parsimony or frugality means nothing else. Practically all the economists who wrote after 1776 deal with it in one way or another, though not all were prepared to accord to it everything that A.Smith claimed for it. It also entered the theoretical schemata of anti-savers like Lauderdale and Malthus. Ricardo’s schema considers waiting rather than abstinence but in any case this schema requires, as our exposition amply shows, a conceptual complement of this kind. Actually, however, the concept was formally established by Read and especially by Scrope. The latter stands to Senior in this respect in the same position as did Rae in respect to the postulate about increasing the productive powers of factors by using their product for further production. No reflection on Senior’s subjective originality is intended, but it is important to note that objectively Senior did no more than is implied in bringing an existing doctrinal tendency to a head. Propelled by the sponsorship of J.S.Mill, Cairnes, and, to some extent, Marshall, abstinence analysis became firmly established in English economics, though it was never so popular elsewhere. It is not difficult to guess the reason why the spearheads of the attack upon it are to be found in the writings of Marx and Lassalle, who saw nothing but the apologetic possibilities that the word Abstinence suggests. But this will be more conveniently dealt with below under the heading of profits. 32 It does not greatly matter whether we call this factor a ‘primary’ or a ‘secondary’ one. Senior did the latter but later tendencies favored the former. 33 Senior defined abstinence as ‘the conduct of a person who either abstains from the unproductive use of what he can command, or designedly prefers the production of remote to that of immediate results’ (Outline, p. 58). The first alone denotes saving or conversion of revenue into capital and is abstinence stricto sensu; the second means only rearrangement within the capital structure and is what we mean by waiting. Senior was evidently aware of the distinction, the validity of which is not impaired by the possibility of reducing one to the other by appropriate wording. History of economic analysis 608 (d) J.S.Mill’s Fundamental Propositions Respecting Capital. Some additional points about the ‘classic’ theory of capital may be conveniently made, and others may be restated by way of comment on the four connected ‘propositions respecting capital’ that J.S.Mill presented in Chapter 5 of Book I of his Principles. 34 ‘The first of these propositions is, That industry is limited by capital,’ though of course industry does not always work up to this limit. Total employment of labor is not so limited, however, since it also proceeds from ‘revenue.’ J.S. Mill believed erroneously that ‘industry’s’ being limited by capital 35 implies that ‘every increase of capital…is capable of giving additional employment to industry; and this without any assignable limit’ ( 3). Carefully stated (and the ‘capable’ properly emphasized), this can be shown to be true, and the use of this proposition against the opinions of Malthus, Chalmers, and Sismondi 36 —who asserted that ‘wealth’ is at any time limited not only by productive power but also by the system’s capacity to consume—is perfectly legitimate. Only it should have been stated as an additional proposition—we might call it the Theorem of Hitchlessness—for it does not follow from industry’s being limited by capital and, though Mill’s argument against those three authors was successful as far as it went, it falls far short of proving this theorem. Moreover, the theorem is interesting only if it be made to hold for the total of technological plus wage capital. But Mill restricted it to the latter so that the proposition he meant to defend amounts only to this: ‘the portion which is destined to their [the laborers’] maintenance, may (supposing no alteration in anything else) be indefinitely increased, without creating an impossibility of finding them employment’ 37 —which is either trivial or 34 This chapter, though disfigured by slips, inadequacies, and clumsinesses, is not without logical beauty. The reader who knows how to mend these shortcomings will be impressed by its symphonic qualities. 35 At the end of this chapter’s section 1, Mill uses this proposition to combat what he considered a popular fallacy about the effects of a protective tariff. The reader’s attention is called to this curious mixture of truth and error, which exemplifies excellently a situation that occurs so frequently in economic arguments: an incontestable truism is used in an inadmissible manner so as to produce a result that should be false and nevertheless is not (wholly) so, because elements of truth extraneous to the logic of the argument may be invoked in order to peg it. Unfortunately, considerations of space forbid us to explain this fully. Mill was no trickster. But the passage nevertheless exemplifies a well-known trick, namely, the trick of making some politically relevant result follow apparently from an obvious truth so that the political opponent is subtly put into a position which it is implied only an absolute fool could take. 36 Mill failed to add Barton. 37 In the same paragraph Mill inadvertently used a nonphysical capital concept when he spoke of the other ‘portion’ of capital that is fixed in machinery, buildings, and the like. Capital that is fixed in machines instead of consisting in machines cannot be capital in the sense of his own definition. General economics 609 false. It is a very interesting question to ask why he should have thus maimed a theorem that was certainly not beyond his range of vision. 38 The answer cannot be that in the short run technological capital is an assemblage of specific goods, the kinds and quantities of which constitute data. For Mill evidently did not intend to write a treatise on short-run analysis. Rather, the answer seems to be that, while of course he was not unaware of the fact that the relation between technological and wage capital is variable, he was inclined on principle—that is, when arguing matters of fundamental principle—to take it for granted, perhaps as technologically fixed, and to neglect the substitutability between the two, the nature and importance of which, though emphasized by Barton and Longfield, were hardly clear to him. This is why he found it so easy (as had Ricardo) to follow the example set by A.Smith and to speak of a ‘portion of capital’ or of a fund that is ‘destined for the maintenance of labour,’ that is, to speak of a wage fund. Let us observe at once that one of the most characteristic features of the so-called wage-fund theory, namely, the implied assumption or at least suggestion that this fund is a sort of datum, thus rests on nothing but primitive technique. 