You are on page 1of 17

Retrospectives: Whatever Happened to the Cambridge Capital Theory Controversies? Avi J. Cohen; G. C.

Harcourt The Journal of Economic Perspectives, Vol. 17, No. 1. (Winter, 2003), pp. 199-214.
Stable URL: The Journal of Economic Perspectives is currently published by American Economic Association.

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.

The JSTOR Archive is a trusted digital repository providing for long-term preservation and access to leading academic journals and scholarly literature from around the world. The Archive is supported by libraries, scholarly societies, publishers, and foundations. It is an initiative of JSTOR, a not-for-profit organization with a mission to help the scholarly community take advantage of advances in technology. For more information regarding JSTOR, please contact Thu Jun 21 12:47:44 2007

Retrospectives Whatever Happened to the Cambridge Capital Theory Controversies?

Avi J. Cohen and G. C. Harcourt

This feature addresses the histoi~ econorrlic nards and ideas The hope is to of deeper1 the norknclnl dialogue of economists, n l ~ i l e perhaps also casting ileu light on orlgoiilg questions If xou hale suggestions for future topics or authors, please nrite to Joseph Pershx , c/o J ~ L n n)l oJ bronomrr P~)sp~rtr-i~~s_), Depart~rlent EconornL of ics (hf/(, 144), L-ni\ersit\ of Illi~~ois C hicago, b01 South Morgail Street, Room at 2103, Chicago, Illi~~ois CiOCi07-7121

Preliminaries: Joan Robinson's Complaints

In "The Production Function and the Theoil of Capital," Joan Robillso11 (19.33-19.54, p. 81) wrote: the production function has beell a ponerful instrument of rniseclucation
The student of ecoilorrlic theoil is taught to n n t e (2 = J ( L , A ) \\here L is
a quantitx of labor, K a qunntitx of caprtnl and Q a rate of outpc~tof
corrl~l~odities is instructed to nssume all norhers alike, and to mensure I
He in man-hours of labor, he is told something about the index-nc~rnher problem
in choosii~g mlit of O L I ~ P L I ~ ,
then h c 1 hcuiiecl oil to the next question,
a 'iiid s in the hope that he nil1 forget to ash in nhat units K is measured Before he
el er cloes ash, he has become n professor, nrlcl so slopph habits of thought are
hnrlclecl on from one geileratioil to the next

w A1 J. C , o h ~ tz ~ Asso( r(~tr ~ O ~ P Sof O ( o t ~ o t ~ r ( I h t k lj7r71rt o f ) , To)onto, ( d a n ( ~ d (( 72 c P S FI s, ; ~

C : .

Hotcour? r c E m r i - z t ~RP(IC/P~ H z s t o t ~oJLrot101~1zr' ~ I P O I C,' t ~ r z ~ ~ t soft ( ( l t ~ b i - z ~(1n(1 ~s rt1 tkr ~ J r ) lg~ Linr)ztzls JPIIOW, P S U (.011~g~, J S botlc zn (,cctrll)trrlg~, C'nrt~d k'rt~g(lotn,n n d P)cg~ssor I I I P ) / ~ Z L ~ , L C't~rz~r)srt) L 4 ~ l ~ l ( l rAd ~ ,P ~ I LAllctt(llr(c T h ~ u ? ~ I I LnZd~d ) ~ s s ~I cP ( c l 7 1 i t o h ~ ~ l @ ) o ) k l l . ( ( ~ ) OJ Z~Y, P ( ) (end ( J P ~ ~ ~ ~ u s - s@ IcP~~P ~ ( L(~ L ? ~ ) , ~ c f.1 ~ ( t i 7 1 ~ 1 1 P Z~S. ( z~k .t ~( (

Her article precipitated illto the public domaill the Cnmbriclge contro\ersies 1 1 capital theon, so-called b\ Harcourt (1C)hC)) because the protagorlrsts uere 1 prii~cipall\ nssocintecl clirecth or inclirectl\ nith Cambridge, Fngland, or Cambridge, AMassnchusetts The contro\ersies raged from the mid-19.50s through the mid-l97Os, nith highh prominent protagonists-Piero Sraffa, Joan Robinson, Lcugi Pnsiiletti and Pierangelo Garegilaili in the "English" corner, \ersus Paul Samuelson, Robert Solon, Frank Hahn nrld Christopher B l i in the "Xrner~can" l~eoclassical ~ or corilerl-slugging it out in first-rnnk journals 4uch a5 the (2unr?rr/~Jour~lctl of Economzt s, the RRI/P~Oc o ~ o t ~ z r of F Itzldzrs and the r t o t / o m / tJ o u r ~ c l l Rlnug (1975) nnd Harcourt (I 972, 1976) co\ er both sides of the contra\ ersx The Carnbndge contro\ersies, if rerrlembered nt all, are usunlh poltra~ecl todax as a tempest 111 a teapot o\er ailomalies inrol\iilg the measurement of capital in aggreg'lte pioduction fui~ctiorlmodels, ha\ing as little sigilif~cancefor rhe neoclnssical mnrgiilnl productnit\ theon of clistr~butionas do Giffcn good ailomalies for the Inn of demaild ll'hen theories of endogenous giou th and real busriless cxcles took off in the 1980s using aggregate ploduction functions, contrib~~tors usunllx \\ 1 ote '1s if the contra\ ersies had 1 x 1er o c c ~ecl r and the ( ,lml~ridge, ~ ~ England contributors had ne\er existed (Kobinson and Si,tffa obligecl b\ d\ ri~g in 1983 ) Siilce l~eoclassicnltheon has sunir ed and the ch,~llenger hn\ e larqelh s disappeared, the usual conclusion is that the "Englrsh" Cantabrigi~~ns \\ere c lcarlx TI rong or a i ongheaded Did the C ambriclge contro\ersies identif\ "slopp\ habits of thoc~glit" t i ~ hale t t been handed dorm to \et ailother generntion, or uere the\ a teapot tempest of concern n o ~ t onl\ to historians of ecoi~omicsi this article, our alm is to put into I11 perspectne \\hat nas at stake nild to argue that the contro\ersies c but thr latest in a series of still-unresol\ed contro\ersies o\er three deep issues The first is the meniliilg and, ns a corollan, the measurement of the concept of capitnl in the anal\sis of industrial capitalist societies The secoilcl is Jon11 Kobinson's coinplniilt that equilibrium nas not the outcome of an ecoilomic process ant1 therefore an inadequnte tool for anal\zing processes of capital acc c~rnulntionand gro\\th The third issue is the role of ideolog and \ision in fuelling contro\eis\ then the results of simple models are not robust Our aim is to coilrirlce the reader of the importailce and rele\ance toda\ of these issues, nhich, n e pl edict, \ti11 inc\itnbl~ erupt in future contro\ers\.'

