keynes labor market (hoover)

7/29/2019 Keynes Labor Market (Hoover)

1/33

Relative Wages, Rationality, andInvoluntary Unemployment in KeynessLabor MarketKevin D. Hoover

I dont like that book. Its not a very congenial book to read. . . . Ifind i t carelessly written, not especially gracefully written, sometimesdishonestly written. -Robert E. Lucas, Jr. in The New ClassicalMacroeconomics: Conversations with the New Classical Economistsand Their Opponents,by Arjo KlamerKeynes was a lucid and resourceful master of English prose, as wereSmith, Bentham, Malthus, the two Mills, Marshall and Veblen. Ricardopossibly excepted, no one of great importanceinthe history of English-speaking economic thought was otherwise. The General Theory ofEmployment Interest and Money, however, is a complex, ill-organizedand sometimes obscure work, as Keynes himself recognized. -JohnKenneth Galbraith,Economics in Perspective: A Critical History

The reputation of John Maynard KeynessGeneral Theory of Employ-ment, Interest, and Money as a badly written book is often exaggerated.But if it is deserved at all, it is because of parts such as chapter 2, ThePostulates of the Classical Economy. Half a century of exegesis andinterpretation have yet to provide a satisfactory and widely accepted ver-Correspondence may be addressed to Professor Kevin D. Hoover, University of California,Davis CA 956 6 or by e-mail to ([email protected]). An earlier version of this paper waspresented at the History of Economics Society meetings in Lexington, Virginia, 23-25 J une1990. thank Thomas Mayer, Nancy Wulwick, Mark Perlman, Steven Sheffrin, William Darity,Axel Leijonhufvud, Victoria Chick, Bradley Bateman, M ichel de Vroey, Don Patinkin, and ananonymous referee for helpful comments on earlier drafts.History of PoliticalEconorny 27:4 @ 1995 by Duke University Press. CCC 0018-2702/95/$1S O


2/33

654 History of Political Economy 27:4 (1995)sion of what this chapter really means. Reactions to chapter 2 fall intofour main strands.One strand holds that Keynes cared very little for the analysisof aggre-gate supply, which is the heart of this chapter, and that he rather botchedthe job in an effort to get on more quickly to the analysis of aggregatedemand, which constitutes the bulk of the book and is its main innova-tion. This view draws some support from K eyness confession to JacobViner that this part of my book is particularly open to criticism (1937,210).

The second strand is related to the first. Some modern commentatorsstress that K eyness essential insight and innovation was his analysis ofaggregate demand (for example, Patinkin 1982, chap. 5, 1987; Tobin1992). Some (for example, Patinkin) recognize that K eyness analysisof aggregate supply also faces interpretive problems. Still others do notregard his analysisof aggregate supply as problematic but rather as fun-damentally insightful and relevant, posing no great interpretive puzzles.The difficulty with this untroubled point of view is not that it is wrongabout Keyness priorities or innovations, but that it fails to appreciatefully the strength of the more hostile reactions to chapter 2 and the needto provide greater clarification of Keyness analysis.Both the first and second strands are too facile, for the very generalityof The General Theory hinges on Keyness dissent from the classics.It takes pride of place as the first substantial chapter of the book; andit forms the linchpin of his economic system. It is here that he tries tojustify his crucial assumption that money wages do not necessarily moveinsuch a way as to clear the labor market. And it is here that he formulatesthe now customary tripartite division of unemployment into voluntary,frictional, and involuntary unemployment.

A third strand is represented by authors such as William Darity andBobbie Horn (1987a, 1987b) M ichael L awlor, Darity, and Horn (1987)and Michel de Vroey (1991), who all argue that chapter 2 is essentiallya mistake. These authors draw a distinction, to use de Vroeys termi-nology, between labor rationing, which is the subject of chapter 2, andunderemployment equilibrium, which is the real subject of The GeneralTheory. The labor market clears in an underemployment equilibrium:labor (and other resources) are fully but not efficiently employed. A nincrease in aggregate demand that redirects malemployed resources(again, de Vroeys term) to better uses increases output and total em-ployment. According to this view, chapter 2could be omitted from The


3/33

Hoover / Keyness Labor Market 655General Theory, rendering Keynes more consistent and freeing up theresources devoted to sorting out the hopelessly tangled skein of its argu-men t.In contrast, a fourth strandof interpretation is that non-market-clearingmoney wages and involuntary unemployment ultimately depend inKeyness system on irrationality. Workers, it is said, suffer from moneyillusion. This view is important as an early reaction to The General The-ory. Economists have never been at home with irrationality, and lately i thas become anathema, not only to the direct heirs to the classical traditionbut also to those who see themselves as the heirs to K eynes. This viewhas renewed its lease with the riseof the new classical macroeconomics.Robert Lucas has goneso far as to argue that the concept of involuntaryunemployment is vacuous and that it draws a distinction with no coun-terpart in the real world. And, as Alessandro Vercelli says, after all, ifKeynes were writing The General Theory in the eighties he would proba-bly have chosen Lucas rather than Pigou as his Turks head (199 , 243).This article reconsiders chapter 2. My method is to provide a closereading of that chapter. The central result of that reading is to take se-riously Keyness repeated claim that workers are concerned with theirwages relative to other workers. I argue that concern for relative wagesis not a secondary matter in Keyness analysis, as some commentatorssuggest, but is, in fact, fundamental. The issue presents to me a seriesof questions: does the relative-wage argument make sense? does it ob-viate the need to appeal to money illusion? what is its place in Keynesssystem? To answer these questions I develop a rational reconstruction ofKeyness argument ina modification of the now-familiar efficiency-wagemodel.It is important to recognize that the model developed in this rational re-construction isnot an attempt to lay hold of the mantle of Keynes for theefficiency-wage model, a central element in the so-called New Keynes-ian Macroeconomics. New Keynesian efficiency-wage models are eitherreal-wage models in which demand-side policies are ineffective or mod-els that explicitly rely on money illusion to yield efficacious aggregate-demand policies. Thus, although the efficiency-wage model providessome ready-made structure on which to hang K eyness arguments, farfrom stretching to find antecedents that validate the K eynesian roots of

1 . To argue, asI shall presently,that Keynes does not assume money illusion is itself nothingnew in the interpretation of Keynes. I t is well worth restating this shopworn point since theerror of attributing adherence to money illusion to Keynes persists after nearly sixty years.


4/33

656 History of Political Economy 27:4 ( I 995)the efficiency-wage model, it is closer to thetruth to see Keyness analysisas suggesting ways to improve and develop an otherwise non-K eynesianmodel.One objection might be that rational reconstructions of this sort arenot really history.2 To which 1 might then reply, Why draw such over-nice distinctions if such exercises are nevertheless informative about thestructure of Keyness argument? Interestingly, however, the method isexplicitly endorsed by Keynes himself. In a letter to Ralph Hawtrey,Keynes writes, Anyhow, the essential point is that my discussion of theclassical theory of interest is merely an attempt to make an hypothesisas to what the classical theory must be if it is to be consistent with theirgeneral views (1973,25). As K eynes was to classical interest rate theory,so is my reconstruction to the analysis of the Keynesian labor market.The model developed below is cast in a modern-looking guise, but thequestion itseekstoanswer is this: what did K eynes think the labor markethad to be like in order for it to be consistent with his general views? Themodel I propose is justified step by step by Keyness insights and itgenerates unemployment exactly in the circumstance in which aggregatedemand is inadequate. It reproduces the relationships between nominalwages, prices, and real wages found in the early chapters of TheGeneralTheory.While the model is an attempt to explicate the relativity argumentof chapter 2, it is not inconsistent in any fundamental way with Keynessmore complete analysis of wage and price behavior in chapter 19.Thefact that I try tounderstand chapter 2and its role in the whole story doesnot in any way imply a devaluation or failure to appreciate the rest of thebook, especially K eyness analysis of aggregate demand and chapter 19.

The power of my interpretive strategy lies in the way in which it helpsto sort through the four strands of interpretation already identified. Thecomplacency of the second strand should surely be shattered by the recog-nition that the proponents of the third and fourth strands do not acceptthe coherence of chapter 2. Proponents of the third strand, from Leon-tief through Friedman and Lucas to the real business-cycle modelers,cite an inadequate analysis of aggregate suppfy as the principal reasonfor rejecting Keyness analysis. The new classical macroeconomists, theintellectually dominant school of macroeconomics today, see this as theprincipal cleavage between themselves and K eynesians. It would be use-

2. This is the essence of Don Patinkins view of the methodology of the current article(Patinkin, personal communication toauthor, 13 August 1992).


