Maurice Dobb,
Theories of Value and Distribution since Adam Smith
Ideology and Economic Theory
Cambridge: Cambridge University Press, 1972.

Some Excerpts

Excerpts from Chapter I: Introductory: On Ideology

... theoretical analysis, at any rate in a social theory like economics, inevitably has a causal story to tell. Different types of causal story may have very different implications for what it is possible to do and to achieve by way of policy and social action; thus it is relevant, indeed crucial, for establishing what alternatives are viable - whether, indeed, there is any viable alternative to the existing socio-economic framework - and this entirely within the bounds of 'positive' and non-normative discourse.

... crucial for this is the contrast between theories that approach the determination of prices, or the relations of exchange, through and by means of conditions of production (costs, input-coefficients and the like) and those that approach it primarily from the side of demand.

Undoubtedly this latter has been the principal and distinguishing contrast between the two main and rival systems of economic thought in the nineteenth century and since;

.....the contrast goes deeper than at first sight would appear because, as we shall see, it involves a difference in the 'frontiers' of the object, or in the factors and influences included within the circle of relevant influences or determining factors. For the classical economists, and especially for Marx, the study of Political Economy and the analysis of exchange-value necessarily started from those socio-economic conditions that shaped the class relations of society.

..."Ricardo saw "the laws which regulate" distribution as "the principal problem in Political Economy" since these explained the principles according to which "the product of the earth is divided among three classes of the community, namely the proprietor of the land, the owner of the stock or capital necessary for its cultivation, and the labourers by whose industry it is cultivated ". One could say that for them Political Economy was a theory of distribution before it was a theory of exchange-value..."

.."Marx always emphasised the distinction between the essential processes and relations in human society and the realm of appearances; identifying exchange, or commodity-money circulation, with the latter, and social relations of production with the former...

Concentration of attention upon exchange per se isolated from its socio-historical setting, was the source of 'false consciousness' and delusive theorising. He said in his polemic against Proudhon: "In principle there is not exchange of products, but exchange of labours which compete in production. It is on the mode of exchange of productive forces that the mode of exchange of products depends." The same idea recurs in his reference to "fetishism of commodities" in Das Kapital: "A definite social relation between men assumes in their eyes the fantastic form of a relation between things"; and again in his Theorien über den Mehrwert (speaking of post-Ricardian vulgärökonomie): "the existence of the revenue, as it appears on the surface, is separated from its inner relations and from all connections. Thus land becomes the source of rent, capital the source of profit, and labour the source of wages. The boundaries of the subject as he drew them were accordingly not arbitrary: they were regarded, consistently with his interpretation of historical development, as necessary for embracing all the factors necessary to any explanation that was to be both complete and substantial.

By contrast with this approach the methodology ushered in by the 'Jevonian revolution' and given more systematic formulation by Menger and the 'Austrian school' sought to derive an explanation of exchange-value from the attitudes of individual consumers towards commodities as use-values catering for the satisfaction of individual wants. The significance of this is not simply (as has commonly been the focus of attention) that emphasis is laid at the opposite end of a chain of interdependent events or processes, but consists rather of two crucial consequences of this approach.

First, it treated individuals, their structure of wants and the choices and substitutions resulting therefrom, as the ultimate and independent data of the economic problem: these were the ultimate atoms of the exchange process and of market-behaviour, beyond which analysis did not go (e.g. it did not, indeed could not, concern itself with the social conditioning or social interdependence of individuals' desires and behaviour-reactions).

Secondly, it derived a theory of distribution as incidental to the pricing-process as the pricing of 'original factors' or productive services according to the role they played in the creation of commodities which at first or second remove were of utility to ultimate consumers. In Menger's conception, as we shall see later, there was a simple hierarchy of 'goods of first order' and 'goods of higher order'; the values of the latter being dependent on the former in a simple manner according to their role in the unidirectional process whereby goods or services of 'higher order' were productively transformed into final consumers' goods and use-values. This, and not the use of the formal device of marginal increments, was the crux of the new trend of the last quarter of the nineteenth century (a reason why the designation of 'marginalism' to describe the latter is misplaced).