Sraffa. Robinson and Pasinetti were at the I;\iq of (:anibridge. Pasinetti and (;aregnani had hot11 been P1i.D. students there in the 1950s. O n the "hrerican" side, Samuelson and solo^\- lrcre at \11T. Hahn was actually at the I'niversity of C:arnbl-idge.and Bliss, though at Essex and tlien Oxford. had been an nndel.graduate and university teacher at (:ambridge. 2003 marks numerous anniversaries-the 30th of Robinson's oiiginal paper. the 100th of her birth and the %It11 of Robinson's and Sraffa's deaths.

Round 1: Meaning and Measurement of Capital in the Scarcity Theory of Price

With the margirlal rerolution, Jerons, hleilger and Walras dereloped pure exchange models in the 1870s that shifted the explailatioil of price auax from the to classical difficultj-of-production focc~s the ileoclassical focus on utility and relatire scarcitr Adam Smith's diamond-nater paradox nas no lorlger a pamclox, since price nas explaiiled as proportioilal to margirlal utilitr , J\ hich deperldecl on scarcitj Neoclassical capital theon uas the arena for exteildiilg the general priilciple of relatir e scarcity to explaiil all prices, including factor prices in models 111th production and time (Hennings, 1985). A common starting poirlt for the neoclassical perspectire on capital is a aggregate prodc~ction fcmction model one-commoclitr S a m ~ ~ e l s o n / S o l o ~ \ / S ~ \ a n

\\here the one produced good ( Q) call be consumed clirectlx or stockpiled for use liren resources as a capital good ( k ) With the usual assumptions, like exogenousl\ k ' and t e c l ~ i ~ o l o constailt returns to scale, dimiilishiilg margiilal productir itj and ~g~, corrlpetitir e equilibrium, this simple model exhibits J\ hat Sarnuelson (1962) called three kex "parables" 1) The real return on capital (the rate of interest) is determilled b\ the technical properties of the diminishing margiilal productiritr of capital, 2) a greater quantitr of capital leads to a loner margiilal product of additiorlal capital ailcl thus to a loner rate of interest, and the same inrerse, rrlonotoilic relation nith the rate of interest also holds for the capital/output ratio ailcl sustainable lerels of consumption per head, 3) the distribution of illcome between laborers and capitalists is explained b\ relatire factor scarcities/supplies and margiilal products. The price of capital senices (the rate of interest) is deterrrlirled b the relatire scarcitj and margirlal productiritr of aggregate capital, x ailcl the price of labor senices (the nage rate) is determilled b the relatire scarcitr x ailcl rnargirlal productnitr of labor ( I ) The three parables of this one-commoditj model depend on a plnsical conception of capital (and labor) for their one-ua) causation-changes in factor ponerfcll, unambiguous quantities cause iilr erse changes in factor prices, allo~\ing preclictiorls like parable 2 But problems for these parables arise in more general models nit11 heterogeneous capital goods Heterogeneous capital goods caililot be measured and aggregated 111 phxsical units, instead, capital raluation must be used, as Wicksell (1911 [I9341) rolurne 1, p 149) told us long ago Their ralue call be measured either as the cost of prodc~ction, ~ h i takes time, or the present ralue of the future output ~ ch stream thex produce In either case, since the measure inrolres time, it presumes a rate of interest-nhich, in the simple model, is determilled in a one-nax maililer b\ the quantity of capital. This additiorlal circularitj, or interdependence, causes

202 Jour nnl of Eco~lotr~ic Perspectru~s

M'icksell effects. Wicksell effects invol~e changes in the value of' the capital stock associated ~vith different intereat ratea, arising fiom either inrentoll revaluations of the same physical stock due to new capital goods prices (price MTcksell effects) or diff'erencea in the physical stock of capital gooda (real Wickaell effects). In the Cambridge controversies, the problem5 created for the ileoclassical parables by Wickaell effects were termed reswitching and capital-re~ersing.Res~vitchingoccurs \vhen the same technique-a particular physical capital/labor ratio-ia preferred at two or more rates of interest while other techniqc~esare preferred at iilterrnediate ratea. At lower values of the intereat rate, the costminirniziilg technique "a~vitchea" from n to hand then ("res~vitches") back to a. The same physical technique is associatecl wit11 two different interest rates, violating parahlea 1 and 2. With capital-reversing, a lower capital/labor ratio ia associatecl with a lower interest rate. In comparing two ateady-state equilibrium positions, it is aa though capital aenrices h a ~ e lo~urr rice when capital is "more acarce." Capital-reversing a p implies that the clernaild cu1l.e for capital is not ahvaya do\vn~varcl sloping, violating parables 2 and 3. T.211y do rea~vitching and capital-re~ersing occclr? Sam~lelaon(1t)CiCi) pro~idea the intuition using the ,;\ustrian collceptioil of capital as titrw, ao that the productivity of capital is the productivity of time itaelf: Figure 1 illustrates two techniques for making charrlpagl~e c~aing only labor and time (and fi-ee grapea). In technique a , 7 uilits of labor make 1 unit of brandy in one period, ~vhich ferments into 1 emit of champagne in another period. In technique b, 2 clnits of labor make 1 unit of grape juice in one period, which ripens into \vine in another period. Then Cunits of labor i shaking the wine produce 1 unit of champaglle in a third period. The cost-minimizing technique deperlds on relatiye factor prices. At high interest ratea ( r > 100 percent), compounded interest on the 2 units of' labor in~ested 3 periods makes b more expensive, so n is chosen. At zero intereat, only for labor costs count, so n is also cheaper. Rut at irltereat rates bet~veen percent and 30 demand for capital cur\.e ~voulcl look 100 percent, h is cheaper. The correspol~diilg like Figure 2. First, notice that at different ~ a l u e of r alorlg any discreet do~vn~vards sloping segment, the ~ a l u e the "capital" ia different for a physically unchanging of technique, clue to price MTcksell effects. Notice also that at lo~ver values of' r, the technique s~vitches from n to b and then rea~vitches back to a , due to real TTickaell effects. And at a value of rjuat below 100 percent, capital-reversing occurs as a lower r is associatecl with a lo~ver capital/labor ratio. Becauae of'Wickael1 eff'ecta, in rnoclela with heterogeneous capital goods (or heterogeneous output), the rate of interest depeilcls not only on exogenous technical properties of' capital, hut also on endogenously determilled prices like the irltereat rate. The endogeneit). of' pricea allo~vs multiple equilibria, ~vhichcomplicates the one-\vay parable explallatioils of income distribution. Differences in quantities no longer yield unambiguously signed price effects. The power and aimplicit). of one-commodity modela emanate5 horn eliminatirlg these endogenous price effect5 and measurement problems (C;ohen, lt189).