5/33

Hoover / Keyness Labor Market 657

ful to know if Keynes can be justly convicted of the flaw that they thinkthat they have located in his analysis. The purpose of this article is toshow that he cannot besoconvicted. The proponentsof the fourth stranddo not dismiss Keynes as the new classicals do, but still do not findchapter 2 consistent within itself or with the rest of The General The-ory. I think that it is both. Therefore my purpose, and my only purpose,is to clarify chapter 2 and to demonstrate how it hangs together. I donot promote a modern model for its own sake-certainly not the bread-and-butter efficiency wage model that, despite similarities to my model,misses Keyness point as surely as do the new classicals. To the extentthat my interpretive strategy succeeds, that is to the extent that Keynessanalysis is shown to be coherent and to serve the main purposes of therest of The General Theory,the skepticism of the first interpretive standcan be rejected as well.

1.Keyness Dissent from the ClassicsTo place the importance of Keyness relative-wage argument, it is usefulto begin on familiar ground. Keynes characterizes the classical theoryof employment-indeed, the core of classical economics-as resting ontwo fundamental postulates:

I .11.

The wage is equal to the marginal productof labor:The utilityof the wage whena given volumeof labor is employedis equal to the marginal disutilityof that amountof employment.(1936,5)3

The first classical postulate is stated more exactly as themoneywage isequal to thevalueof the marginal product of labor. The second can beput into more modern terminology: the real wage is equal to the marginalrateof substitution between consumption goods and lei~ure.~

Keynes accepts the first classical postulate at least for purposes ofargument (GT,17).The heart of his dissent from the classics is that herejects the second classical postulate, except as a limiting case that definesfull employment (7-13, 15,28). Keynes gives two reasons for rejectingthe second postulate. The first is that the actual attitude of workers issuch that the money wage as well as the real wage may be important

3. Hereinafter references to Keynes 1936will appear as GT.4. Keyness statement of postulate I 1 is inexact. He surely means themarginal utility of thewage, rather than the utility of the wage (see Fender 1981,14n. I ) .


6/33

658 History of Political Economy 27:4 (1995)

to them (%lo), although he says this is not a fundamental reason. Thesecond reason, which he does say is fundamental, is that wage bargainsare madein terms of money wages,sothat even if workers accept cutsinmoney wages these may not result in cuts in real wages or lead to greateremployment (10-12).

Presently, I will consider Keyness reasoning in greater detail. Forthe moment, consider just one aspect: Keynes argues that workers areconcerned not only with what their wages will buy but also with theirrelative position with respect to other workers (13-15).5 A cut in oneworkers (or a group of workers) money wages, if other workers wagesremained unchanged, would erode his (or their) relative position. Becausecuts are seldom uniform, individual workers or groups have reasontocareabout their money wages independent of their real wages.

Keyness argument about relativities has typically been associated withhis non-fundamental objection to the second classical postulate: namely,that the actual attitude of workers is concern for nominal wages. SinceI will argue in due course that it is important to associate this argumentwith his fundamental objection (namely, that wage bargains are made interms of money), i t is critical to observe that this argument is developedin section 3 of chapter 2-following the main discussion of his twoobjections and well separated from the non-fundamental objection. Itsplace in the text is ambiguous; it is not at all clear that it refers only, oreven primarily, to the non-fundamental objection.

Having rejected the second classical postulate, Keynes is in a positionto introduce the concept of involuntary unemployment. His definition iscommonly taken to be that a worker is involuntarily unemployed whenthe real wage offered similar workers na ob he is qualified to do is higherthan his marginal rate of substitution between consumption and leisure:that is, he would work for the going wage, or even somewhat less, were ajob offered to him. Keynes himself sometimes refers to the involuntarilyunemployed as having a marginal disutility of labor below the goingwage (for example, 10, 14). Nevertheless, he does not use this conditiondirectly to define involuntary unemployment. Instead, he adopts whatappears at first sight to be an extremely convoluted definition:Men areinvoluntarily unemployed ij in the event of a small rise in the price ofwage-goods relatively to the money-wage, both the aggregate supply oflabour willing to work f or the current money-wage and the aggregate

5. K eynes makes this argument as early as 1925 ([I9311 1973,211).


7/33


demand for it at that wage would be greater than the existing volume ofemployment (15).John Nicholas Smithin (1987) and Victoria Chick (1987, 72, 73)demonstrate clearly that this definition implies the well-known inequalitybetween the real wage and the marginal rate of substitution for consump-tion and leisure. I will argue presently that, far from being convoluted,Keyness definition is perfectly adapted to answer the classical objectionthat all unemployment is really voluntary, which has recently been force-fully restated by Lucas and other members of the new classical school.

2. Some Standard InterpretationsInterpretations of Keyness dissent from the classics that center on ag-gregate demand (for example, Hicks[19371 1967; Friedman 1974) to thecontrary notwithstanding, Keyness own claims in chapter 2 (and chap-ters 3 and 18) suggest that an acceptable interpretation should centeron aggregate supply and the labor market. This by no means denies thatKeyness greatest innovations may have beenin the areaof aggregate de-mand, as Patinkin, for example, maintains (1982, chap.5; 1987). Rather,it underlines that an economy with a Keynesian aggregate demand sectorand a classical aggregate supply sector would nevertheless not delivercharacteristically Keynesian behavior. Critics such as Patinkin do notdeny this point; but they do take the adequacy of Keyness labor marketanalysis for granted.A long interpretive tradition questions its adequacy.It is to those critics that I now turn.

2.1 Money IllusionInvoking such passages as within a certain range the demand of labouris for a minimum money-wage and not for a minimum real wage (GT,8), Leontief (1936) interprets Keynes as assuming the supply functionfor labor is not homogeneous of degree zero inprices and wages or, in thecommon parlance, that workers suffer from money illusion. Becausethey do not understand the effect of prices on real wages, they willinglysupply labor when real wages fall because of a riseinprices, although theywithdrawit when real wages fall because of a cut in money-wages. Thisinterpretation still retains great currency n textbooks of macroeconomicsand the history of economic thought (see, for example, Blaug 1985,663-65;compare to note 1above).


8/33

660 History of Political Economy 27:4 (1995)Leontief notes that the homogeneity postulate is not simply assumed

in classical economics but is derived from more fundamental assump-tions, which, he believes, Keynes does not directly attack. The funda-mental assumption that delivers homogeneity is the rationality of theworker: workers maximize utility derived from real goods subject tobudget constraints.6 Leontief can be read, therefore, as asserting thatKeynes assumed workers to be irrational. Professional unease with thefoundations of K eynesian economics is reflected in the widespread accep-tance of Leontiefs interpretation: rationality is the keystone of economicanalysis.While it is clear that Keynes was willing to entertain the possibilitythat workers sometimes act under money illusion, this was not essential;and he took pains to claim that his argument works when workers arefully rational (GT,9, 14).Beyond this, however, all passages that seemto support the interpretation based on money illusion refer to Keynessnon-fundamental objection to the second classical postulate. In otherwords, even under the best case for his critics, money illusion in Keynessown view could not be theoretically fundamental, whatever its practicalimportance.

2.2 Rigid WagesBlaug (1985,663) asserts that the importance of money illusion is as anexplanation for wage rigidities. He goes on to claim that any other sourceof rigidity, such as powerful trade unions or minimum wage laws, willdo just as well to support the Keynesian model of income determination.Whatever the merits of these claims with respect to income determina-tion, it is clear that trade unions, minimum wage laws or, for that matter,money illusion will not support Keyness analysis of involuntary unem-ployment. Keynes (GT,5, 8, 16) concedes to the classics that unemploy-

6.This construes rationality in an extremely narrow sense, one completely appropriatetothe nature of the issues addressed in this paper but one far too narrow to do justice either toKeyness rich epistemology ortorecent Keynesian scholarship (cf. Carabelli 1988 and Bateman1987).7. Some evidenceto the contrary might be found in the introduction to Keyness reply to hiscritics (1937.209) inwhich he concedes the theoretical correctness of Leontiefs point. I do notthink that this should be taken too seriously as Keynes never develops his concession in thatpaper or, I think, elsewhere. In this case, he really does seemto want to push quickly throughthe discussion of aggregate supply toget on with the theory of investment and interest, whichare his chief concerns on that occasion.


