You are on page 1of 7

Prelude to a Critique of Economic Theory Author(s): Joan Robinson Source: Oxford Economic Papers, New Series, Vol.

13, No. 1 (Feb., 1961), pp. 53-58 Published by: Oxford University Press Stable URL: http://www.jstor.org/stable/2661830 . Accessed: 22/07/2013 14:00
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Oxford University Press is collaborating with JSTOR to digitize, preserve and extend access to Oxford Economic Papers.

http://www.jstor.org

This content downloaded from 143.107.252.196 on Mon, 22 Jul 2013 14:00:02 PM All use subject to JSTOR Terms and Conditions

TO A CRITIQUE PRELUDE THEORY OF ECONOMIC


By JOAN ROBINSON IT is no wonder that this book' took a long time to write. It will not be read quickly. Addicts of pure economic logic who find their craving ill journals products peddled in contemporary satisfiedby the wishy-washy for drop, can by drop that enjoy, elixir they a double-distilled have here many a day. For some, indeed, the logic may be too pure. We plunge immediately discussionof assumptionsand into the argumentwithoutany preliminary a capitalist economy,but to in we are of Evidently delimitation topics. avoid the ambiguities which have clustered around the word, capital is but no enterprises;wages, but no paynever mentioned. There is profit, packets; prices, but no markets. Nothing is mentionedbut the equations of productionand the necessary conditionsof exchange. There is a great deal to be said forthis method of exposition (over and above its lapidary style),foreveryattemptby an authorto explain himself in termsof the preconceptionsof one reader confusesanother. Best leave each to work it out forhimself. To finda clue, let us go back a stage and pick up the argumentfrom Sraffa's Preface to Ricardo's Principles. Postulate that corn is the only commodity consumed by workersand that the corn-wagerate is fixed. or equipment Cornis requiredalso as seed, and thereis no othercommodity in corn existenceat a of stock Then of corn. the for production necessary the beginningof a year has reproduceditselfwith a surplus at the end of the year. The ratio of the surplus to the stock is the rate of profit. The workers are, so to speak, intermediategoods, like machines, necessary forthe process by which corn produces corn. to produceluxuries, may be used to employmoreworkers The corn-profit or to carryout investment;or it may rot in the barns. The way it is used whichis fixedby technical conditions,and the rate of profit, cannot affect the equilibrium prices of all other products are determinedin terms of corn (and so in termsof each other)by theircosts of production,including at the corn-rate upon the capital (valued in corn)requiredto produce profit them. Can the propositionsderivedfromthis model survivethe removal of the postulate that only corn is required to produce corn?
1 Productionof Commodities byMeans of Commodities. Prelude to a CritiqueofEconomic Theory,by Piero Sraffa(Cambridge UniversityPress).

This content downloaded from 143.107.252.196 on Mon, 22 Jul 2013 14:00:02 PM All use subject to JSTOR Terms and Conditions

54

PRELUDE

TO A CRITIQUE

OF ECONOMIC THEORY

The firststep-here the present argument begins-is to introduce a variety of wage goods. Let there be a number of distinct commodities each of which is required,in a particular quantity, to be consumed by a worker, just as particularquantities of oil and fuelare requiredto operate a machine. The commodities are also required to produce each other and themselves.(To set us offon the righttack, wheat, iron,and pigs are mentioned. But they soon become commoditiesa, b, . .. , k.) The same argumentapplies as before. The commoditiesreproducethemselveswith a physicalsurplus. The conditionthat the rate ofprofit is uniform throughout the economy settles their relative prices. The value of the stock of ofthe year and ofthe surplusaftertheyhave commoditiesat the beginning been replaced can be expressed in terms of any one of the commodities. The value of the real wage (which is fixed in physical composition by technical necessity)is also determined,and the cost of productionof any commoditiesthat do not enter into the real wage (subject to the condition that they yield the rulingrate of profit)settles their prices. This merely elaborates the corn-wagemodel without alteringits essence. The next step takes us much further. Instead of the real wage being fixed by physical necessity, the workersreceive a share of the surplus. The author toys with the idea of separatingthe wage into a part which is necessaryand the rest; he rejects it in deferenceto ordinaryusage. He makes this concessionwith evident reluctance,but readers may welcome it, not only to avoid verbal clumsinessbut also because we could hardly imagine that, when the workers had a surplus to spend on beef, their physical need for wheat was unchanged. Wage goods thus cease to be necessary for production in technicallyfixedproportions. There remain, however, commoditieswhich are necessary as means of production for themselves and each other. (The pigs and wheat presumably drop out, but the iron remains.) They reproducethemselveswith the aid of labour and yield a surplus out of which the labour is paid. We are now launched on the main problem-the effect upon prices of changes in the division of the surplus between wages and profits. the division. We are to consider Nothingis said about what determines the consequences,not the causes, of changes in the real wage. It is this, not the austere style,that makes the book difficult.We are uniform concernedwith equilibriumprices and a rate of profit throughout the economy,but we are given only half of an equilibriumsystemto stand on. We need a fence to prevent us plunging offinto the abyss. The author suggests as a helpful (but not necessary) provisional assumption that constant returnsprevail. I, for one, found that this only made me all the more dizzy. It seems betterto assume that changes in the share of the compositionof output. wages do not affect