Fiprp 1 Samuelson's (1966) Example of Wicksell Effects in a Simple Austrian Model

Technique 6


The cost equations are:

Technique (1 ' i L ( 1 A F ) "
Technique 6 2/.(1 A r ) ' 61.(1 r)
\Vhen compar-ing cost\. I. cancels our for I~otli tecliniques. Suitchpoint\ occur when the cob6 ar-e equal

Figz~r~ 2 Demand for Capital (Per Unit of Labor) in Samuelson's (1966) Example

A\ earl\ as 1936, Sraffa urote a letter to Kobiilsoil explaining the essence of this corrlplication for rleoclassical capital t h e o i ~Res~\itching and capital-re\ersiilg \\ere noted in the 1950s b\ Dn\icl Champernonne (1953-1954) and Robinson, but their full sigilificance nas realized onh 111th Sraffn's 19bO book Smffa (19b2, p 4'79) powd the kex question regarding the meailiilg 'uld measurement of capital %\llat is the good of a quantih of capital . ~ \ h i c h ,since it depends on the rate of

interest, cannot be used for its traditional purpose . . . to determine the rate of interest[?]


Round 2: Equilibrium and Time, Differences Versus Changes

Capital is fundarnentallv intertnined with issues of time. A! Bliss (1975, p. 39) urote: "One of the esseiltial tasks of a theon of capital is . . . to make clear nhr a purelv static and timeless economic theon could not be adequate." Questions about the measurement of capital in aggregate prodc~ction fimction gronth models segued to questions about hen, if at all, ma\ clrnamic processes of accumc~lation and distribution be analhzed nithin an essentiallr static equilibric~rn frarnenork. The rleoclassical approach to capital comrnonlv exarnirles accumulation and rates of return using comparative statics exercises-inclctdil~g comparisons of steaclv-state gro~vth paths-~vhich reflect differences in initial conditions. Robinson irlslsted that such cornparisoils clicl not re\eal anrthing about p r o c ~ s s ~ s accumuof lation and gror\th, eren their ultimate outcomes. She "frequenth had occasioil to complaiil of the inabilitr of rleoclassical nriters to distinguish betnee11 a in the parameters of an eqcullbrium model and the effects of a changt. taking place at a rnornerlt of time" (Kobinson, 1980, \olume 5, p. lil) Her phrase "histon \ ersus equilibrium" summed up this methodological critique. Kobinson (19'74 [1980], p. -5'7) argued: The real source of trouble is the confi~sionbet~veencomparisons of equilibrium positlolls arltl the hlston of a process of accumulation. We mlght suppose that n e can take a number of still photographs of economies each in statiorlan eqc~ilibriurn; . . This is an allor\able thoc~ght . experiment. But it is not allonable to flip the stills through a projector to obtain a moriilg picture of a process of accumulation. Thus, marlr hears ago, Robillsoil (1933, p. 390) put back on the agenda \\hat u e nor\ call path-dependent equilibria: "the lei7 process of moving has an effect upon the clestlnatiorl of the molernent, so that there is no such thing as a position of long-run equilibrium r\hich exists inclependentlr of the course ~111ich econthe ornr is follo~virlg a particular date." The title of her 1975 paper, "The Unimporat tance of Resnitching" (Robinson, 19'75a),reflected her belief that J\ hile resr\itching arld capital-re\erslng nere problematic for neoclassical capital theon, her methotlological critique uas far Inore important.

" 110 similar valuation problems arise for heterogeneous labor? The crucial difference wit11 capital is tliat tliere is no theoretical presumption tliat competition \<illequalize wages across different t y e s of labor, in the xvav that rates of return \\,ill equalize (arljnsted for risk) across im.esrments in different capital goodszindustries. To the extent tliat liererogeneous labor reflects differences in lluman capiral, the valuation problems for the neoclassical parables due to interest rate changes are only exacerbated.

Avr J Cohpn a n d G. C. Harrour?


The neoclassical comparative statics exercises are clearlx an abstraction, as are all econornic rnotlels. The crucial question is nhether the abstraction o f comparati\e statics ca11fu)es or obscutes essential features of the accumulation process."