9/33

Hoover / Keyness Labor Market 661ment resulting from obvious market imperfections such as trade unions,closed shops, and minimum wage laws is rightly classified as voluntary.That unemployment due to money illusion is also voluntary can be in-ferred from Keyness discussion of the second classical postulate. Disu-tility [of labor], he writes, must be here understood to cover every kindof reason which might lead a man, or a body of men, to withhold their la-bor rather than accept a wage which had to them a utility below a certainminimum (6). Money illusion renders the labor supply curve nonhomo-geneous; butso long as the supply of labor actually forthcoming is deter-mined on the labor supply curve-homogeneous or not-unemploymentis voluntary.Whatever justification might be given for rigid wages, it is clear thatK eynes did not mean his theory to depend upon them. To be sure, heassumes that money wages are fixed in chapter 3 but only with the caveatthat this is a simplification to be relaxedindue course (27). And chapter 19is entitled Changes in Money-Wages.

2.3 Heterogeneous GoodsThe obvious alternative to concentrating on Keyness non-fundamentalobjection as the arguments from money illusion and fixed money wagesdo, is to try to make sense of his fundamental objection. K eynes deniesthat the wage bargains between the entrepreneurs and the workers de-termine the real wage so that there is no longer any reason to expecta tendency towards equality between the real wage and the marginaldisutility of labour ( 1 1).Leijonhufvud (1968,97,98) argues that K eynes wants to draw a fun-damental distinction between monetary and barter economies. In mone-tary economies, demands must be communicated through offers to paymoney.A worker may desire to work at a real wage less than his marginalproduct, but a firm will not hire him unless it sees demand in the formof money on offer for its output. This demand cannot simply be directbarter with the worker: he cannot feed his family on a ton-and-a-half ofcold-rolled sheet a week (90).Leijonhufvuds argument is an application of Clowers ( 1 965) dise-quilibrium analysis.8 As such, it is most appropriate to the analysis of8.T his was prefigured in Patinkin [19561 1965. Clower himself shows some distaste for the

term disequilibrium and for the ways in which his ideas were developed by Robert Barroand Herschel Grossman (197 ; 1976) and others. He only reluctantly admits to being the


10/33

662 History of Political Economy 27:4 (1995)quantities traded, given prices that are not market clearing. In itself, itdoes not explain why a firm does not cut its real wages andsoreduce anygap between the real wage and the marginal disutility of labor. It is notenough to say that demand is deficient at current prices. For the outputof the firm can be sold atsome lower price, however small, and that pricedetermines the value of the workers marginal product and the demandfor labor. If labor isforthcoming at the implied lower wage, the firm hasevery incentive to employ it; if it is not forthcoming, labor is voluntarily,not involun taril y, ~nemployed.~

Although Leijonhufvuds analysis is incomplete, itremains important.It is an attempt to take seriously the truth that modern economies arecomplex with many heterogeneous goods and that money is necessaryto cope with such heterogeneity. After all, in a subsistence economy,workers eat their own output, sodemand failures are not an issue.Victoria Chick (1987, 141, 142) also appeals to heterogeneous goodsin order to interpret Keyness fundamental objection. She points out thatfrom the workers point of view, the real wage depends on the pricesof the many goods he consumes and not on the price of his employersoutput. Surely the worker must bargain for money wages, not real wages,because the real wage cannot even be known until after the worker is paid.While it is clear that Chicks point is correct, it is not obvious thatit is enough to explain persistent, substantial deviations from full em-ployment. A ny form of indexation would be some improvement oversimple money-wage bargains. K eynes could not have been unaware ofvarious experiments with indexation (for example, those in the Britishcoal industry just after World War I ).

2.4 Nominal Wage AdjustmentFender interprets Keyness fundamental objection as an argument notabout static equilibrium but rather about dynamic adjustment (Fender1981, 20-23). The classics assume that, when the labor market is notin equilibrium, the real wage adjusts to clear the market. Fender readsKeynes as accepting the classical description of static equilibrium as afunction of real wage, but as arguing that, when the labor market is not inequilibrium, the nominal wage adjusts. Since profit-maximizing pricesgrandfather, if not the father, of these models (see Clower 1984a.267).

9. I take this tobe the gist of Fenders(1981, 3 , 32) dismissal of L eijonhufvuds argument.


11/33

Hoover / Keyness Labor Market 663are largely determined by wage costs, prices tend to fall along with fallingnominal wages. The real wage may, therefore, not fall-or even rise-sothat there is no tendency to clear the labor market. Fenders interpretationis supported in part in The General Theory. Keynes writes, Thus, ifmoney-wages change, one would have expected the classics to arguethat prices would change in almost the same proportion, leaving the realwage and the level of unemployment practically the same as before (GT,12). In a footnote to this passage, Keynes writes, This argument would,indeed, contain, to my way of thinking, a large element of truth, thoughthe complete results of a change inmoney-wages are more complex (1 2,n. 1); he then points the reader to chapter 19,in which the complexity ofwage/price interaction is explored. Fender recognizes that this dynamicargument is not Keyness whole story; for he writes: The argumentisnotthat prices will change to exactly the same extent as money-wages, hencekeeping real wages constant. The argument is rather the much weaker onethat money-wage adjustments may not produce the adjustment in the realwage required to restore full-employment equilibrium (198 1,22,23).Patinkin (1982, 14142; 1987,27, column 2) argues, against interpre-tations such as Fenders, that K eynes did not find a direct causal linkbetween nominal wages and prices but rather an incompletely speci-fied causal chain connecting nominal wages through aggregate demand,unemployment, and the marginal product of labor (invoking the first clas-sical postulate) to prices. Fenders argument can be seen as insufficientfor another reason as well. If, as Fender supposes, the nominal wage fallswhen labor is in excess supply, the real wage will move toward marketclearing if prices also fall less than proportionally. If they fall propor-tionally, the same situation of excess supply will continue; conversely, fthey fall more than proportionally, unemployment will become progres-sively worse. In the first case, the classics would be correct: the marketclears. Neither Fender nor K eynes puts stock in the second case of perfectproportionality. And there is no evidence that K eynes thought that thethird case, which amounts to progressive collapse of the labor market,was in any sense general. Rather K eynes thought that an outstandingcharacteristic of the economic system in which we live [is that] whilst itis subject to severe fluctuationsin respect of output and employment, it isnot violently unstable. Indeed it seems capable of remaining in a chroniccondition of sub-normal activity for a considerable period without anymarked tendency either towards recovery or towards complete collapse(GT,249).


12/33

664 History of Political Economy 27:4 (1995)Clearly, Fender is right to ascribe the dynamic argument to K eynes;

but it cannot be the whole of his fundamental objection. M ore is neededto explain the subnormal activity that Keynes believed to be the chroniccondition of many economies.3. The Placeof the Relative-Wage ArgumentWhat more is needed, I believe, is attention to K eyness oft repeated-indeed emphasized-point that i t is the relative wage rate as well asthe real wage rate that matters to workers: any individual or group ofindividuals, who consent to a reduction of money-wages relatively toothers, will suffer a relative reduction in real wages, which is sufficientjustification for them to resist it (GT,14; compare to 252, 253, 264).

As I have already observed, whether this claim is meant to supportKeyness non-fundamental or fundamental objection to the classics isunclear. The reason, as I show presently, is that it is meant to support bothobjections. To associate the relative-wage argument exclusively with thenon-fundamental objection, as many commentators do without adequatetextual warrant, renders Keyness argument in chapter 2 obscure andinconsistent, or simply incorrect. To associate it with the fundamental,as well as the non-fundamental, objection reveals Keyness argument tobe both consistent and richly suggestive.

3.1 Two ConfusionsTwo confusions arise from K eyness claim for the importance of rela-tive wages. The first is that resistance to money-wage cuts is organizedthrough unions. o The confusion arises because Keynes writes such sen-tences as: The effect of combination on the part of a group of workersis to protect their relative real wage (GT, 14) and Every trade unionwill put up some resistance to a cut in money-wages, however small(15). But these are elaborations on a more general point. The quotationabove clearly states individuals or groups of individuals, so that anysuccessful interpretation of Keyness point must apply with equal forceto unorganized as well as organized labor.