Anticipating discussion in later chapters: it is perhaps deserving of incidental mention that this inclusion of a theory of distribution within the theory of the pricing-process, as a constituent set of the larger set of market processes viewed as an interrelated whole, is question-begging in an important respect. A structure of market demands can only be derived from consumers' desires, preferences or behaviour-reactions on the assumption of consumers being equipped with a given money-income. Hence an initial distribution of income between individuals is implicit in the general pricing-process, in the sense that it must be included as one of the determinants of the structure of demand from which all prices (including prices of productive factors) are derived: the whole pricing process being relative to this postulated distribution. A theory of distribution, in other words, if it is conceived as a theory of derived prices of productive services or factors, cannot be independent of initial income-distribution as essential premise.

The important, but much too-seldom noticed, consequence of this contrast between the classical approach and the modern follows directly from what has been said: namely, that in the former income-distribution is treated as being the result of social institutions (e.g. property-ownership) and social relations, whereas in the latter it is determined by the conditions of exchange. In the one case it is determined from outside and in the other case from inside the process of market prices (Marx would have expressed the former by saying that social conditions and class forces were more fundamental than relations of echange).

Classically, income-distribution (e.g. the wage-profit ratio) was a pre-condition of the formation of relative prices. Per contra, in post-Jevonian and Austrian theory income-distribution is derived as part of the general pricing process - as a constituent set of equations in the total equational system of market equilibrium (although not without circularity, as we have seen, to the extent that an initial distribution of income has to be assumed for the translation of consumers' wants or preferences into terms of market demand). Thus income-distribution is made to appear as something independent of property institutions and of social relations: as something supra-institutional and supra-historical so far, at least, as income-distribution between factors is concerned. We shall see later that this is the substance and essence of the critique of the theory of marginal productivity in modern discussion (the polemic against the so-called 'neo-classical' school); although the discussion itself has been mainly concerned with formal matters (of consistency and the like). As one writer has recently put it: "The theory of production relations was meant to be independent of the institutions of society; that is, relations between men were treated as irrelevant for an explanation of distribution. It was Marx's insight that this separation is invalid, even in the world of pure logic, and the significance of this distinction for the case of more than one captial good has been emphasised by the modern critics of the neo-classical parables."

Excerpt from Chapter 9
on Pierro Sraffa's 'Production of Commodities by Means of Commodities'

Revival of interest in the classical economists of the Ricardian age may well have had a joint parentage in the shift of focus towards 'macroscopic' questions of which we have already spoken and the new light on Ricardian doctrine represented by Piero Sraffa's edition of the Works and Correspondence of David Ricardo of the early 1950s (especially the novel interpretations of Ricardo's theory of profit and of his search for an invariable measure of value, modestly embodied in the editor's General Introduction).

...Apart from its special corollaries, what is particularly striking (some might say revolutionary) about the Sraffa-system viewed as a whole is its rehabilitation of the Ricardo-Marx approach to problems of value and of distribution from the side of production; with the consequential result that relative prices are independent of the pattern of consumption and of demand. Indeed, it was "ignoring the influence of the commodity mix that consumers wish to have" and its "neglect of the composition of consumer demand" that puzzled Sir Roy Harrod in his review of the work (The Economic Journal, December 1961 pp.784, 785.), and which he considered unjustified. (In the following June Mr Sraffa had a Note in the same journal (Economic Journal, June 1962, pp. 477-9) to show that the criticism rested on a misunderstanding not in the sense that demand did play a rôle in the Sraffa-system, but that the allegation made by Sir Roy was mistaken, to the effect that a shift in the quantities produced would affect prices in terms of Sraffa's own theory.) We have already remarked that the failure of demand to appear as a determinant was also a characteristic of the von Neumann model; and there was similar bewilderment at this feature of the model when it first appeared. The peculiarity in this latter case was commonly, if wrongly, attributed to the fact that it was a growth-model and that all outputs were inputs, including wage-goods which were implicitly treated (as by Marx) as simple replacement of labour used-up in the production-process (profits being wholly ploughed-back as new investment). There was accordingly no capitalist consumption and no discriminating luxury-demand. It has been not uncommon since then for mathematical models, without these specific restrictions, to exhibit the same general characteristics of prices derived directly from conditions of production and unaffected by the pattern of consumption. Thus a year after the appearance of Production of Commodities by Means of Commodities a formal system was presented by an American economist (Jacob T. Schwartz) which resulted in the conclusion: "We are led to conclude that price-ratios are determined by the technological conditions of production; in particular, no considerable rôle seems to be left for the 'supply and demand' considerations which are so central to the customary economic theory of price ... We emphasise once more... that, considering the almost vanishing rôle played by consumer preference in the above analysis, we have before us very strong presumptive evidence against the marginal utility theory (or more precisely, against its special significance)." Later on the same writer summarised "the proper conclusion at this point" as being "that the rate of profit p is not successfully determined by the Walrasian theories from consideration of production coefficients, utility functions and so forth. What our analysis shows, in fact, is that the determination of the rate of profit is not purely a question of economics at all, but is rather a social-political question... Thus an initial scepticism about classical equilibrium analysis is justified ... The Walrasian determination of this rate is questionable."