Round 3: Neoclassicals Fight Back: Aggregate Production Functions-1 956 -1 966

Solow (1955-1956) irnrnetliately recogrlizetl that problems in measuring aggregate capital clue to MTcksell effects could be overcorne only "in vei? special cases" and presciently commented that "the real difficulty of [capital] comes not from the physical diversity of capital gootls. It cornes horn the intertwining of past, present and future . . ." He cocmteretl with an enz11im'ral tlefense of one-cornrnotlity motlels as captclring the essential features of the growth process, a position heltl consistently to this day (Solow, 2000; belt see also Pasinetti, 2000). With characteristic xvit, he clefends his choice by saying that "if God had meant there to be more than tsvo factors of protluction, He ~voultlhave matle it easier for us to tlra~v three-tlimensional diagrams." Solow's (1956, 1957) one-cornmotlity proclc~ction fc~nctionrnoclel enabled him to measure the respective contributions of capital deepening and technical progress to gro~vth output per head over time. in There were also three, less successfc~l,theoretical attempts dclring the period from about 1956 to 1966 to fend off the problems of heterogeneoc~scapital. First, Sxvan (1956) introducetl into this round of capital controversies what came to be kno~vn the metaphor of "putty capital." He collapsetl the ever present tension as between capital as physically heterogeneous capital goods and as homogeneous funds flowing to equalize rates of return through his metaphor of meccano sets, the pieces of which can be timelessly and costlessly reshaped into appropriate quantities of "capital" in response to the pull of relative factor prices. These metaphors originated with Biihm-Bawerk (1907, p. %0), who wrote in the context of his controversy with J.B. Clark: "Clark thinks of capital as a quantum of value 'imputed' in material goods. He strips off everything which may suggest material existence, and retains only a value jelly, existing eternally." Subsequent metaphors included "leets" (steel spelled backward), butter, leg0 and putty (as opposed to clay). But all of these metaphoric feints, which effectively collapse heterogeneous capital goods into a one all-purpose commodity, only avoid, but d o not solve, bVicksell's problems. Solow (1963) re-entered the ring with a second theoretical response from the
'Robinson also aimed this critique at her otherwise allies, Sraffa's followers. Garegnani, for example, staunchly defends the view that rigorous results mav only be established within a franlework capturing the effects of persistent forces that characterire long-period positions-for example, the natural prices of the classical political economists, the prices of production of Marx and the long-period normal equilihriunl prices of hlarshall, the economist's counterpart of the natural sciences' traditional centers of gravitation. Garegnani's (1970) insistence on using this long-period nlethod is why Robinson directh and Iialdor indirectlv fell out with the Sraffians.


Journal of Economzc Perspectzm

neoclassicals, attempting to aloid problerns of capital by focusing on the rate of return on in~estment.In the tradition of Ii-ring Fisher (1930), this was capital t h e o q without any inention of either "capital" or "its" marginal product. Solow's model addressed the question "what is the expected marginal return to a little rnore saving/investment in a fully employed economy?" and sei-red as the basis for empirical estimates of rates of return in actual econoinies. Pasinetti (1969, 1970) argued that neither Fisher's nor Solo~v's approach provided an intuitively satisfying explanation of the rate of return unless an "unobtrusive postulate" that disallo~ved capital-reversing was slipped into the analysis, although Solow (1970) disputed this. The third theoretical neoclassical response attempted to extend the onecornlnoctitv lesults to inore general heterogeneous coinmodit). models. Sarnuelsoil'\ (1963) attempt in the "surrogate production function" included what appealed to be a Tariet! of ph!sicallv distinct capital goods, but he also assumed equal factor proportions in all industries, rnaking relatile prices independent of changes in distrib~~tion between wages and profits. As Sainuelson subsequently realized, this collapsed his model back to one cornmodit! .' effecti~el~ By the late 1960s, Sainuelson's (1966, p. 568) judicious "Summing Up" article admitted that outside of one-commodity models, reswitching and capital-re~ersing may be usual, rather than anomalous, theoretical results and that the three neoclassical parables "cannot be universally valid." On a theoretical level, the "English" Cantabrigians won the round over aggregate production functions. Even neoclassical~ like Hahn (1972, p. 8) sllorved no mercy for aggregate production functions, which "cannot be sho~vn follo~v to from proper [general equilibrium] t h e o n and in general [are] therefore open to severe logical objections." They fell out of favor in the 1970s and early 1980s until their revival with elldogerlous growth and real business cycle theories.

Round 4: General Equilibrium-1966 and Beyond

X final neoclassical theoretical counteroffensi\e moTed into the arena of general equilibrium, with Bliss and Hahn replacing Solow and Samuelson as k e ~ protagonists General equilibrium models sustain the general lleoclassical principle of explainillg all pnces, including factor prices, bv relatiTe scarclt~in that prices are , detelnuned b preferences, endo~tments x and technolog, and factor returns are q u o 1 to or nzeosz~redby disaggregated marginal products. CompetitiTe equilibrium prices are also corlslstent 1~1th Pareto-efficient dlspositioll of output. a The three parables, howel er- especiall~the in\ erse, monotonic relation between the quantit~ capital and the rate of interest-\\ere not rescued. Bliss's 1975 of book (which most obseners consider the definitiTe lleoclassical treatment of capital theon that ended the Cambridge contro\ersles) exarnines this relation uslng

" Ironically, Sanluelson's simplifying assunlption also supported Ifarx's labor theoq of value!

Tt71at~oerHnfifiened to the Cnnzbridge Cnfiital T h e 0 7 Controverszes?


intertemporal general equilibriurn models to generate comparative static results." Bliss (1975, p. 85) collcludes that "there is no support from the t h e o l ~ general of equilibrium for the proposition that an input to productioll will be cheaper in an econorn: where rnore of it is available." Sraffians get the same result (Schefold, "00). The general equilibriurn round was motivated by Samuelson's quest, in his surrogate production filnction model, "to pro\ide some rationalization for the validit: of the simple J. B. Clark parables" (Sarnuelson, 1962, p. 194, emphasis in original). Clark (1891, p. 312) made straightforward one-way causal claims: " [A] s capital increases, while other things rernained unchanged, interest falls and as the labor forces increases, if other things remain the same, wages fall." Samuelson's failure prompted a "retreat" to general equilibrium models. But the switch to general equilibriurn, rather than saving the lleoclassical parables, abandoned them for simultaneous equation price systems, and correct statements about factor returns being equal to or measured b disaggregated marginal producti~ities. ! Relinquished, ho~vever, were one-way causal claims about unambiguously signed differences in the interest rate associated with differences in the quantity of capital. As Hahn (1981, p. 128) put it, lleoclassical general equilibriurn "is not cornmitted to a relatile scarcity theon of distribution." hloreo\er, the general equilibrium approach re\italized Robinson's concerns about equilibrium. Theoretical work, specifically, the disappointing SonnenscheinMantel-Debreu stability results, found no particular reason to believe in the stabiliq of the general equilibrium outcome. In discussing these results, Hahn (1984, p. 53) wrote: " [ T l h e Arrow-Debreu construction . . . must relinquish the claim of prolidi~ig rlecessan descriptions of terminal states of ecollonlic processes." The lack of adequate stability results raised questions about the conception of equilibrium as the end of an economic process and the adequac! of comparati\e statics as explanations of the process of change following a parameter shift (Fisher, 1989; Irlgrao and Israel, 1990).