10.Fender (1981, 18) seems to read Keynes this way, although it is not clear that he meansto imply that only unions resist money-wage cuts.


13/33

Hoover / Keyness Labor Market 665The second confusion about Keyness claims for the importance of

relative wages is the belief that worker resistance should be glossedas strikes or withdrawal of labor. This would make nonsense ofKeyness analysis of involuntary unemployment; for earlier, he con-cedes to the classics that if people are unemployed due to an openor tacit agreement amongst workers not to work for less, and if labouras a whole would agree to work for less more employment would beforthcoming . . . such unemployment, though apparently involuntary,is not strictly so,and ought tobe ncluded under the . . .category of vol-untary unemployment due to the effects of collective bargaining, etc.(8). Inasmuch as the claim about the importance of relative wages is in-tended to explain involuntary unemployment, worker resistance cannotbe entirely a matter of strikes or withdrawal of labor, as these are indica-tors of voluntariness. Of course the quotation also reinforces the earlierpoint that the claim about relative wages is not simply a claim aboutthe behavior of unions. To put it another way, open or tacit agreementsnot to work for lower than current money wages modify the shape ofthe labor-supply schedule, but they do not cause involuntary unemploy-ment because that requires workers to be off their labor-supply sched-ules.

It is clear, therefore, that, while Keynes was willing to believe thatrelative wages affect the labor-supply schedule, he must also have be-lieved that they affect labor demand; otherwise the claim about relativewages lends no support to his central contention that involuntary unem-ployment is not only possible but important and common. Of course, ifrelative wages affect labor demand,,they must do so, in Keyness view,because of worker resistance. Worker resistance in turn must be con-strued more widely than strikes or other withdrawals of labor if theresulting unemployment is to be involuntary.3.2Labor SupplyKeynes confronts heterogeneity in the economy squarely. In introducinghis claim about relative wages, he says, there is imperfect mobility oflabour, and wages do not tend to an exact equality of net advantagein different occupations (GT, 14), which is to say that he takes theheterogeneity of the labor force seriously. His model of the economy isnot the single-sector model implicit in textbook IS-LM-A S K eynesianmodels. At minimum, it is a two-sector model with a capital-goods sector


14/33

666 History of Political Economy 27:4 (1995)and a consumption-goods sector. In the same spirit, relative wages maybe introduced into the labor-supply decision in order to clarify Keynessnon-fundamental objection.The utility of a representative worker in the consumption-goods sectorcan be written

where C is consumption, L is labor and w,/wk is the ratio of the wagein the consumption-goods industry (w,) to the wage in the capital-goodsindustry (wk).The labor-supply schedule can be derived in the usualway: the worker chooses consumption and leisure, taking wage rates asgiven, subject to the budget constraint,

w,L - pcc=0,wherepc is the price of the consumption good. The resulting labor-supplyschedule will, in general, have the form

that is, labor supply depends (directly) on the real wage (as judged bythe worker) and (directly) on relative wages.A property of the labor-supply function with relative wages is that ifthe labor supply is plotted against real wages then, holding wk constant,any change in w, will shift the entire curve. Fender (1 981, 17, 18) makesa similar observation to confirm Keyness statement that when the labor-

supply curve is a function of nominal wages, the supply curve for labourwill shift bodily with every movement of prices (GT,9). Fender musttake Keynes to include wages in prices if this explanation is to makesense. Context gives no reason to suppose that K eynes in fact meantwages, especially when one follows up the reference to the appendix tochapter 19as suggested in a footnote to this passage,The appendix suggests an alternative interpretation consistent with rel-ative wages inthe labor-supply function. Fluctuations in investment causefluctuations in the demand for labor in the capital-goods sector. Employ-ment and therefore wages in the capital-goods sector (wk) luctuate. Thisin turn causes fluctuations in the demand for and price of consumptionI I . I nterestingly, J ohn Hickss ([19371 1967) original IS-L M model is a two-sector model.Santi Chakrabarti (1979) also stresses the two-sector natureof K eyness model.


15/33


goods. A change in pc,which changes the real wage, of course leads tomovements along the supply curve for labor in the consumption-goodssector. But the change in wk,which changes relative wages, also causesthe curve to shift. Thus, just as Keynes argues, the supply curve shiftswith every change in prices.

Observe that when labor supply depends on relative wages, workersin the consumption-goods sector will wish to withdraw their labor whentheir wage falls even if prices fall proportionately, leaving the real wageunchanged. Also observe that this is fully rational inthe sense of Leontief.If all prices, including all wages, change by equal proportions, thenneither real wages nor relative wages are affected, so the labor supplyis itself unchanged. The utility function is homogeneous of degree zeroin prices and wages, and this implies that the labor-supply function isalso homogeneous. Keyness belief that no irrationality is involved inworkers concern about nominal wages is vindicated: they do not sufferfrom money illusion.

So far so good. Still, if employment is given as the intersection oflabor supply and labor demand, even if relative wages appear in thelabor-supply function, all who wish to be employed are employed; thereis no involuntary unemployment. Equally, if disutility comprises all thereasons one might have to withdraw labor, equality between the sup-ply of and the demand for labor-even if relative wages appear in thelabor-supply function-rules out the real wage exceeding the marginaldisutility of labor. Keyness non-fundamental objection, interpreted as apoint about the importance of relative wages in labor supply, does notthen by itself explain involuntary unemployment.

Keynes claims that, when labor supply depends on relative wages, theclassical labor market is indeterminate. Fender interprets this as resultingfrom the shifting of the labor supply curve with changesinwages (prices)as discussed above. He points out that the system is quite determinateeven with a shifting labor-supply function, so that Keyness non-funda-mental objection is simply wrong, and i t is fortunate that he laid notheoretical weight on it (Fender 1981, 18). Once again, if we followKeyness suggestion to consult the appendix to chapter 19, t is clear thathis point is not about the shifting labor-supply functionper se but aboutthe way in which the classics achieve closure in their models by assumingthat some mechanism ensures full employment(GT,274).If they are notentitled to assume this, then Keynes is correct, their labor markets areindeterminate; and, if they are entitled, then Fender is correct, the labor


16/33

668 History of Political Economy 27:4 ( I 995)market is determinate even when relative (or nominal wages) appear inthe labor-supply function.This suggests, however, that the reason the objection is non-fundu-mental is that Keynes knew that it did not by itself explain unemployment.All Keynes wished to show, and all he showed satisfactorily, was thatworkers might rationally care about nominal wages independently ofreal wages. The classics have no right in K eyness view to insist thatonly real wages count. It is a matter of the actual attitude of workerstowards real wages and money-wages respectively and is not theoreticallyfundamental (GT,8).Clearly the theoretical weight must rest on K eyness fundamental ob-jection: workers bargain for money wages, not for real wages. Workers,Keynes reiterates, resist cuts in their relative wages and, therefore, cuts intheir nominal wages. As already shown, if such resistance were merelystrikes or withdrawal of labor, then it would form no basis for an ex-planation of involuntary unemployment; yet K eynes associates workerresistance with involuntary unemployment. For this reason, if for noother, it is clear that the claim about the importance of relative wagesmust be associated with the fundamental objection (as well as with thenon-fundamental objection).4.A Relative-Wage Model of the Labor MarketWhat form does worker resistance take? Keynes does not go into detail,but possibilities are slowdowns, sabotage of equipment, lack of cooper-ation with management, or simply poor performance due to low morale,all lowering productivity. This immediately suggests the modern theoryof efficiency wages. I will show that it is possible to give a consistent andcomplete interpretationof Keyness fundamental objection as a relative-efficiency-wage model.I2It may be worth reiterating that the interpretive strategy here is to useconsiderations explicitly discussed in The General Theory to adapt theefficiency-wage model into a vehicle to showcase important features of

12. Akerlof and Yellen 1986 (especially the introduction) and Stiglitz 1987 provide excellentintroductions to efficiency-wage models. Akerlof and Yellen say that efficiency-wage modelsare necessarily models of the determination of the real wage (1986, 16, 17). Joseph Stiglitz(1987, 35, 36) disagrees in principle, although he does not work out a complete nominalefficiency-wage model. Stiglitz is correct; the relative efficiency-wage model presented belowexplains nominal wage stickiness.