Already in earlier chapters, in connection with Dmitriev's defence of Ricardo's system, we have referred to his demonstration that profits and hence relative prices can be immediately determined once the real-wage and the conditions of production (labour expenditures and their dating in time) are known. It was also mentioned that, while conditions of production can be expressed in terms of dated labour (thereby giving a time-pattern to production), they can also be expressed in terms of simultaneous production by means of a set of input-output equations, with labour as one of the inputs (thereby avoiding any problem of infinite regress).

This in essentials can be said to be the method adopted by Sraffa. If in the Dmitriev-equation one were to substitute for the labour terms (the N's) the quantities of the wage-good (A) needed in the course of producing each good, one would certainly have the nucleus of the Sraffa-system. The price-equations in Chapter II of Production of Commodities by Means of Commodities provide for a series of products that are also inputs, some of them in the form of subsistence for workers; the price-equation for each product consisting of the sum of the various input-quantities multiplied by their several prices, with the addition of the rate of profit times that sum. These equations accordingly have this form

(A_aP_a+B_aP_b+...K_aP_k) (I+r) = AP_a

(A_bP_a+B_bP_b+...K_bP_k) (I+r) = BP_b

.      .      .

(A_kP_a+B_kP_b+...K_kP_k) (I+r) = KP_k

There are k equations for the k products, all of which appear both as inputs in some or other products and as outputs; and the k independent equations suffice to determine the k-i price-relations and the rate of profit. These products he calls 'basics'; and it is emphasised that only the conditions of production of 'basics' play a part in determining prices and the rate of profit. If there are products that do not play a productive rôle as inputs (they are called 'non-basics' or 'luxuries'), then "these products have no part in the determination of the system. Their rôle is purely passive. If an invention were to reduce by half the quantity of the means of production which are required to produce a unit of a luxury commodity of this type, the commodity itsef would be halved in price, but there would be no further consequences; the price-relations of the other products and the rate of profit would remain unaffected."

In a later chapter it is explained that each of the price-equations of which we have been speaking could be replaced by a series of labour-terms each with its appropriate date. This is called "Reduction to dated quantities of labour"; and the two formulations are essentially equivalent expressions for the same production-situation, looked at from different angles or in different perspectives, as it were. The reduction-equation for each commodity then consists of a series of labour-terms, each multiplied by the wage, with the addition to this of the rate of profit for the period intervening between the date of the labour input in question and the emergence of the final product. The price-equation for product A would then have this form:

L_aw+L_a1w(1+r)+...L_anw(1+r)ⁿ+... = APa.

What this amounts to is that there are two main ways in which prices can be derived from production-conditions: two ways, as we have just said, that are substantially equivalent. First; prices can be derived by describing production in terms of labour expenditures per unit of output, with a time-period attached to these expenditures. Secondly, they can be derived by describing the situation in terms of total commodity-inputs per unit of output; in which case both the rate of profit and the prices are simultaneously determined as resultants. In the first case the level of real-wages has to be explicitly postulated, in terms of cost in labour or of product, and in the second case this has to be included as one of the commodity-inputs. It should be added that with durable fixed capital in the picture the first alternative cannot be used without some ex cathedra postulation of a depreciation principle, since outputs of different time-periods, or dates, are joint products of the durable equipment in question.