This content downloaded from 143.107.252.196 on Mon, 22 Jul 2013 14:00:02 PM All use subject to JSTOR Terms and Conditions

JOAN ROBINSON

55

There is a further difficulty. The wage 'changes' only in the sense that the value of x changes as we run our eye up and down a curve. In the year that we are examining,each change has already happened. So long as all commodities reproduce themselves within a year, this is easy to accept; but when long-lived machines come into the picture (in a later chapter) it causes discomfort. Can the equalization of the rate of profit throughoutthe economy come about except throughthe equalization of expected profits on new investmentin various lines ? If the rate of profit has changed duringthe life-time of machines in existence this year, there is no equality between expected and realized profitsin any one linelines? why should there be equality between realized profitsin different Let us add to the protectivefenceof provisionalassumptionsthat we need not take the word 'change' literally. We are only to compare the effects of having differing rates of profit, with the same technical conditionsand the same compositionof output. Thus reassured,we can remainon the narrow ledge withoutvertigo. When the wage is not given by technical conditions,what do prices mean? A change in the division of the surplus between wages and profits alters relativeprices. But we need to know the prices to value the surplus that is to be divided. This was the problem that flummoxedRicardo. Sraffa's solution is ingenious and satisfying. He isolates those basic commoditieswhich enter directlyor indirectlyinto the production of all commodities and, from the technicalequations whichshowhow each enters into the productionof the others,he constructsa standard of value in the formof a composite commodityinto which each particular item enters, as means of production,in the same proportionas it appears as output. The beauty of this is that, as the wage reckonedin terms of this standard rises, the prices of some of the commoditiescomposingit (in which are a wages are a highproportionof cost) rise,and others(in whichprofits of cost) fall,to just such an extentas to balance each other, highproportion and leave the ratio of the value of the surplusto the value of the means of production unchanged. This provides a technically determinedratio of surplusto means of productionwhichis independentof the division of the surplus between wages and profits. Now, given the n technical equations forn commodities,and the wage rate in termsof the standard, the n-I prices and the rate of profitare the wage determined. Or, given the n equations and the rate of profit, is determined. Assumingthat wages are paid at the end of the year (no capital is required to financea wage fund) there is a linear relationshipbetween the share of wages in the surplus and the rate of profit. This having been established, the standard commoditycan be left to

This content downloaded from 143.107.252.196 on Mon, 22 Jul 2013 14:00:02 PM All use subject to JSTOR Terms and Conditions

56

PRELUDE

TO A CRITIQUE

OF ECONOMIC THEORY

look after itself and the argument is conducted in terms of the rate of to zero wages (that is, the ratio of surplusto means of profitcorresponding withthe wage rate that it entails. production), and the actual rate ofprofit, In order to constructthe standard commodityit must be possible to find a quorum of basics-commodities that enter directly or indirectly into the productionof all commodities.So long as thereare necessarywage goods there are bound to be basics, for,via labour, the wage goods enter into all production. But when wages are part of the surplus we have to fall back on an assumption that there is at least one basic commodity. Certainlythat is plausible enough, but it is natural to ask what would happen if there were none. Does the whole method stand or fall on this assumption? I thinknot. Suppose that technical equations could be divided into two systems into withoutany overlap, in one of whichiron entersdirectlyor indirectly the production of all commodities, and in the other, wood. The two systems of equations belong to the same economy in the sense that the rate of profit and the wage rate are the same in both. Now, when the rate is determined of profit is given,the wage rate in termsofthe iron-standard for the iron system and the wage rate in terms of the wood-standard is determined for the wood system. The fact that the wage is uniform determinesthe price of iron in the wood-standard. The assumption of at least one basic commoditythus appears to be a mere simplification, not a crucial step in the argument. After exploring the propertiesof a system in which each productive process takes one year and produces one commodity,we are shown the application of the methodto joint products,fixedcapital and land, and to the choice of techniquewhen alternativemethodsare available forproducing a singlecommodity.The argumentthen ceases as suddenlyas it began. In elaboratingthe method to deal with complexitiessuch as long-lived machines,many points of great interestare turnedup (includinga version of the formulaforthe relation of the value of a machine to its cost which was worked out, presumably, much later, though published earlier, by Kahn and Champernowne),but the main point of dealing with these problemsis just to show that it can be done. The essence of the argument remains that which is exhibited with circulatingcapital only. The sub-titlegives a hintofthe purpose forwhichit has been established to -Prelude toa Critique ofEconomicTheory.In the preface,afterreferring a draftof the book which he discussed with Keynes in 1928, Sraffawrites:
As was only natural during such a long period, othershave fromtime to time taken up points of view which are similarto one or other of those independently from directions or in different adopted in thispaper and have developedthemfurther now those pursuedhere. It is, however,a peculiar featureof the set of propositions publishedthat,althoughtheydo not enterintoany discussionofthemarginaltheory

This content downloaded from 143.107.252.196 on Mon, 22 Jul 2013 14:00:02 PM All use subject to JSTOR Terms and Conditions

JOAN ROBINSON

57

been designedto serve as the basis ofvalue and distribution, theyhave nevertheless fora critiqueof that theory.