17/33

Hoover / Keyness Labor Market 669Keyness argument and their interrelationship. I do not argue that K eyneshad (implicitly) worked out an efficiency-wage model. Rather, the modelI present here is a consistent synthesis of what has appeared to some tobe Keyness incoherent discussion of the labor market.It is perhaps useful to note that this enterprise is not entirely anachro-nistic or Whiggish. Alfred Marshall already uses the term efficiency-wages to refer to work measured inefficiency units(1920,548). Further-more, he had the idea that higher real wages increase efficiency: highlypaid labour is generally more efficient and therefore not dear labour; afact which, though it is more full of hope for the future of the humanrace than any other that is known to us, will be found to exercise a verycomplicating influence on the theory of distribution (1920,5 10).Marshall (1920,529-33 passim) seems largely, though not exclusively,to find the mechanism for increasing efficiency in nutrition and betterphysical conditions for workers, and to find efficiency considerationsmore relevant to the underdeveloped southern clime (cf. Leibenstein1957). Keynes knew his Marshall. Indeed, in a letter to Hawtrey in 1936,Keynes (1973, 35) cites this very passage from Marshalls Principles.Marshall pervades The General Theory,even where he receives no ex-plicit recognition. The model here differs sharply from what K eynesfound in Marshall, in that it is a nominal efficiency-wage model and nota real efficiency-wage model, and in that it is used to address questionsof economic fluctuations and not long-term distribution.Keynes is careful to distinguish between consumption goods and capi-tal or non-wage goods, and he often refers to the variety of capital goods.Keyness analysis ismultisectoral at root. A two-sector model is, there-fore, the minimum appropriate to Keynesian analysis. This is alreadyevident in the analysis of labor supply in section 3 above. Still, to dis-cuss some of the issues involvedin analyzing involuntary unemploymentadequately, I will have to consider a three-sector model. I arbitrarily la-bel two sectors as different capital-goods sectors and the third sectoras the consumption-good or wage-good sector. I assume that relativewages matter only in the capital-goods sectors. One might imagine thatrelative wages were important to workers in every industry; Keynes,however, clearly recognized that there are degrees of bargaining power:some workers are in a better position to protect their relative wages thanothers (see GT,267). To capture this insight in an extreme case, I modelthe consumption-goods sector in the traditional neoclassical manner: thereal wage adjusts to clear the market; labor supply is always equal to


18/33

670 History of Political Economy 27:4 (1995)labor demand. This has the added advantage of malung the model closeto many modern efficiency-wage models in which there areprimary la-bor markets, in which efficiency wages are paid, and secondary labormarkets, in which market-clearing wages are paid (for references, seeAkerlof and Y ellen 1986,3,4). It will thus be easier to discuss moderncriticisms of the Keynesian definition of unemployment in the next sec-tion. Whether it isreally the consumption-goods sector in which workershave the least bargaining power is of no importance in what follows.

4.1 The Capital-Goods SectorsThe two capital-goods sectors will be designated k l and k2. I shall firstpresent a model of sector k 1. Appropriately altering the subscripts givesthe model for sector k2.In common with all efficiency-wage models, assume that effectivelabor is the productof the numberof workers employed (or hours worked)and the efficiency of each worker (or hour). Let efficiency depend directlyupon the perceived relative wage. Thus for k 1

where wkp2 is the perception of the workers in kl of the wage paid ink2. There are assumed to be many firms in each sector, sothat the rulesfor profit-maximization under perfect competition apply wherever theassumption of efficiency wages does not alter them. The short-run pro-duction function for the firm may be written

and is assumed to have the usual neoclassical properties, except that thelabor input isnow measured in efficiency units.Taking prices and thewage ink2 as given, the firm now has two meansof obtaining higher output: to hire more workers or to pay higher nominalwages. The firm chooses both the wage rate and its demand for labor inorder to maximize profits,

The first-order conditions for this problem are:


19/33

Hoover / Keyness Labor Market 671and

or

Optimal choices of wkp andLk for givenPk andwk are easily determined. 3Divide equation (4)by (5)and rearrange to yield the Solow conditionfor this model (see Akerlof and Yellen 1986,3):

which may be solved for wil, the equilibrium value of wk 1 . Substitutingwl l into (4) or (S),he equilibrium demand for labor, Lkq*, may berecovered.

If at the real wage, w k ] P k I , equilibrium labor supply, L,S:, is greaterthan Lkq*,then some workers are unemployed, for Keynes assumed thatthe short side of the market dominates. Workers wish to work at thecurrent wage or even somewhat less, but firms will not hire them, for anygains from additional workers at lower wages are more than offset bylosses from the reduced productivity of the existing workers who suffera drop in their relative wages.

Firms choose an optimal money-wage and will not lowerit in order toemploy redundant workers. On the other hand, if prices rise, holding wagerates constant, efficiency is constant, the real wage falls and firms arewilling to hire more workers, ust asinthe standard neoclassical analysis.Again, we should notice that no irrationality in the sense of Leontief (nomoney illusion) is involved. f all wages (including perceived wagesintheother sectors) and prices change by the same proportion, the demand forlabor and the output of the firm are unaffected. Instead, Keynes providesa reason why firms are reluctant to lower the money-wages they pay inthe face of given money-wages paid by other firms. If all firms feel suchreluctance, wages are sticky downward, despite the fact that a coordinatedwage cut might be good for all firms and for the unemployed.

It is precisely to rule out such coordination that firms are modeled assetting nominal wages and demand for labor conditional on the wage

13.The problem yields sensible solutionsonly if the functionekl (wk l / w , ) as the propershape (see Stiglitz 1987,5): e k l / ( w k l / w & ) must first increase then decrease as ( ~ k l / w ; ~ )increases. An example of such a function, a modification of one given by Akerlof (1982),p. 561, is e =-a +b( wkl / w[ 2) u, herea,b >0and 05 CY 5 I .


20/33

672 Historyof Political Economy 27:4(1995)perceived to prevail in the other sector. This specification aims to captureKeyness view that a firms nominal wages are sticky because workerswould perceive a unilateral cut as eroding their relative position. Clearly,as K eynes points out, inflation would lower real wages without affectingefficiency. A coordinated cut in nominal wages would have preciselythe same effect. One should think of firms as playing a game in whichthey choose their nominal wage simultaneously with other firms whileignorant of the other firms choices. Their perception of the other firmswages is therefore based on the last move. So long as the demand forlabor is less than the supply of labor, no firm will find any advantagein being the first to raise its nominal wage, and every firm will find anefficiency loss in being perceived to have lowered its nominal wagesrelative to other firms. 4Actually, this is true only if the efficiency functions in the two sectorsare related in a special way. To see what is at stake, imagine that eachsector had an efficiency function that achieved peak efficiency only whenthe wage was twice what it was n the other sector. Firms would have toembark on a series of competitive wage increases to attempt to capturerelative advantage. To be fully satisfactory for economic analysis, a modelwould have to include a mechanism through which efficiency wouldalter endogenously until somesortof steady-state were reached. But thepurposes of the current model are expository and considerations of thatdynamic process are beyond the scope of this article. It is clear fromKeyness discussion of relative wages that he imagined some sort ofsteady-state relative wage structure to be prevailing already. To capturethis for comparative static exercises, I impose the condition that theefficiency functions in k 1 and k2be related in such a way thatThis means that the optimal relative wage in kl is precisely the inverseof the optimal relative wage in k2, so that, if each sector chooses itswage based on its perception of the other sectors wage, its perceptionwill be confirmed, and there will be no incentive to embark on a path ofcompetitive adjustment.

14.T he sticky nominal wage provides the samesortof handle for aggregate demand policiesas do sticky nominal wages in models such as those in F ischer 1977 and Taylor 1979. Thenominal-efficiency-wage model differs from these, however, in that i t does not assume anyfixed contract length. The game is played continuously, and the gap between when one firm setsits wage and when this is perceived by workers in the other firm can shrink to an infinitesimalduration, so long as changes in wages are never coordinated.


21/33

Hoover / Keyness Labor Market 673To complete the supply side of the capital-goods sectors, assume that

the supply of labor is derived from a utility-maximization problem as insection3 above and depends on the nominal wage paid and the price ofwage goods, pc :

Lks,=L;,(wl/Pc). (8)Workers are assumed to be completely differentiated by skill, so thatthose employed in sector kl are never employed in k2 nor those in k2in k l . This is a stronger assumption than Keynes in fact makes or thanis really needed, but it captures the flavor of Keyness view of the labormarket and simplifies the concrete model presented in the appendix tothis paper.