The fact that the level of wages has to be independently postulated as a datum in this mode of price-determination ("regarded as consisting of specified necessities determined by physiological or social conditions which are independent of prices or the rate of profits") means that we are back at the methodology and approach of the (truly) classical system. No attempt is made to derive a theory of distribution from within the circle of exchange; and in the abandonment of this attempt we witness a reversion to the pre-Jevonian order or pattern of determination: prices are derived from (or in part dependent upon) conditions of distribution rather than distribution being derived from the structure of prices treated as being in turn a resultant of demand. True, when handling an equational system, one is dealing with a case of mutual determination; but this, as we have seen, does not negate or exclude direction-pointers to the determination in any substantial theory; and what is of substantial importance here is that among the given conditions of the problem, or postulated data, a social datum is introduced from outside (or, as some might prefer to express it, beneath) the market process. Thus the boundaries of economics as a subject are ipso facto drawn differently and more widely: they are drawn so as to include social, and moreover institutional and historically-relative, changing and changeable, conditions that were excluded from Economics as viewed in the post-Jevonian tradition. The ideological implications of such a difference are evidently extensive and quite crucial; and we have already commented on them in an earlier chapter.

It is in this light that we should, I believe, view an aspect of Sraffa's system - or perhaps one should say of his mode of exposition of it - that some readers appear to find puzzling. Initially labour is treated on a par with material inputs, valued qua input at a subsistence wage, "on the same footing as the fuel for engines or feed for cattle". Surplus then has the same significance as Marx's surplus-value or Ricardian net revenue. Afterwards, to allow for the possibility that wages "may include a share of the surplus product", he adopts the device of "treating the whole wage as variable", excluding it from among the inputs and treating wage-goods consequentially not as 'basics' but as 'non-basics'. Net income then conforms to the conventional definition of national income as including both wages and profits. The wage, however, is still explicitly stated in the equations of price-determination, being introduced there along with the quantities of labour used in the various industries (instead of appearing in the guise of inputs of necessary subsistence). Actually this change is made for reasons of formal convenience, as making it easier to define maximum profit for purpose of the Standard Commodity and demonstrating the effect of a changing wage-profit ratio upon relative prices; and nothing in principle is involved in the change. (As he himself says translation of things into "the more appropriate, if unconventional, interpretation of the wage" can easily be made at the cost of some additional circumlocution.) But it could be regarded as a device for handling situations in which collective bargaining had grown to be a significant influence in the labour market and trade unions were able to encroach upon surplus-value in the interest of higher wages. If one were to seek a Marxian analogy, it might be this. Marx's concept of the value of labour-power could be regarded as being within the context of 'pure' Capitalism, with labour-power sold competitively by individual bidding. Collective bargaining, as soon as it developed, introduced a new element into the situation; and as a result the price of labour-power need no longer correspond with its value but could rise at the expense of surplus value. For theoretical purposes one will now have to adopt as Datum the degree and market-influence of labour-organisation, since this determines how much of what in the 'pure' case was surplus-value is now included in the wage.

A central place is occupied in the Sraffa system by a concept at is devised to solve what we have seen to have been a crucial (and unsolved) problem for Ricardo: that of finding an 'invariable standard', or measure of value, which will be invariable to changes in the ratio of profit to wages. Ricardo sought this in Labour as his Absolute Value; but found that there were difficulties in using this simpliste, to the extent that the proportions of capital (Marx's 'organic composition of capital') differed as between industries. Sraffa's unique contribution has been to solve this problem by means of his so-called 'Standard Commodity': namely by selecting (hypothetically) a commodity, or set of commodities, that have the required properties, so that if this were to be selected as money, or as numeraire, measurement in terms of it would be invariant to distribution-shifts in two crucial respects. First, if wages are defined in terms of it, there is a linear relationship between changes in wages and resulting (and inverse) changes in profit. This is, indeed, the case whether profit is expressed in terms of the standard system or alternatively in terms of the actual system. Secondly, and consequentially, "the ratio of the net product to the means of production would remain the same whatever variations occurred in the division of the net product between wages and profits".