And the Winner is . . . ?

Not so fast. The fight was far from oler because there was no agreement on the ~z~p7zficance all of these result^. The two sides used different criteria to judge the of agreed upon outcomes of the contro\ers:. The different criteria involve another ongoing and unresolved controvers:: Has there been continuity in the e\olution of economic t h e o q frorn Adam Smith to the present or discontinuit!, with the marginal relolution setting lleoclassical economics on a different path from earlier classical political econonq and Marx (Bharadwaj, 1978)'i The "English" Cantabridgians, who viewed Sraffa's 1960 book
" Dixit ( 1 9 i i ) said in effect that Bliss's arguments made the quasi-rents of most prelious writing on capital theor? either rero or, with regard to those of Cambridge, England, negative.

as a revival of classical theo1-1 (Sraffa also edited Ricardo's collected works), belie1e in discontinuity; most neoclassicals, in continuity. il.lnile neoclassical economics enlisions the lifetime utility-maximizing consumption decisions of individuals as the driving force of econonlic activity, with the allocatioll of given, scarce resources as the fundamental econoinic problem, the "English" Cantabrigians argue for a return to a classical political econom: vision. There, profit-making decisions of capitalist firins are the driving force, with the filndanlelltal economic problem being the allocation of surplus output to ensure reproduction and growth (bl'alsh and Gram, 1980). Because individuals depend on the market for their livelihoods, social class (their position within the dilision of labor) becomes the fundainental unit of analvsis. The potential rate of profits on capital arises from differing power and social relationships in production, and the realiration of profits is brought about b7 effectke demand associated with saling and spellding behaviors of the different classes and the "aniinal spirits" of capitalists. The rate of profits is thus an outcome of the accumulation process.' Robinson argued-citing I'eblen (1908) and raising the specter of hlarx-that the mealling of capital la7 In the properq owned b the capitalist class, which confers on ! capitalists the legal right and economic authoritv to take a share of the surplus created by the production process. Imagine for a moment the Cambridge contro\ersies as a crucial thought experiment between two competing visions of econoinics. From a Cambridge, England, perspective, how much inole decisi\e could the results have been? Capital t h e o n was the arena for extending the principle of scarcity to explain the return to capital through rnarginal productivity. It was precisely on this key point of what determines the rate of return that the "anomalous" reswitching and capitalre\el sing results occurred. The three neoclassical capital parables were sho~vll only to hold in a one-commodiq model (where classical theon was equal17 lalid). A1 attempts to extend the parable results to inore general nlodels of heterogeneous goods failed, because 'CZ'icksell effects made the links between capital and interest bidirectional rather than one-~va! lloreover, the stability literature of general . equilibrium called into question the neoclassical vision of the lifetime utilit7inaxinlizillg decisions of individuals driving an optimal allocatioll of resources through the inechanisnl of prices as scarcit: indexes. M7hat else would it take to collvillce an ecorloinist to shift visions? For neoclassicals, none of this was obvious. For them, the controversies were conducted largely in lleoclassical terins about lleoclassical models. Reswitching and capital-reversing prompted inuch useful neoclassical work to t1-1 and refine the t h e o q through secondan hypotheses and additional assumptions; Burmeister's

' In the neoclassical vision, rates of interest and profits are interchangeable terms. "English" Cantabridgians differentiate profits (the return on investment in capital goods) fi-om interest (the hire price of finance) and stress the theoretical importance of profits. Outside the one-conlmodity model, the price of capital's sei~ices-its rental-is the rate of profits multiplied by the price per unit of capital goods (Harcourt, 1972, pp. 37-39).

A U Z JCohen and G. C. Harcourt


(2000) "regular economies" are a good example. But there was little sense of a viable alternatke vision waiting in the wings and even less sense that the neoclassical ~ l s i o n at stake. was Furthermore, neoclassical one-commodity models remained intact and fruitful as a basis for empirical work. As explicit simplifications, the: could get b with the ! less rigorous notion that relatile scarcities must be the emnpimcnlb do~nznant determinant of relative prices, even if Wicksell effects are theoretically p ~ s s i b l eSolow's .~ rationale for his empirical work has always been straightforward and frank: assuming that the data mav be regarded "as if" the! were generated b the underl~ing ! simple model, the estimation procedures s e n e to provide orders of magnitude of the ke! parameters of the model. These "lo~vbro~v" models remain heuristicallv important for the intuition thev provide, as well as the basis for empirical work that can be tractable, fruitf~il and policv-rele~ant.~ In contrast, the "English" camp was sorely lacking empirical work on the causes and impact of capital inlestment. The Cambridge, England, rejoinder on the empirical issue, baffling to many "American" economists, was that the empirical likelihood of Wicksell-type feedback effects was beside the point. This was a theoretical debate, in which all models assume a gilen technology. An: time series or cross-count17 data would hale different technology bases, which could be used to explain awa! anv contra^^" empirical tests by ad\ocates on either side.'' According to Sraffa (1961, pp. 305-306): "Theoretical measures require absolute precision. Any imperfections . . . were not merelv upsetting, but knocked down the whole theoretical basis. . . . The work of J. B. Clark, Bohm-Bawerk and others was intended to produce pure definitions of capital, as required by their theories. If we found contradictions. . . these pointed to defects in the theon."ll Another weakness on the "English" side was that neither Robinson nor her fellow Cambridge critics del eloped an alternatk e set of theoretical (as opposed to descriptke) tools that avoid her concerns about the limitations of equilibrium analysis. Or e\en where thev have-we think here of Kalecki's (1968) and Good-