Sofar, I have assumed thatL s is greater thanL D. But we must add thecondition that the inequality cannot be reversed.At the point at which thesupply of labor ust equals the demand for labor, the efficiency condition,equation (6),must be replaced by

for all higher levels of demand. Essentially, this says that nominal wageswill rise to clear the sectoral labor market oncefull employment has beenreached.

Within the capital-goods sectors, this model almost agrees with theaccepted explications of Keyness analysis of unemployment: when thereis unemployment, the real wage exceeds the marginal rate of substitutionof consumption for leisure (or, in Keyness own terms, the marginaldisutility of labor). It differs from the accepted explication only in thatthe demand for labor isa particular point, not a schedule, because takingthe price level as given, the fact that the firm chooses the most efficientnominal wage implies that it chooses the real wage as well.

4.2 The Consumption-Goods SectorAssume that efficiency wages are not paid in the consumption goodssector. Therefore, the short-run production function for consumptiongoods is

and has the usual neoclassical properties. Profits are


22/33

674 History of Political Economy 27:4 (1995)A firm chooses L , to maximize ncr,taking p, and w, as given. Thisproblem has the single first-order condition,The demand for labor as a function of the real wage in the capital-goodssector can be derived from inverting this first-order condition:

L p =L,D(wc/p,).Keynes assumed that there was imperfect mobility between occupa-tions (GT, 14). At the extreme, each of our three sectors could haveindependent supplies of labor. It is more useful in addressing the moderndebate over the existence of involuntary unemployment, however, to as-sume that, while there is no mobility between k l and k2,workers in thecapital-goods sectors are the more skilled and the more highly paid, andthat workers in the consumption-goods sector need not be skilled. Qual-ified workers unable to find work in the capital-goods sectors constitute(part of) the supply of labor to the consumption-goods sector. Thus, thesupply of labor in the consumption-goods sector depends on the supply

and demand for labor in eachof the capital-goods sectors as well as uponthe real wage:

Because efficiency wages are not paid in the capital-goods sector, it canbe assumed to clear in the usual manner:L,D =L ; .

If i tcan be shown that the concept of involuntary unemployment makessense even when there is a secondary sector which clears, it follows afortiori that the concept of involuntary unemployment makes sense whenrelative efficiency wages are pervasive.4.3 Aggregate DemandAssuming that the supply of labor exceeds the demand for labor in thecapital-goods sector, total consumption would depend not only uponrelative prices and wages, as it would in the neoclassical model, but alsoupon total employment in the capital-goods sector, which is rationed. 5

15.Consumption multipliers arise because of this rationing ust as they do in Clower's(1965)reinterpretationof the Keynesian consumption function. Were labor demand greater than labor


23/33

Hoover / Keyness Labor Market 675Another way of putting this is that consumption depends on total currentincome-namely, there is a standard Keynesian consumption function:

To close the model, I ignore most of the complications of Keynessanalysisof aggregate demand simply by specifying that, aside from con-sumption, the only other components of aggregate demand are nominalautonomous expenditures (Akl and Akz), which may be thought of asgovernment expenditure on capital goods. A ll goods markets clear, sothatand

and similarly for sector k2. A more sophisticated (and more Keynesian)analysis of aggregate demand would not add anything of significance tothe analysis.4.4Comparative StaticsThe comparative statics of the relative-efficiency-wage model are fairlyeasy to explain. The model is fully laid out in the appendix. Concretefunctional forms are posited, and for a particular set of parameters, anequilibrium solution and the resultsof several comparative static experi-ments are presented. The key results are precisely what Keynes leadsusto expect.Consider the case in which the supply of labor exceeds the demandfor labor in the capital-goods sectors. Additional spending on capitalgoods-Keyness preferred method of combating the slump (see GT,116, 117, 119, 129)--clearly reduces unemployment in the relative-efficiency-wage model. The Keynesian prescription to expand aggregatedemand works as follows. A proportional increase in Akl and Ak2 (sayan increase in government expenditure) increases the demand for capitalgoods, driving up their prices. Because of the relativity game played byfirms, wk I andwk2 do not rise, soefficiency remains constant. The rise insupply, consumption would depend only upon relative prices, but firms would find themselvesrationed in the labor market.


24/33

676 History of Political Economy 27:4 (1995)Pk I and P k2 lowers the real wage from the firms point of view, encourag-ing them to hire redundant workers and to increase output. The additionalwages paid to workers in the capital-goods sectors increase nominal con-sumption demand. Prices (and wages) riseinthe consumption-goods sec-tor tomaintain market clearing. Total unemployment falls, both becauseof the increased demand for labor in the capital-goods sectors and be-causetherisein pc n the faceof fixed nominal wages in the capital-goodssectors lowers the total supply of labor. 6

A continued increase in aggregate demand reduces unemployment,until either k 1 ork2 reaches the point at which the supply of and demandfor labor are ust equal. A t that point, nominal wages in both sectors mustrise. They rise in the sector that first becomes constrained because theshort side of the market dominates. So, as Keynes understood (GT, 3),the limiting case of his model was a classical model. Nominal wages risein the other sector because firms find it profit maximizing to preservethe existing relativities. In The General Theory,Keynes argues that histwo definitions of unemployment-one based on rationing, the other onthe elasticity of output to aggregate demand-were in fact consistent(cf. Keynes 1973, 71, 86). In a letter to John Hicks in 1936, Keynes(1973,74-75) clarifies the point, noting that in the short run, when laboris specialized by skill or location, inelasticity of output will be reachedbefore each separate specialized market clears (1973, 74-75). This is anatural property of the model developed here.The only prominent feature of the Keynesian account that is not foundin the model as it is set out here is K eyness recognition that nominalwages are not strictly rigid, but in fact do fall in the slump, though not insuch a way as to ensure market clearing. The model could be modified insuch a way that efficiency becomes a function of the state of the economy.As Keynes writes, L abor is not more truculent in the depression thanin the boom-far from it (GT,9). With some modification, the modelwould allow nominal wages to rise somewhat in the boom, even beforethe capital-goods sectors reach full employment, and fall somewhat inthe slump.

16.After TheGeneral Theorywas published, K eynes 1 973,26) wrote that a better definitionof unemployment might in fact be that unemployment exists when holding the marginal propen-sity to consume constant, an increase in investment increases consumption. The comparativestatic experiment here corresponds toa test for unemployment under that definition, and showsthat, as Keynes believed, his various definitions were consistent.


25/33

Hoover / Keyness Labor Market 6775. The Concept of Involuntary Unemployment5.1 A New Classical ObjectionAgainst Keynes is a central attitude in the new classical worldview.While it would be anachronistic to build new arguments for Keynes torebut the anti-K eynesian positions of the new classicals, it is instructiveto see how well the old arguments hold their own. Continued relevanceis a fundamental rationale for history of economic thought.The new classical macroeconomics seeks market-clearing microfoun-dations for macroeconomics. 7 In general, new classical economists dis-sent from K eynesian macroeconomics as relying on various sorts ofirrationality-contrary to the spirit of neoclassical microeconomics. Therelative-efficiency-wage model shows that, at least in the labor market,such irrationality need not be associated with K eynes.More particularly, however, the new classicals challenge the sound-ness of Keyness distinction between voluntary and involuntary unem-ployment. This is most clearly stated by Robert Lucas:

Nor is there any reason why one would want to draw this distinction.Certainly the more one thinks about the decision problem facing in-dividual workers and firms the less sense this distinction makes. Theworker who loses a good job in prosperous times does not volunteerto be in this situation: he has suffered a capital loss. Similarly, thefirm which loses an experienced employee in depressed times suffersan undesired capital loss. Nevertheless the unemployed worker at anytime can always find somejob at once, and a firm can always fill a va-cancy instantaneously. That neither typically does sobychoice is notdifficult to understand given the quality of the jobs and the employeeswhich are easiest to find. Thus there is an involuntary element in allunemployment, in the sense that no one chooses bad luck over good;there is also a voluntary element in all unemployment, in the sensethat however miserable ones current work options, one can alwayschoose to accept them. (1978,242)

Lucas goes on to accuse Keynes of indulging in mere wordplay in hisdefinition of unemployment. The phrase the existing money wage,he argues, is completely ambiguous. Unless it is defined as the pricesomeone else is willing to pay him for his labor. . .,what is it? (Lucas17.See Hoover 1988 for a complete analysisof new classical macroeconomics.