In loose and popular language this standard measure consists of a commodity produced under some kind of average-type conditions of production. But what kind of average? On close inspection definition of such a type is less simple than at first might appear for reasons with which practitioners of input-output analysis are familiar. If one were to look for an actual individual commodity, the qualities it would need to possess for this purpose are explained as follows: The key to the movement of relative prices consequent upon a change in the wage lies in the inequality of proportions in which labour and means of production are employed in the various industries. It is clear that if the proportion were the same in all industries no price-changes could ensue, however great was the diversity of the commodity-composition of the means of production in different industries. For in each industry an equal deduction from the wages would yield just as much as was required for paying the profits on its means of production at a uniform rate without need to disturb the existing prices.

What one accordingly needs to look for is a critical proportion" of labour to means of production, such that, if it were possible to find a commodity that was produced with this proportion, its price would be invariant to a change in wages, since any wage-change yielded just that addition to, or subtraction from, profit as was needed to yield the new uniform rate of profit. This 'critical proportion', be it noted, would need to apply to each 'layer' in its vertical chain of production: to the production of the means of production themselves and to that of the means of production in turn used to produce the former, and so on.

Definition of this crucial proportion is then reduced to two alternative "'pure' ratios between homogeneous quantities...namely the quantity-ratio of direct to indirect labour employed, and the value-ratio of net product to means of production". A standard "composite commodity" is then defined as a set so chosen from existing ones that "the various commodities are represented among its aggregate means of production in the same proportions as they are among its products"; or alternatively a set of commodities arranged "in such proportions that the commodity composition of the aggregate means of production and that of the aggregate product are identical ". Reflection should make clear that the ratio of net product, or surplus, to means of production, or inputs, of this system has a unique meaning - a meaning capable of expression in product-terms as much as in Ricardo's simple product-case of Corn as both input and output.

The relation between this and Ricardo's 'Absolute Value' is then most ingeniously shown by taking a quantity of labour as an alternative but equivalent standard. "A more tangible measure for prices of commodities... is 'the quantity of labour that can be purchased by the Standard net product'." This quantity is given "as soon as we have fixed the rate of profits, and without need of knowing the prices of commodities, a parity is established between the Standard net product and a quantity of labour which depends only on the rate of profits; and the resulting prices of commodities indifferently regarded as being expressed either in the Standard net product or in the quantity of labour which at the given level of the rate of profits is known to be equivalent to it". As wages change, and with them (inversely) profits, this quantity of labour serving as standard likewise changes. The conclusion is that "all the properties of 'an invariable standard of value' ... are found in a variable quantity of labour, which, however, varies according to a simple rule which is independent of prices ". This quantity of labour serving as a unit of measurement "increases in magnitude with the fall of the wage, that is to say with the rise of the rate of profits, so that, from being equal to the annual labour of the system when the rate of profit is zero, it increases without limit as the rate of profit approaches its maximum value".

Thus what was conceived to be a central problem of classical political economy in Ricardo's day has been solved a century and a half later. In the absence of a solution there could be no way of distinguishing in the case "of any particular price fluctuation whether it arises from the peculiarities of the commodity which is being measured or from those of the measuring standard ". But there is a reflection occasioned by this signal achievement that may appear to many as of even more general interest than the details of the solution itself. This is that for so long this problem should have been entirely misconceived, even to the extent of denying its existence as a real problem, and the reason for this misconception and neglect. The reason evidently was the same as for blindness to the possibility of 'reswitching' of productive methods: the inability of post-Ricardian economists to appreciate the dependence of the price-structure on distribution, in their preoccupation (at any rate since Jevons and the Austrians) with the converse dependence of distribution upon a demand-determined price-structure. In this we seem to have yet further illustration (if it be needed) of the bias imparted to thought by the given conceptual framework of one's subject, whether this is inherited or acquired - a framework or 'picture' which at the outset we suggested was ideologically permeated if not prompted and inspired.

What the sequel to all this will be it is too early to say with any assurance; and to remain silent is better than to indulge in speculation that is lacking in sure foundation. But whatever the future course of discussion and analysis, little but clarification can result from the animated critique of the past decade, whether its main significance is destined to be primarily negation and overcoming or mediation and transcendence in some novel synthesis. Whatever the particular outcome, one can say that the discussion of the 1960s was manifestly a turning-point. If only because what had been widely accepted as an orthodoxy of the textbooks has been shaken, and an older, discarded tradition revived, nothing can ever be quite the same again as it was before.

