Increasing Returns and Unemployment Theory

Increasingreturns and unemploymenttheory

Increasingreturns and the foundationsof unemploymenttheory: an explanation MARTIN L. WEITZMAN

I would like to express my gratitudeto the contributorsfor their stimu- lating comments and to the editor for kindly providing me with an opportunityto reply. There is much in the comments with which I agree, and from which I have learned. I certainlywish I had expressed myself more clearly on certain points in the article. I hope that my remarks here will help to clarify some of these points. At first I thoughtI would replyto each commentby addressing,more or less systematically and exhaustively,the specific issues they were raising. Now I am not so sure that is the best approach.The issues are too disparate, and some of them, if I may be excused for expressing myself so directly,seem to me beside the point. Instead,I thinkit would be more fruitfulif I tried to restatemy general thesis, in words, taking into accountas best I can the criticisms that seem to me most relevant. My immediatepoint of departureis the "Keynesian message." By that term I have in mind not anotherrendition of what Keynes meant in The General Theory. (I am not particularlyqualified, in any event, to give such a rendition.) I have in mind more the spirit of that which constitutes what might be called the Keynesian outlook or economic philosophy. Now this "spirit" is hardlyan objective entity, and need- less to say it is my interpretationof it which is here recorded.Neverthe- less, and allowing for normalhuman differences in perception,I would hope there might be rough agreementon those features I am about to highlight. All this is by way of leading up to where I think increasing The authoris Professor of Economics at the MassachusettsInstitute of Technology.

Journal of Post KeynesianEconomics/Spring 1985, Vol. 7, No. 3 403 404 JOURNALOF POST KEYNESIANECONOMICS

returns (and imperfect competition) fits into the picture-why I think this has been an extremelyimportant (if hidden)element in the Keynes- ian story up to now and why it is likely to be even more importantin the future.

The Keynesian message The starting point of a Keynesian outlook is the observationthat the advanced capitalist countries frequently suffer from "involuntary unemployment."'While most attentionis naturallyfocused on the labor market, in such a condition labor and capital are unemployed. A strikingthing aboutthis phenomenonis thatthere does not appearto be any deep-seated rationale or justification for it. When there is an oil shortage energy prices shoot up, but that is an understandable,if sometimes annoying, aspect of the capitalistsystem, and it does serve an importantpurpose. When there is involuntaryunemployment, it is as if the system itself is malfunctioningor operatingin a self-destruc- tive mode. The welfare losses associated with this form of "market failure," if that is the best term for it, are typically astronomical compared with other forms of market failure. I mention all of this explicitly because there has been some tendency,which is now becom- ing more pronouncedagain, to deny the problem. So right from the beginningthere tends to be a difference betweenKeynesian and classical perceptionsof the importanceof the problemof involuntaryunemployment. I hope to show, among other things, that the issue of returns to scale is involved even at this initial perceptuallevel. As we all know, there has been no small effort expendedon finding the "source" of such a massive marketfailure. Literallythousands of articles have been written on the microfoundationsof macroeconomics. Naturallyenough, there is no unanimityhere. But there is plenty of repetition. All of this is as it shouldbe, because the issue is an extreme- ly important one. Some of the main themes in this "failure of coordination" literaturewhich tie into The General Theorymay perhaps be summarizedas follows. The Keynesian view looks out over a capitalist world that is full of uncertainty.In such a world, expectationsplay an importantdual role as both a manifestationof uncertaintyand a cause of it. An economic agent's expectationsin thatkind of environmentare arbitraryto a large degree because they can be based on almost anything, including self- fulfilling expectations of the behavior and expectations of other INCREASINGRETURNS AND UNEMPLOYMENT:EXPLANATION 405

agents. "Being based on so flimsy a foundation," as Keynes put it, such expectations are "subject to sudden and violent changes." The system itself is poorly equippedto handle uncertaintyabout the future because of the conspicuous absence, undoubtedlyfor good reason, of most Arrow-Debreutype marketsfor future and contingentdeliveries of commodities. Money, or near-moneygovernment debt, then serves as a store of value linking the present with an uncertainfuture. This kind of money, unlike its classical counterpart,has importantnon- neutralityproperties which can be exploitedby policy makersfor better or for worse. The fact that most prices, including wages, are denominatedin the monetaryunit of account also has importantimplications in a genuine monetary economy. Wages tend to be rigid or sticky. An important strandof the post-Keynesianliterature has sought to justify this sticki- ness, or at least to remove from it the dreadedtaint of ad hoc money illusion. This strandhas led to severalsub-theories-implicit contracts, efficiency wages, interdependentutilities, and others too numerousto mention-with the commontheme of trying to prove, in some broader- than-usualsense, the "rationality" of sticky wages. I will end this tiresome litany here, even thoughit could be extended in yet furtherbyways and directions. The bottom line to most believers in the Keynesianmessage is that there are more thanenough reasonsto justify governmentintervention to recoordinateprivate failures of coordination.And classical macroeconomists,on the other hand, are not likely to be persuadedby what they are proneto see as one set of ad hoc reasons replacing another.What, then, you may ask, is the reason for yet another story of coordination failure based upon, of all things, economies of scale?