The significant word is 'however'. Others have developed input-output systemsand process analysis to higher degrees of elaboration than are shownhere,but theyhave not broughtthem to bear on the foundationsof orthodoxdoctrine. Can we divine what the critiquewill be? There are threemain propositions which can be derived from the corn-wagemodel and which have been shown to survive all the necessary modificationsthat follow from elaboratingits assumptions. is that, when we are provided with a set of technicalequations The first for production and a real wage rate which is uniformthroughoutthe of economy,there is no room fordemand equations in the determination and allow that fence, equilibrium prices. (Whenwe take downourprotective changes in distributionaffectthe composition of output, we shall need a freshset of equations relatingthem,but that is quite anothermatter.) a dead Marshallianhorse Some mightcomplainthat this is only flogging (which Sraffahimselfhelped to kill, even before1928). But to my mind it emphasizes a point which,both in its scholastic and in its political aspect, is of great importance; in a market economy,eitherthere may be a tenin different of wages and the rate of profit lines dency towards uniformity of production,or prices may be governedby supply and demand, but not both. Where supply and demand rule, thereis no room foruniform levels of wages and the rate of profit.The Walrasian system makes sense if we interpret it in termsof an artisan economy,where each produceris committedto a particularproduct,so that his income depends on his output and its price. Each can have a prospectiverate of returnon investment in his own line, but there is no mechanism to equalize profitsbetween one line and another. In real life,no one expects to see an equalization of obtainable fromsugarin Cuba and cocoa in Ghana or can the rates of profit even say what an equal rate of profitwould mean. of demand equations into the theoryofthe wage economy, The intrusion onto the exchange economy,have and the attemptto foista rate of profit led to endless confusion;a critiqueto clear it up is long overdue. to the The second propositionis mentionedby Sraffain his References Literature.It is the rejectionof the claim 'that the price of everycommodity,either immediatelyor ultimately,resolves itself entirely(that is to and rent.' residue) into wage, profit, say, withoutleaving any commodity In the corn-wage economy,the productionof cornthis year requiresthat there should be a stock of corn already in existence,to provide seed and the subsistenceof the workersuntil the next harvest. Sraffahas removed the assumption of a technically determined physical real wage. This

This content downloaded from 143.107.252.196 on Mon, 22 Jul 2013 14:00:02 PM All use subject to JSTOR Terms and Conditions

58

PRELUDE

TO A CRITIQUE

OF ECONOMIC THEORY

throwsgreat weightupon commoditiesregarded as means of production, a weight made all the greater by the assumption that capital is not required for a wage fund. Productionof Commodities by means of Commoditiesis his centraltheme. It leads to the very strikingproposition that there is a technically determined maximum notionally possible rate of profit,which would obtain at zero wages. (It is only notionally possible, for even when the postulate of a precise physically necessary wage has been abandoned, there is still a vague but tough lower limit to possible real wages and so an upper limit to the possible rate of profit.) The thirdproposition,if we may indulge in a loose mode of expression that the author carefully avoids, is that the marginalproductivitytheory of distribution is all bosh. Sraffadoes not deny any sensible argumentsthat can be expressed in marginal terms. His treatmentof diminishing returnsfromland and of the choice of technique makes room forlegitimateuses of the concept of a productivefunction. What he demonstratesdecisively (though doubtless the deaf adders will take no notice) is that there is no such thing as a 'quantity of capital' which exists independentlyof the rate of profit. It is importantto realize that the third proposition does not depend upon the second. Certainlythe propositionthat no production,by the methods known today, could take place without some pre-existing commodities,is highly plausible, but it is a matterof fact,not of logic. It does not mean that if prices could be reduced, without residue, to wage, profitand rent, then the marginalproductivity theoryof distribution would be cogent. Flint mines were dug with antlers picked up in the forest. If this economy was run on capitalist lines, it must have been necessary to advance wages to the men collecting antlers (otherwisethey would be self-employedtraders). Men dug the pits and shaped the flints. All processes could be reduced to termsof dated inputs of labour. To findthe capital requiredforproduction(in the sense in whichcapital is the principal on which profit is the interest)we must know eitherthe wage in termsof axes or the rate of profit. Certainly,Sraffa is right that in Ricardo's time, or our own, commodities are necessary to produce commodities. But even the neolithic rate of profitwas not determinedby the 'marginal product of capital'. Presumably, it will be a little time beforethe critique to which this is the prelude will be published. We mighthave some self-criticism meanwhile. Cambridge

This content downloaded from 143.107.252.196 on Mon, 22 Jul 2013 14:00:02 PM All use subject to JSTOR Terms and Conditions

You might also like