'Following Stigler, Cohen (1993) argues that the neoclassicals held a "93% scarcity theoq of value" after

the Cambridge controversies, akin to Ricardo's admission of a 93 percent lahor theoq of value after the
effects of capital were taken into account.
"here are important limitations on the empirical support for the neoclassical parables. Fisher (1971,
p. 325) has shown that as long as factor income shares remain constant, an aggregate Cobb-Douglas production function will fit the data well "even though the underlying technical relationships are not consistent with the existence of any aggregate production function." '"The recent endogenous growth literature o n convergence has grown out of the lack of empirical evidence for the predictions of the neoclassical one-commodity rnodel (that is, countries with higher capital/labor ratios should, but d o not, have lower rates of return and growth). Robinson's (19751>, p. .54) comment presaged this literature: "In comparisons across countries, American industry is generally found to have the highest ratio of inputs per worker . . . but I do not know that anyone has ever suggested that the rate of profit on capital is exceptionally low in the United States." D a ~ l d aidler suggests that Sraffa was ahead of his time in pointing out that neoclassical models based L on an aggregate production function lack proper microfoundations, because this is a far more telling criticism of modern real business cycle theory, whose exponents make strong claims about such matters, than it was of 1950s vintage growth models.


21 0 Journal of Et ono~nirP ~ r s p ~ c t z v ~ s

win's (196'7) cyclical growth models, Kaldor's (1996) cumulative causation processes and Pasinetti's writings (1981, 1993)-the profession by and large ignored them. Thus, the two Cambridges could not agree about the sigrlificallce of either the results nor the supporting e\idence. Such disagreeinerlts about sigllificarlce are an endernic problem in economic anal!sis. What is the meaning of a simple model whose clear-cut results are not sustained when restricti\ e assumptions are loosened? Is it nonetheless a valuable parable, useful heuristicall! and empirical17 to isolate crucial terldellcies that get obscured in rnore general models? Or is it a mistake whose insights must be discarded while searching for a better explanation in a completel! different direction? Blaug (in Caravale, 1976, p. 38) captured beautifully the typical neoclassical response to these questions: "The Cambridge School has this crazy idea, that if we ha\e a rigorous simple theon, and then we disco\er one little flaw in it, that inakes it more complicated to use it, we are finished. If we need five tyres to run a car instead of four tyres, we haven't got a car any more, so we must give up e\en.thing and start using an aeroplane." bl'ith neither side able to deli\er a knockout punch, issues of faith and ideolog entered the ring xvith clairns about the sigrlificallce of the results and coinpeting \isions of economics. 12'hen one-corninoditv results are not robust in inore general models, the lack of definitive e\.idence leaves room for i d e o l o p to pla! a role in the decision to hang on to a t h e o q or vision. The intensit) and passion of the Cambridge contro\ersies were generated not by abstract technical questions about FVicksell effects, but by strong ideological undercurrents like the ethical justification of the return to capital and fundamental methodological questions about coinparing deepl! differing ~isions economics and the extent to which equilibof rium is a useful tool of econoinic analysis. Ideology and methodology, two subjects most ecollomists would rather avoid, were pervasive ulldercurrellts fueling the contro\.ersies (Bliss, 19'75, chapter 15).

Whence and Whither the Cambridge Capital Theory Controversies?

The Cambridge controversies were the last of three great twentieth-centun. capital t h e o q controversies. Earlier contro\.ersies occurred at the turn of that c e n t u q among Bohm-Bawerk,J. B. Clark, Ining Fisher and ITeblenand then in the 1930s among Knight, Hayek and Kaldor. Similar issues recurred in all three controversies, and we will sketch some examples here.'' At the turn of the twentieth century, J. B. Clark and Biihm-Bawerk were consciously countering hlarx's theory that the return to capital in\.olved exploita-

l 2 Colien and Harcourt (forthcoming) provide a fuller treatment of these early episodes and more detail on all arguments in this paper.

Whatever Hafipened to the Cambrzdge Capital The091 Controverszes?

21 1

tion of labor.13 Clark's response, that wages and interest were simply prices stemming from the respective marginal products of labor and capital, is best expressed in his famous claim that "what a social class gets is, under natural law, what it contributes to the general output of i n d u s t l ~ "(Clark, 1891, p. 312)). Veblen disputed Clark's marginal productivity theoq, arguing instead that profit was institutionallv grounded in the social power of the capitalists that enabled them to appropriate the technological achievements of the society as a whole. Irving Fisher (1907) believed that the interest rate could be viewed as the equilibrium outcome of' simultaneous equations. Bohm-Bawerk disagreed, arguing that simultaneous equations, while useful, in\.olved circular reasoning and failed to provide a causal explanation of interest. Bohm-Bawerk, in defending an alternative Austrian vision of economics, sought a one-way explanation tracing interest determination back to the original physical factors of labor and land. In the 1930s controversies, Hayek insisted that decreases in the interest rate prompt more roundabout, capital-intensive production, even though he could not prove this in heterogeneous goods models. Hayek (1941, pp. 141-142) freely acknowledged: "All attempts to reduce the complex structure of waiting periods . . . are bound to fail, because the different waiting periods cannot be reduced to a common denominator in purely technical terms." Kaldor and Knight agreed that the inverse, rnonotonic relation between capital intensity and the interest rate is not sustained in heterogerleous commodity models. The! disagreed about which of their respective one-commodity models pro\ided better insights. All three authors (Bohm-Bawerk and I'eblen, too) expressed concerns about equilibrium. For example, Hayek's (1941, p. 17) "dpnamics" emphasized historical causation over mutual interdependence. Kaldor (1938, p. 14) argued against comparative statics and for a "process of change." Knight (1931, p. 210) believed that capital and growth are "long-run historical changes [that] must be faced as problems of historical causality and treated in terms of concepts very different from those of given supply and demand functions and a tendency toward equilibrium under given conditions." Looking back over this intellectual history, Solo~v (1963, p. 10) suggested that "when a theoretical question remains debatable after 80 vears there is a presumption that the question is badly posed-or very deep indeed." Solow defended the "badly posed" answer, but we believe that the questions at issue in the recurring capital controversies are "\.en deep indeed." The Cambridge controversies were not a tempest in a teapot. M e agree with 7 Bliss's conclusion (1975, p. 346) in viewing "the theory of capital not as some quite separate section of economic theory, only tenuously related to the rest, b u t . . . as an exterlsiorl of equilibrium theory and production theory to take into account the role of time." Major issues-explaining (and justifying) the return to capital, visions of accumulation, limitations of equilibrium tools-were and are at stake. M'hile many of the key Cambridge, England, combatants stopped asking questions
'"or a short introduction to the earliest controversy in this journal, ~vitha focus on Clark hut some discussion of the other participants, see Persky (2000).