26/33

678 Historyof Political Economy 27:4 (1995)1978, 242). And if it is defined in this way, there is no such thing asinvoluntary unemployment; for there is always some wage available toany worker; and, if they do not accept it, they are voluntarily unemployed.Lucas makes the same point in a more homely way in K lamers Con-versations with the New Classical Economists . . . (1984):

K L A M E R : y taxi driver here is driving a taxi, even though he is anaccountant, because he can t ind a job . . .LUCAS: would describe him as a taxi driver [laughing], if what he isdoing is driving a taxi.K L A M E R :ut a frustrated taxi driver:LUCAS:Well, we draw these things out of urns, and sometimes we getgood draws, sometimes we get bad draws. (40)

5.2 A Modern ResponseThe new classicals are not known to favor efficiency-wage models. Still,even in most such models, Lucas would have to say that there is no in-voluntary unemployment. In the relative-efficiency-wage model of sec-tion4,the secondary sector clears and all workers are potentially workersin the secondary sector (taxi drivers); therefore, there is no involuntaryunemployment. This is much like Thomas Hobbess point that fear andliberty are consistent ( I 651, chap. 21). When one chooses after ratio-nally weighing the consequences, however unpalatable, ones action isvoluntary (cf. Lindbeck and Snower 1987,7; de Vroey 1991, 19, 20).In answering a point about an efficiency-wage model, Shapiro andStiglitz attempt to deflect this same sort of criticism:

At one level, the issue here is just a matter of semantics. It is littledifferent from the old story that so long as there is a competitive la-bor market anywhere in the economy (grape picking in California)all unemployment must be voluntary, since any individual could havemoved to California (purchased a job). Tous the fact that during theGreat Depression20or25percent of the labor force in Chicago, work-ers who were once gainfully employed, were sitting at home, willingto work, at the going wage in Chicago, suggests a massive marketfailure, regardless of whether one says that, because of their decisionnot to migrate to California, they were voluntarily unemployed. ( 1987,1215)


27/33


This line of argument-though it strikes a responsive chord-is not en-tirely effective against a point such as Lucass. It does not answer directlyhis charge about the ambiguity of a phrase such as the going wage. Itclaims a point is merely semantic, as if that meant totally insubstantial.At heart involuntary unemployment is here defined to be that a worker isinvoluntarily unemployed when he is out of work and the wage exceedsthe marginal disutility of his labor. And, when one asks, as Lucas does,what sort of labor? we must answer, the sort the worker most wants todo if he is qualified to do it. But, as Lucas senses, this may be difficultto pin down.

And Keynes has a stronger response at his disposal.

5.3Keynes Anticipated LucasKeyness definition of involuntary unemployment is not that the wage ex-ceeds the marginal disutility of labor of those out of work. This is simplya deduction from his primary definition-namely, a small rise in pricesincreases employment in aggregate or, equivalently, the elasticity of em-ployment to aggregate demand is not zero (cf. Leijonhufvud 1968,94).

Keyness definition of involuntary unemployment, especially the firststatement of it as a test of what happens when the price of wage goodsrises, strikes many as extremely convoluted, andithas been the source ofmuch confusion. However, once we interpret Keyness labor market as arelative-efficiency-wage model, the value of such a definition becomesclear. By proposing a test of what happens to total employment, Keynesdirects our attention away from the individual unemployed worker to-ward the work force and the economy as a whole: menare involuntarilyunemployed (GT, 15; original italics, emphasis added) not a man isinvoluntari ly unemployed. Of course, if men are unemployed, someparticular men must be unemployed as well; and Keynes points to ex-actly those people whom common sense suggests-workers who wouldwork in their chosen profession at the going wage, but who are not of-fered work. If an earlier Lucas had challenged Keynes over the ambiguityof the going wage, Keynes could merely have answered, perform thetest: consider a single market and ask, if the price of wage goods rises,would more people be employed? Or consider the economy as a wholeand ask, if aggregate demand rises, would more people be employed?Those potential accountants who refuse to work as taxi drivers are in-voluntarily unemployed if, in a higher state of aggregate demand they


28/33

680 History of Political Economy 27:4 (1995)would be employed as accountants; even those who work as taxi driversare involuntarily unemployed as accountants if, in a higher state of aggre-gate demand, they were offered desks and ledgers instead of taxis. Therelative-efficiency-wage model encapsulates both phenomena, which isperhaps the strongest reason for taking it to be an adequate summary ofKeyness fundamental insights.The relative-efficiency-wage model also suggests why the proposal ofthose Keynesian critics who would surgically remove chapter 2 fromThe General Theory is unnecessarily drastic. Concern for relative wagescan force the economy into a stable configuration below full employ-ment. If there is a sector such as the consumption-goods sector in themodel above, or like taxi-driving in K lamers exchange with Lucas, inwhich the wage clears the market, then the economy will appear to befully employed. There is malemployment in such an economy as someof the employed could be better employed. Such malemployment de-pends, however, precisely as Keynes suggests, on the violation of thesecond classical postulate (that is, on labor rationing), and his appar-ently convoluted definition of unemployment provides the acid test forits existence. Chapter 2, therefore, provides the firm basis for the elastic-ity of employment and output to aggregate demand, which is the aspectof underemployment these critics prefer to stress. Chapter 2 is neitherredundant nor incoherent; it is indispensable.Lucas and his allies revive classical modes of economic analysis. Be-cause he knew his classical opponents intimately, Keynes tailored hisdefinitions to answer their criticisms and, in sodoing, fully anticipatednew classical objections as well. In the quotation above, Lucas arguesthat unemployment is voluntary albeit unfortunate. Keynes understoodthis point of view: A classical economist may sympathize with labour inrefusing to accept a cut in its money wage, . . .;but scientific integrityforces him to declare that this refusal is, nevertheless, at the bottom ofthe trouble (GT,16). But K eynes continued, The classical theorists re-semble Euclidean geometers in a non-Euclidean world who, discoveringthat in experience straight lines apparently parallel often meet, rebukethe lines for not keeping straight-as the only remedy for the unfortunatecollisions which are occurring. Y et, in truth, there is no remedy exceptto throw over the axiom of parallels and to work out a non-Euclideangeometry (GT, 16).

A metaphor perhaps more apt than Keyness own is that the clas-sical economists are like the Ptolemaic astronomers, who can fit their


29/33

Hoover / Keynes's Labor Market 681models to most any phenomenon, providing, of course, they can in-vent enough clever epicycles. Of course, Keynes is then our Copernicus,leaving plenty of room for latter-day K eplers and Newtons to correct anddevelop his insights. This seems to identify the new classicals with theInquisition, cleaving to theoldscience, while denouncing the new, in theface of overwhelming evidence to the contrary.Appendix: The Relative-E fficiency-Wage M odelWhat follows is an illustration of the model described in the text withconcrete functional forms replacing the general mathematical forms pre-sented there.

production functionfirst-order conditionfor profit maximizationlabor supplyefficiency function

Solow condition

labor market clearinggoods market clearing

production functionfirst-order conditionfor profit maximizationlabor supplyefficiency function

Solow condition


30/33

682 History of Political Economy 27:4 ( I 995)L f2=L ; market clearing

goods market clearing

production functionfirst-order conditionfor profit maximizationlabor supplylabor market clearinggoods market clearing

(A.14)

(A.15)(A.16)(A. 17)(A.18)(A.19)

The labor-supply function for sector c (A. 17) is derivative of the linearlabor-supply functionsin sectorsk 1andk2.It is based on the assumptionthat firmsin k l andk2 choose the workers they actually hire uniformlyfrom those whose reservation wage is at or below the offer wage: thatis, they cannot discriminate on the basis of the reservation wage. Uis the conventionally calculated rate of unemployment: in this modelU = 1 - L : +L L +' L : ) /( L i , +L i2+L:.).Thebi,'s are chosen toensure that equation (7) (in the main text) is fulfilled withwkI /wk2=1.