Note to §11 of Chapter Nine

Some reference should perhaps be made in conclusion (virtually as a footnote to what has been said) to something that may well have sprung to mind in the course of reading the previous section. Some may deem it sufficient to say simply that a postulate as to the level of real wages (or alternatively the ratio of surplus to wages) is introduced 'from outside' as a sociological datum (dependent, e.g. on the state of class relations existing at a given time and place). Others, however, may feel some impatience with this, even to the point of regarding it as question-begging, and may sense a need to 'close the model' by introducing some more explicit explanation of the forces that determine the division of the total product between profit (or property-income) and wages. While it may have been legitimate at an early stage of capitalism to accept that wages were governed competitively by something akin to Marx's 'value of labour-power', and surplus accordingly treated as a residual, does not such an approach (which certaily had the merit of including labour-power, and hence wages, within the circle of value-relations) lose its relevance as soon as capitalism develops beyond its earlier competitive stage? Does it not lose both relevance and plausibility in a stage of monopoly-capitalism, with monopolistic (or oligopolistic) firms endowed with the power to pass on a wage-rise into prices and to enforce something like a minimum profit-margin or surplus-ratio? Some might feel inclined to maintain that in such circumstances real wages are determined as the residual rather than profits.

A complaint of this kind would seem to have inspired a recent statement by Dr Nuti about "the relation between the real wage rate and the profit rate uncovered by Sraffa and before him by the Russian economist Dmitriev". This, he observes, "provides scope for the concept of class struggle in the determination of relative shares". But he goes on to say: "unfortunately, however, there is no simple way of closing his system, of determining which point of the wage-profit relation is actually reached and how in any economy"; and this for two reasons. First, "the real wage-rate cannot be taken as exogenously determined as in the classical thought, fixed at a subsistence level in conditions of elastic labour supply". Secondly, "it cannot be determined directly by the class struggle... because after Keynes we have to recognise that wage bargaining determines money wages, while the real wage rate is determined by the behaviour of the priceleve1". This latter is a cogent objection, and, in conjunction with the former, might incline one to think that under modern capitalism, with its high degree of concentration and monopoly, one may well need to look, in explaining distribution, towards factors that set a minimum to profit, rather than as formerly a minimum to wages. Otherwise (it may well be asked) do not critics of the orthodox theory of distribution place themselves in as vulnerable a position as those they criticise and seek to dethrone?

We have seen that Kalecki provided an explanation of distribution of this very type, and with the situation of modern capitalism evidently in mind. According to this the share of profits in (gross) output was determined by the degree of monopoly, which gave the firm or entrepreneur the power to exact a mark-up on prime cost by price-raising. Whatever the level of money-wages, the ratio of prices to it (and hence the real-wage and profit- margin) will be dependent upon the price-raising power with which firms are endowed - something that varies in inverse relation to the amount of effective competition. By such means, it would seem, capitalism has aquired the power to negate the growing influence of trade unionism over money-wages, and the ability to tolerate conditions in which the industrial reserve army can no longer play its former 'stabilising' rôle.

There are some formal difficulties about this theory as to how the degree of monopoly is to be appropriately defined. Kalecki himself interpreted this in terms of the inelasticity of the demand curve (and hence the ratio of average demand-price to marginal revenue and to marginal and average cost). But how is this to be transferred from the level of a particular product to the macroscopic level of the economy as a whole; and what real content does the notion have if the transfer can be satisfactorily achieved? A more substantial difficulty is that by implication the mark-up would be zero under conditions of perfect competition. If surplus-value is sole1y the creation of monopoly, its emergence under the 'normal' competitive conditions envisaged by the classical economists and by Marx would seem to be denied. This objection would not be a serious one if the theory were clearly labelled as one appropriate to monopoly capitalism per Se; an alternative explanation of surplus-value being admittedly appropriate to an earlier, competitive stage of the system, in conditions of a reserve of labour and elastic labour supply.

Alternatively, and with a somewhat different emphasis, one might say that, while the classical Marxian explanation for the emergence of surplus-value continues to apply to modern capitalism, as to its earlier stage, the influence of monopoly enters in as an additive element in the stage of monopoly capitalism - an influence reminiscent of forms of exploitation characteristic of pre-capitalist stages of development.