The macroeconomicrole of increasing returns Suppose,just for the sake of argument,we lived in a worldwhere there was strict constant returnsto scale-down to the level of a person or even a grain of sand, or even, if that were necessary, an atom. The standardreasons for macroeconomic failure listed above, and many

'A remarkableseries of importantrecent papers can be interpretedas supportingthe contention that the concept of "rational expectations" in a macroeconomiccontext can be a very tenuous construct. See, e.g., Farmerand Woodford(1984) and the nu- merous references cited therein. 406 JOURNALOF POST KEYNESIANECONOMICS others, would continue to hold. Money might be non-neutral,there could be a genuine role for governmentas a Pareto-improverof social welfare, etc., etc. But, and here is the crucial distinction, there can be no involuntaryunemployment with strict constantreturns to scale in all aspects of technology. Note carefully the claim: some form of increasing returns is a necessary, but by no means sufficient, condition for genuine involuntaryunemployment. Under constantreturns, the macroeconomic inefficiencies would show themselves in the form of "wrong" labor-leisurechoices (or, more generally,wrong substitution choices among variousfactors and commodities). An economy of blue- berry pickers, mushroom gatherers, and clam diggers cannot exhibit involuntaryunemployment no matterwhat else is presentor absent. It can show fluctuations, inefficiencies, poverty, even starvation,but it cannot show involuntary unemployment. Any "involuntarily unemployed" resourcewould merely form itself into a mini-firm, hire (with non-increasingreturns to scale in borrowing) a few grains of cooperating input, and sell its mini-outputon competitive markets. Balancedexpansion would take care of the rest. Involuntaryunemploy- ment is logically impossiblein a strictconstant returns to scale worldof one-person firms. Now it seems to me that the really damagingmacroeconomic inefficiencies of advanced capitalist countries are caused by involuntary unemployment,not by the wrong labor-leisure choice. It seems to me that Keynes would never have written his book if he had lived in a constant-returns-to-scaleworld because there would never have been a Great Depression. It seems to me that if we could magically turn our economy into a constantreturns systems-if the automobileworker laid off from a 10,000-man plant could produce in his home workshop 1/10,000th of what thatplant produces (by using in his home workshop 1/10,000th of its capital)-we would have eliminatedthe lion's shareof macroeconomiclosses due to coordinationfailures. Othercoordination inefficiencies admittedlymight remain, but my casual empiricismtells me they would be orders of magnitude smaller in terms of welfare losses. As JamesTobin has remarkedin a differentcontext, it takes a lot of Harburgertriangles to fill up an Okun gap. So I see increasing returnsand imperfectcompetition as not just anotherminor detail, but as crucial aspects of the Keynesian story. That story simply cannot be told at all credibly or completely without something like increasing returnsblocking unemployedlaborers from workingon their own or in small groups. It was simply to focus as sharply as possible on the INCREASINGRETURNS AND UNEMPLOYMENT:EXPLANATION 407 underlying"real" role of increasingreturns and imperfect competition that I attempted(perhaps unsuccessfully) to trim away as much as possibleof all else fromthe model.(Certainly there is a crucialplace, in anycomplete story, for money,sticky wage contours, expectations, andso forth-my basicpoint being that increasing returns constitutes a necessary,but by no meanssufficient, condition for the existenceof involuntaryunemployment.) Furthermore, as I triedto show in the paper,the quantity-adjustmentmechanisms which play such an impor- tantrole in the operationalpart of Keynesiantheory can be grounded muchmore solidly in imperfectcompetition than in perfectcompeti- tion, wherethey reallydo representan artificialintrusion. In mostreasonable models of an economywith non-trivial increasing returnsto scale,there is goingto be a theoremshowing that higher levels of equilibriumemployment are associatedwith higher real wages. Thisaspect comes out quiteclearly in the modelunder discus- sion andI believeit obtainsunder fairly general circumstances.2 The existenceof economiesof scalewill generallymean that higher levels of long-runeconomic activity go togetherwith higherreal pay. Now pro-cyclicalreal wages is a very un-classicalfeature which not only correspondsempirically to what we typicallyobserve in the real world,3but makes it extremelydifficult theoretically to acceptthe idea thatthe economic system can automatically, and relatively easily, adjust itselftoward full employment.After all, the classicalargument is that unemploymentwill be eliminatedby downwardpressure on wages. Arguingwhere burdens of prooflie is alwaystricky, but it seemsto me thatthe burdenof proofhere rests squarelyon whoeverwould assert that downwardpressed (money) wages spontaneouslycause the in- creasedreal wages that accompanyhigher employment. For that to happen,prices must decline even faster than wages. It is possible,but some good storieshave to be told. Hereis yet anotherindication that economiesof scaleform a naturalbackdrop for Keynesianmacroeco- nomics.4 As I perhapsdid not sufficientlystress in my paperbut have been at painsto emphasizehere, it is not increasingreturns alone that causes involuntaryunemployment. The astute commentsof de Meza and