212 Journal of Economic Perspectives

because they died, the questions have not been resolved, only buried. M7hen economists decide to delve again, we predict controversies over these questions will be revisited,just as they were time and again in the 80 years prior to the Cambridge controversies.

The authors a7e moct grateful to the edztors of thejournal, Ma7k Blaug; ChristopherBlzss, Harnld Hagemnnn, Heznz Kun, Dauid Lnidler, Muwny Milgcite nnd the pnrticipants in senzznn7s at Queens' College, Cnnz(,mdge,Brock CTniz~ersztj, York/Toronto Workshop in the the Hzrto? of Economzcs, the Histo? of Economic Thought conference at the CTniue~sity ,w of \ ' e England, i Y South Wales, and the Hzstory of Economzcs Societj meetzngs at Wake Forest ~ ~ Lizzuersztj for thezr comments on drafts of thzs artzcle. Thanks to Bonn3 Stevensenfor creating t h jgures. ~

Bharadwaj, Krishna. 1978. Classical Political Economy and Rise to Dominance of Supply and DP ?nand Theories. New Delhi: Orient Longman. Blaug, Mark. 1975. The Cambridge Reuolution: Surcrss or Failure? 4 Critical Analysis of Cambridge Throric.~of V a l u ~ Distribution, Reuised Edition. and London: Institute of Economic Affairs. Bliss, ChristopherJ. 1975. Capital Themy and thr Distribution of Income. Amsterdam and New York: Elsevier North-Holland. Bohm-Bawerk, E. von. 1907. "Capital and Interest Once More: 11. A Relapse to the Productijip-Theory." Quarter4Journal of Economics. Februaly, 21:2, pp. 247-82. Burmeister, Edwin. 2000. "The Capital The017 Controversy," in Critical Essays on Pirro Sraffaai I.rgaq zn Economics. Heinz D. Kurz, ed. Cambridge: Cambridge University Press, pp. 305-14. Caravale, Giovanni, ed. 1976. The Cambridge Drbatr on the Theoy of Capital and Distribution. Perugia: Universita degli Studi di Perugia. Champernome, David G. 1953-54. "The Production Function and the Theory of Capital: A Comment." Rn~ieu,of Economic Studies. 21:1, pp. 112-35. Clark, John Bates. 1891. "Distribution as Determined by a Law of Rent." QuartrrlyJournal o j Economics. April, 5:3, pp, 289-318. Cohen, Avi J. 1989. "Prices, Capital and the One-Commoditv Model in Neoclassical and Classical Theories." Histoy of Political Economy. Summer, 21:2, pp. 231-51. Cohen, Avi J. 1993. "Samuelson and the 93% Scarcity Theory of Value," in Tht Dynamics of the Il'ealth of LVations. Growth, Distribution and Structural Change. Essays in Honour ofLuig-2Pasinetti M . Baranzini and G. C. Harcourt, eds. London: Macmillan, pp. 149-72. Cohen, Avi J. and G. C. Harcourt. Forthcoming. "Capital Theory Controversy: Scarcity, Production, Equilibrium and Time," in Capital The09.C. J . Bliss, A. J. Cohen and G. C. Harcourt, eds., in the series Critical Ideas in Economics. K.D. Hoover and M. Blaug, eds. Cheltenham, U.K.: Edward Elgar. Dixit, Avinash. 1977. "The Accumulation of Capital Theoly." OxfordEconomic Paprrs. 29:1, pp. 1-29. Fisher, Franklin. 1971. "Aggregate Production Functions and the Explanation of Wages: A SimofEconomics and Staulation Experiment." R n ~ i m tistics. 53:4, pp. 305-25. Fisher, FrankLin. 1989. "Adjustment Processes and Stability," in The PalLgave: Genwal Equilibrium. John Eatwell, Murray Milgate and Peter Newman, eds. New York Norton, pp. 36-42. Fisher, Irving. 1907. The Rate of Interest: Its ,Vature, Determination, and Relation to Economic Phenomena. New York: Macmillan. Fisher, Irving. 1930. The The09 of Intrrest. New York: Macmillan.

Aui J. Cohen and G. C. Harcourt


Garegnani, Pierangelo. 1970. "Heterogeneous Capital, the Production Function and the The0 7 of Distribution." h i m of Economic Studies. 1 July, 373, pp. 407-36. Goodwin, Richard M. 1967. "A Growth Cycle," in 5oczalzsm Capztalzsm and Cconomzc Growth Es s a y Prrsrntrd to Llaurzcr Dobb Charles Feinstein, cd Cambridge Cambridge L nnersih Press, pp 54-58 Hahn, Frank H. 1972 Thr Share of il'uges zn thr ,Vatzonal Incomc London Tl'eide~lfeld and \icholson Hahn, Frank H. 1981 'General Equilibrium Theon," in The CTISZS hconomzc ?he09 D Bell zn and I histol, eds hew Yorh Basic Books, pp 123-38 Hahn, Frank H. 1984 Cquzlzbnu~nand dlacro econolnzcs Oxford Basil Blackwell Harcourt, G. C. 1969 Some Cambridge Contro\ersies in the Theon of Capital " Journal o j Cconomzc Lzterature 7 2, pp 369-405 Harcourt, G. C. 1972 Some Cambridge Contro vnszes zn the Theor? of Capztal Cambridge Cambridge University Press. Harcourt, G. C. 1976. "The Cambridge Controversies: Old Tl'ays and New Horizons-or Papws. 28:1, pp. 26Dead End?" OxfordEcono~nic
C. r