Consider the four comparative static experiments presented in table 1.The parameter settings illustrate the existence of an equilibrium and thebehavior of the model; they confirm the comparative static results dis-cussed in the text. Column I is taken to be the benchmark solution, andshows that the model is not vacuous. Column IT shows that when thedemand parameters, dll and dZ1 ,and all prices and wages (includingthe perceived wages) are raised proportionally, there are only nominal,not real, effects. This confirms the homogeneity of the model of degreeone, demonstrating that Leontief's objection does not apply. Column I11shows a coordinated symmetrical wage cut (including perceived wages)with no change in demand. This demonstrates that, were such coordi-nation is possible, unemployment could be lowered. Column IV showsa symmetrical increase in the demand for capital goods without coor-dination of response (that is, initial perceived wages do not change).Actual wages do not change either and demand is increased enough sothat unemployment once again falls-this time to zero.


31/33

Hoover / Keynes's Labor Market 683

Table1 Comparative StaticsParameter Settings

CaseParametera1 Ia2a3a12a22a32blb2 Ib12b22613CI Ic2 Ic3 Idl Idl 2g31g32

Wkq4 2

Lk"lLks

Variablewk

Y kP k Iwk2Lk"2Lk"2Yk2P k2WCL:L:YePcU

I2oooo2oooo2.00000.50000.50000.25001 .oooo1 .oooo4.00004.00000.75002oooo2oooo2.0000I .oooo1 .oooo0.00000.9000

4.00004.0000

I4.00000.25000.49001.73211.15474.00000.25000.4900I .732I .27026.09221.47731.47732.2049

16.32700.1953

I 12.00002.00002.00000.50000.50000.2500I .ooooI .oooo4.00004.00000.75002.00002oooo2oooo1.10001.10000.00000.9000

Initial C4.40004.4000

I 14.40000.25000.49001.732I .27024.40000.25000.49001.732I .27026.70 14I .4773I .47732.2049

17.95970. 953

I112.00002.00002.00000.50000.50000.25001 .ooooI .oooo4.00004.00000.75002.00002.00002oooo1 .ooooI .oooo0.00000.9000

hditions3.50003.5000

1113.50000.28570.4 1701.85161.0803.50000.28570.4 I701.8516I .080

16.78701.335I .3352. I4998.37270.1218

IV2.00002.00002.00000.50000.50000.2500I .ooooI .oooo4.00004.00000.75002.00002.00002.00001.10001.10000.00000.9000

4.00004.0000

IV4.00000.33330.33332.00001.33344.00000.33330.33332.0000I .3334

12.00160.99990.9999I .9999

24.00 140.0000


32/33

684 History of Political Economy 27:4 ( 1 995)ReferencesAkerlof, GeorgeA. 1982. Labor Contracts as Partial Gift Exchange.Quarterly J our-Akerlof, GeorgeA,, and Janet L. Yellen. 1986.EfJ iciency Wage Models of the LaborBarro, RobertJ ., and Herschel I . Grossman. 1971. A General Disequilibrium Model

. 1976.Employment, Output and Inflation. Cambridge: Cambridge Univer-Bateman, Bradley W. 1987. Keyness Changing Conception of Probability. Eco-Blaug, Mark. 1985.Economic Theory in Retrospect. 4th ed. Cambridge: CambridgeCarabelli, Anna M. 1988.On Keynes Method. London: Macmillan.Chakrabarti, Santi K. 1979.The Two-Sector General Theory Model. Delhi: Macmil-Chick, Victoria. 1987.Macroeconomics after Keynes. Deddington, Oxford: PhilipClower, Robert W. [I9651 1984. The Keynesian Counter-revolution:A Theoretical

nal of Economics 97.4543-69.Market. Cambridge: Cambridge University Press.of Income and Employment.American Economic Review 61. :82-93.sity Press.

nomics and Philosophy 3.1 :97-119.Un vers ty Press.

Ian.Allan.Appraisal. In Clower 1984b.

. 1984a. Money and Markets. In Clower 1984b.

. 1984b.Money and Markets: Essays by Robert M Clower. Edited by Don-ald A. Walker. Cambridge: Cambridge University Press.

Darity, William A., Jr., and Bobbie L. Horn. 1987a. Involuntary UnemploymentReconsidered.Southern Economic J ournal 49.3:7 17-33..1987b. nvoluntary Unemployment Independent of the Labor Market. our-nal of Post Keynesian Economics 10.2:216-24.

De Vroey, Michel. I99 . Labour Rationing, Involuntary Unemployment, Underem-ployment, and Malemployment:A Conceptual Elucidation. Working Paper No.91 19, Institut de Recherches Economiques et Sociales, Department des SciencesEconomiques, Universite Catholique de Louvain, Belgium.

Fender, John. 1981. Understanding Keynes: An Analysis of the General Theory.Brighton: Wheatsheaf.

Fischer, Stanley. 1977. Long-term Contracts, Rational Expectations, and the OptimalMoney Supply Rule. J ournal of Political Economy 85.1 9 1-206.

Friedman, Milton. 1974.A Theoretical FrameworkforMonetary Analysis. n MiltonFriedmans Monetary F ramework. Editedby RobertJ . Gordon. Chicago: Univer-sityof Chicago Press.

Galbraith, ohn Kenneth. 1987.Economics in Perspective: A Critical History.Boston:Houghton Mifflin.Hicks, John R. [1937] 1967. Mr. Keynes and the Classics. In Critical Essays inMonetary Theory. Oxford: Clarendon.Hobbes, Thomas. 1651. Leviathan, or the Matter; Forme and Power of a Common-


33/33

Hoover / Keyness Labor Market 685wealth Ecclesiastical and Civil.Oxford: Black well.Money. London: Macmillan.

Hoover, K evin D. 1988. The New Classical Macroeconomics: A Sceptical Inquiry.Keynes, John Maynard. 1936. The General Theory of Employment, Interest and

. 1937.The General Theory of Employment.Quarterly J ournal of Economics

. [I 9311 1973.The Economic Consequences of Mr. Churchill. In Keynes1973.. 1973. The General Theory and After: Part I / Defence and Development.Vol. 14 of The Collected Writingsof J ohn Maynard Keynes. London: Macmil lanand St. Martins Press.

K lamer, Arjo. 1984. The New Classical Macroeconomics: Conversations with theNew Classical Economists and Their Opponents. Brighton: Wheatsheaf Books.Lawlor, M ichael S., William A. Darity, J r., and Bobbie L.Horn. 1987.Was Keynes aChapter Two K eynesian? J ournal of Post Keynesian Economics 9.4516-28.Leibenstein, Harvey. 1957. Economic Backwardness and Economic Growth. NewY ork: Wiley.Leijonhufvud, Axel. 1968.On Keynesian Economics and the Economics of Keynes:A Study in Monetary Theory. New Y ork: Oxford University Press.Leontief, Wassily. 1936.The Fundamental Assumption of M r. Keyness MonetaryTheory of Unemployment. Quarterly J ournal of Economics 51 . 1 :192-97.L indbeck, Assar, and Dennis Snower. 1987.The Insider-Outsider Theoryof Employ-ment and Unemployment. Cambridge: M IT Press.Lucas, Robert E., Jr. 1978. Unemployment Policy. In Studies in Business-CycleTheory.Oxford: Blackwell.Marshall, A lfred. 1920.Principles of Economics. 8th ed. London: M acmillan.Patinkin, Don. [19561 1965.Money, Interest, and Prices. 2d ed. Harper & Row.

52.11209-23.

.1982.Anticipationsof the General Theory?Chicago: University of ChicagoPress.

. 1987. J ohn Maynard K eynes. In The New Palgrave Dictionary of Eco-nomics. Edited by J . Eatwell , M . Milgate, and P. Newman. London: Macmil lan.Shapiro, Carl, and Joseph E. Stiglitz. 1987. Can Unemployment Be Involuntary:

Reply. American Economic Review 75.5: 215- I7Smithin, John Nicholas. 1987. The Definition of Involuntary U nemployment inKeynes General Theory: A Note. HOPE 17.2:219-22,Stiglitz, J oseph E. 1987.The Causes and Consequencesof the Dependence of Qualityon Price. J ournal of Economic Literature25.1:148.Taylor, J ohn B. 1979.Staggered Wage Setting ina MacroModel.American EconomicReview 69.2:108-1 3.Tobin, J ames. 1992.An Old Keynesian Counterattacks. Eastern Economic J ournalI 8.4:387400.Vercelli, Alessandro. 1991. Methodological Foundations of Macroeconomics:

keynes labor market (hoover)

Documents