2This issue is further explored in the working papers of R. M. Solow (1984) and J. E. Meade (1984). 3See the recent work of Bils (forthcoming)and the references cited there. 4This has long been the contention of N. Kaldor. See Kaldor (1984) and the references there to his earlier work. 408 JOURNALOF POST KEYNESIANECONOMICS

Perlman, and Davidson (without the polemics), note that ingredients like money, expectationsof other firms' responses, reasonablespecifi- cations of tastes, a sticky wage contour of equal pay for equal work, and so forth are also needed to give a credible account of involuntary unemployment.Yet, I would maintain,it is importantnot to lose sight of the major truth that large scale division of labor makes for an economic environmentin which, as contrastedwith constantreturns, it really is quite fundamentallydifficult to tell reasonable stories about how a market economy naturallyadjusts to create full employment.

Beyondpost-Keynesian economics There is at least one additional reason why increasing returns and imperfect competition are essential ingredients in the Keynesian approach. This reasonI believe to be the most importantof all. It has to do with where the subject of macroeconomics is going, or ought to be going. The currentfailure of economic policy is not so much a failureof the Keynesiancase for governmentintervention per se, as it is a failure to find new ways of permittinga high level of employmentto coexist with low inflation. The general considerations put forth by Keynes and others for justifying government interventionto improve the macroeconomic environmentdo not indicate the best form of government policy. Keynes naturallythought in terms of standardfiscal and monetary instrumentsto manage aggregatedemand. But there is no reason why such sledgehammertools should remain equally appropriatefor dealing with the macroeconomicproblems of our own day. It is my contentionthat Keynesianeconomists have not been nearly imaginative enough in devising new mechanisms for dealing with stagflation. This lack of imaginationgoes completelyagainst the spirit of Keynes himself. As E. A. G. Robinson has expressed the thought recently,there has been a "failure of our generationto analyse clearly the essential preconditionsof reconciling a high level of employment with avoidanceof inflation, to identify the institutionalchanges necessary to achieve this, and then to establish first a consensus and then action regarding the making of the institutionalchanges." What is most desperately needed, in my opinion, is an improved frameworkof microeconomicincentives to automaticallyinduce better output, employment, and pricing decisions at the level of the firm, therebygreatly reducing the need to rely so exclusivelyon discretionary INCREASINGRETURNS AND UNEMPLOYMENT:EXPLANATION 409 macropolicy. In this search for an improvedeconomic mechanism, the ideasof increasingreturns and imperfect competition within a Keynes- ian macrostructurewill, I am convinced,play a centralrole.5

5My own opinions on this matter have been developed in my book The Share Economy.

REFERENCES

Bils, Mark. "Real Wages Over the Business Cycle." Forthcomingin Journal of Political Economy. Farmer,Roger, and Woodford,Michael. "Self-Fulfilling Prophesies and the Business Cycle." CARESS WorkingPaper #84-12, University of Pennsylvania, 1984. Kaldor, Nicholas. "Keynesian Economics After Fifty Years." In Keynes and the Modern World.Ed. by Trevithickand Worswick. CambridgeUniversity Press, 1984. Meade, James E. "The MacroeconomicImplications of Monopolistic Competition." Unpublishedpaper, 1984. Solow, Robert M. "Monopolistic Competitionand the Multiplier." Unpublishedpaper, 1984. Weitzman,Martin L. "Increasing Returns and the Foundationof UnemploymentTheory." Economic Journal, December 1982, 92, 787-804. . The Share Economy. HarvardUniversity Press, 1984.