39 We take another step toward understanding the wage-fund theory when we consider Mill’s ‘second fundamental theorem respecting Capital [which] relates to the source from which it is derived’ and which makes capital ‘the result of saving’ ( 4): capital increases by revenue’s being converted into it. We know already (see above, ch. 5, sec. 6) that the ‘classic’ schema of economic evolution was at fault both through overrating the importance of mere increase in the items that constitute capital and through overrating the role that (voluntary) saving plays in this increase. Also, the ‘classics’ in their anxiety to emphasize the fundamental meanings of economic mechanisms, were given to drawing together much too closely decisions to save and decisions to invest. These decisions, though never quite identified, 40 tended to shade off into each other to the exclusion of all that may intervene between them, 41 38 This theorem should read that on the long-run trend (that is to say, neglecting the effect of temporary disturbances) there are no assignable limits to investment opportunity at appropriately falling rates of interest, except possibly, institutional ones. Lauderdale, Malthus, Storch, and others denied this, but Mill, of course, accepted it, and there was no good reason for him not to put it in so many words, especially since it had already been stated by James Mill. 39 It is not true, however, that Mill’s refutation of Malthus’ argument depends upon this or on any other part of the wage-fund doctrine, as Lord Keynes seems to have believed (General Theory, p. 364). 40 What they may indeed be said to have identified is (the schedule of) savings and (the schedule of) loanable funds: what was saved was ipso facto made available for real investment either in the saver’s business or in someone else’s except in times of deep depression; and for them there was no other source of loanable funds except saving—the money created by bank credit never being taken into account when fundamental principles were under discussion. 41 Malthus, on this point, fully shared the prevailing view. See, e.g., his turn of phrase: ‘parsimony or [my italics] the conversion of revenue into capital…’ (Principles, p. 369n.). History of economic analysis 610 so that saving unconditionally enriches and spending unconditionally impoverishes both individuals and nations. Like Say, Mill reasserted all this; in other words, he reasserted— with added emphasis even—the Turgot-Smith theory of capital formation. 42 But then how was it that he was able to hold (as he evidently did)—also in conformity with dominant tradition—that saving, and saving alone, invariably increases not total capital only but also wage capital, the wage fund, ‘without assignable limit’? No difficulty arises immediately, because the fixed capital, to the production of which the labor is assigned that is compensated from the new addition to savings, must be first produced. Provided that acts of investment follow sufficiently closely upon the decision to save, it is indeed true, at that first turn of the wheel, that demand for productive services—and let us grant their reduction to labor services only—is immediately increased to the full amount of the new addition to savings. That is, the wage fund is increased so as to give, to this amount, ‘to labour either additional employment, or [an increase in wage rates, amounting to] additional remuneration,’ 43 which means either a larger aggregate produce, if the laborers in question have been unemployed before, or a larger share of labor in the same ‘aggregate produce,’ if they have to be drawn from other employments. But, so far as new technological capital results from this employment, things may obviously look very different when all the adjustments have taken place. We may be faced by a different ‘organic composition of capital,’ possibly even an absolute decrease in the variable capital or wage fund. Again, neglecting the possibility that Mill was thinking of short-run effects only, we once more fall back upon the explanation offered before: like all the ‘classical’ leaders, he took the relation between technological and wage capital as a datum, so that in the final result saving would increase both of them in the same proportion. If this were so, 44 then and only then could we speak of a wage fund in any sense other than that the sum total of wage incomes is uniquely determined under the same conditions as is any other economic quantity, for example, the sum total ‘destined’ for the purchase of motor cars. Replacement of workers by machinery was, of course, not overlooked. But it was treated, except 42 This must, however, be taken cum grano salis. Mill admits freely ( 2) that many persons are maintained from capital (not from the return to capital) who produce nothing and that there is such a thing as unproductive consumption by productive workers. If saving provides capital that goes into unproductive consumption, it cannot be held, without qualification, that it ‘enriches’ society or, for that matter, that it is all but synonymous with expenditure on maintaining and aiding productive labor. 43 Which on Mill’s own showing would in general mean unproductive employment of part of the new savings; see preceding footnote. 44 Strictly it can never be so. For savings, other things being equal, must affect the rate of interest and the rate of interest must affect the rate of turnover of capital, i.e., the relation of technological capital and wage fund, as well as the structure of the former, rare cases excepted in which rates of turnover are actually uniquely determined by technological necessity such as that of the period that must elapse between sowing and harvesting. General economics 611 . essential point involved. History of economic analysis 604 used in different lines of production and the different rates of turnover of different kinds of circulating capital, such as. value of such products than the mere fact, if it is a fact, that a profit of £100 accruing every second year is not economically equivalent to a profit of £50 accruing every year. The profit. its explicit recognition belongs History of economic analysis 606 to Jevons and Böhm-Bawerk, though implicitly it is also present in one of the various aspects of the wage-fund theory. 30 Such