Hayek, Freidrich A. von. 1941. The Pure Theoq of Capztal. London: Routledge. Hennings, Klaus. 1985. "The Exchange Paradigm and the The017 of Production and Distribution," in f.bundatzons ofC'conomics. bf. Baranzini and R. Scazzieri, eds. Oxford: Blackwell, pp. 221-43. Ingrao, Bruna and Giorgio Israel. 1990. The Invisible Hand: C'conomic Equilibrium in the Hzstol?; ofscience. Cambridge: MIT Press. Kaldor, Nicholas. 1938. "On the The017 of Capital: A Rejoinder to Professor Knight." EconoA met~ica. pril, 6:2, pp. 163-76. Kaldor, Nicholas. 1996. Causes of Growth and Stagnation in thr lZhrld Economny. The 1984 R a f a ~ b Cambridge: Cambridge UniverA!JattioliI.c>cturc>s. sity Press. Kalecki, Michal. 1968. "Trend and Business Cycles Reconsidered." C'conomzcJournal. June, 78: 310, pp. 263-76. Knight, Frank H. 1931. "Professor Fisher's Interest Theory: A Case in Point." Journal ofPolztzca1 Economy. April, 39:2, pp. 176-212. Pasinetti, Luigi L. 1969. "Switches of Technique and the 'Rate of Return' in Capital The017." Etonomzc Journal. September, 79:315, pp. 503-31. Pasinetti, Luigi L. 1970. "Again on Capital

Theon and Solow's 'Rate of Return "' Economzc Journal June, 80 '318, pp 428-31 Pasinetti, Luigi L. 1981 5tructural Change and hconomzc Growth 4 Throretzcal Essa) on the D3namzcs of the ltealth of LVatzons Cambridge Cambridge Uni~ersihPress Pasinetti, Luigi L. 1993 Str~~cturalhconomzc Djnamzcs A Theop of the Ccono~nzcConsequences of Human I rarnzng Cambridge Cambridge UniLerslh Press Pasinetti, Luigi L. 2000 "Crihque of the Neoclassical Theory of Growth and Distribution " Banca ,Vazzonale del Lazloro Quartwb h u m 53 215, pp 383-431 Persky, Joseph. 2000 "The Seoclassical Advent: American Economics at the Turn of the 20th Centul7." Journal of Economic Perspectives. M'inter, 14:1, pp. 95-108. Robinson, Joan. 1953. "Imperfect Competition Revisited." C'conomzcJournal. September, 63: 251, p p 579-93. Robinson, Joan. 1953-54. "The Production Function and the T h e o n of Capital." R n ~ i m of C'conomic Studies. 21:2, pp. 81-106. Robinson, Joan. 1980 [1974]. "History Versus Equilibrium," in Collected C'conomic Papers, Volume 5. Cambridge, Mass.: MIT Press, pp. 48-58. Robinson, Joan. 1975a. "The Unimportance of Reswitching." Quarterb Jourrzal of Economics. Februan, 89:1, pp. 32-39. Robinson, Joan. 1975b. "Reswitching: Reply." Quartrly Journal ofEconomics. Februalv, 89:1, pp. 53-55. Robinson, Joan. 1980. Collected Economic Paprrs, lblu~nes1-5. Cambridge, Mass.: MIT Press. Samuelson, Paul A. 1962. "Parable and Realism in Capital T h e o n The Surrogate Production Function." R n ~ i e uofEconomic Studies. June, 293, pp. 193-206. Samuelson, Paul A. 1966. "A Summing Up." Quarterb Journal of C'conomics. November, 80:4, pp. 568-83. Schefold, Bertram. 2000. "Paradoxes of Capital and Counterintuitive Changes of Distribution in an Intertemporal Equilibrium Model," in C ~ i t ical Essays on Pirro Sraffak I,c>gacyzn C'conomics. H. Kurz, ed. Cambridge: Cambridge University Press, pp. 363-91. Solow, Robert M. 1955-56. "The Production Function and the Theory of Capital." h i m of Economic Studies. 222, pp. 101-08. Solow, Robert M. 1956. "A Contribution to the The017 of Economic Growth." QuarterlyJournal of Economics. Februalv, 70:1, pp. 65-94. Solow, Robert M. 1957. "Technical Change and the Aggregate Production Funchon." Ra~zeu of C'conomzcs and Jtatzstzcs. 39 3, pp 312-20.


Journal of Economic Perspectiues

Solow, Robert M. 1963 Capztal ?heal?, and thr Ratt of Return Amsterdam North-Holland Solow, Robert M. 1970 ' O n the Rate of Re" Economzc Journal June, turn Replv to Pas~nett~ 80 '318 pp 423-28 Solow, Robert M. 2000 "The Neoclass~cal a \ Theon of Growth and Distribution " Banca L zzonalr drl Lazloro Quarter4 Rn~zeu,December, 53 215, pp 349-81 Sraffa, Piero. 1936 Letter to Joan Robinson, October 1936 k c h i ~ e s ,IGng's College, Cambridge Sraffa, Piero. 1960 Productzon of Co~n~nodztzes hi Ltleans of Commodztzes Prelude to a Critzque of Lco nomzc Theon Cambridge Cambridge L ni\ ersit) Pre\\ Sraffa, Piero. 1961 Comment," in 7he The09

of Capital. F.A. Lutz and D.C. Hague, eds. London: blacmillan, pp. 305-06. Sraffa, Piero. 1962. "Production of Commoditia: a comment." Economic Journal. June, 72:286, pp. 477-79. Swan, Trevor W. 1956. "Economic Growth and Capital Accumulation." Economic Record. 322, pp. 343-61. Veblen, Thorstein. 1908. "Professor Clark's Economics." Quarterly Journal of Economics. Februaly, 22:2, pp. 147-95. Walsh, Vivian and Harvey Gram. 1980. Classical and LVeoclassical Theories of Genwal Equilibrium. New York: Oxford University Press. Wicksell, Knut. 1911 [1934]. I.ectures on Political Economj, Volume 1. London: George Routledge & Sons.