An analogous difficulty applies to the post-Keynesian theory, according to which the share of profit in national income is dependent on the growth-rate of the economy together with the saving (or alternatively spending) propensity of capitalists. The implication of this would seem to be that in static conditions with zero-capital accumulation (Marx's 'simple reproduction') profits could be zero. This, however, could only be the case if capitalists persistently refused not only to invest but also to spend their (potential) incomes: surplus-value would then fail to accrue simply because it could not be 'realised', even though conditions favourable to its 'creation' were in existence. Since there must necessarily be an equality between aggregate capitalist spending (including investment) and total output less what is sold to wage-earners, it follows that total realised surplus must vary with the size of capitalist expenditure (including investment). This equality-condition, however, could only become a sufficient theory of profit if (say) investment were singled out as the independent variable: then it would appear to follow that amount of profit or surplus is consequent on investment, and determined by the size of the latter. But while this might well be a plausible view in a planned socialist economy, where growth, and hence investment, is postulated as a key policy-element in the plan, it is plainly inappropriate to an unplanned atomistic market economy. In the latter investment as much as individual capitalist spending is a variable that has to be explained in terms of the total situation, and is affected by expectations of future capitalist income, which is in turn related to what has come to be regarded as 'normal' income on the basis of past income-experience. The equality- condition in question, accordingly, cannot be treated as a substitute for a theory of profit and distribution in socio-economic terms (explaininge.g., the rate of surplus-value, or profit-wage ratio, which is otherwise left unexplained).

Yet another plausible hypothesis is that a minimum rate of profit is somehow fixed by some kind of quasi-political or institutional decision-mechanism. This is a plausible interpretation, perhaps, of a tentative hint given by Mr Sraffa when he suggests that in his system the rate of profit, rather than the real-wage, could be postulated as the independent variable: the former is "susceptible of being determined from outside the system of production, in particular by the level of the money rates of interest". The latter would presumably be fixed, in the main, by the Central Bank, whether acting on its own initiative or as an instrument of governmental monetary policy. If one is inclined to view State policy as an instrument or reflection of class interest, or of powerful pressure-groups within the ruling class, one will tend to regard monetary policy as a way of enforcing (more or less consciously), on behalf of capital-owners as a whole, such a profit-share in the proceeds of production as existing circumstances permit. True, they may at times decide to lower interest-rates in pursuit of a 'cheap money policy', as in time of war or of economic depression, or under the influence of international economic relations. (At times of excess capacity, for example, it may well be in the collective interest of profit-receivers to lower interest and profit-rates if thereby investment-expenditure and hence capacity-working can be augmented.) But the very fact that the idea of a 'normal' long-term rate is so stubbornly long-lived, apparently, even in disturbed times, lends support to the notion that the conventional upshot of banking policy is to set a substantial minimum to the profit share.

Finally, before leaving such questions mention should no doubt be made of a difficulty that some feel about postulating real wages as an independent variable if this is done (as by Marx) in value terms (i.e. of labour) or in terms of Sraffa's Standard Commodity. In the former case objection is made that wage-goods are not, in fact, bought by workers at their values but at their prices of production; in the latter case it is objected that the actual consumption of workers will probably not consist of the Standard Commodity but of a quite different assortment of goods. In either case substantial content is lost from the postulate. This difficulty, however, would seem to be more apparent than real provided one is willing to accept the notion of a given standard of living as consisting of a variety of assortments of wage goods that are considered as equivalent by a typical worker's household, and to interpret 'a given level of real wages' in this sense. Then a definition in terms of one member of an equivalent set (whether in terms of labour necessary to produce it or of Standard Commodity) will retain its meaning when translated into another member of the same set of equivalents.

One can only conclude, at the time of writing, that such alternative explanations of distribution in our twentieth-century world are sub judice in current economic discussion, and that discussion (or even elaboration) of them has proceeded insufficiently far as yet to make final judgement possible, still less to speak of a consensus. Unsatisfactory this may be as a concluding note; nonetheless it would appear unavoidable. At least it may be an indication that political economy is not a closed text and that it remains open to the creative moulding of controversy with which its past is so richly endowed. Indeed, this is probably more true today than it was half a century ago when Keynes could write of "the general principles of thought which economists now apply to economic problems" as though these were an agreed corpus of theory.