閤 Second Thoughts on Keynesian Economics OR。 Robert J. barro Annual Meeting of the American Economic Association. May, 1979),pp 54-59 vinety-First The American Economic Review, Vol 69, No. 2, Papers and Proceedings of the Stable url: tp/ links. istor org/sici?sic=0002-8282%628197905%2969%3A2%63C54%63 ASTOKE%3F20C0%3B2-6 The American Economic Review 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 http:/lwww.istororg/about/terms.htmlJstOr'sTermsandConditionsofUseprovidesinpartthatunlessyouhaveobtained 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 JSTOR is an independent not-for-profit organization dedicated to and preserving a digital archive of scholarly journals. For more information regarding JSTOR, please contact support(@jstor.org Thu mar1522:36:082007
Second Thoughts on Keynesian Economics Robert J. Barro The American Economic Review, Vol. 69, No. 2, Papers and Proceedings of the Ninety-First Annual Meeting of the American Economic Association. (May, 1979), pp. 54-59. Stable URL: http://links.jstor.org/sici?sici=0002-8282%28197905%2969%3A2%3C54%3ASTOKE%3E2.0.CO%3B2-6 The American Economic Review 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 http://www.jstor.org/about/terms.html. 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 http://www.jstor.org/journals/aea.html. 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. JSTOR is an independent not-for-profit organization dedicated to and preserving a digital archive of scholarly journals. For more information regarding JSTOR, please contact support@jstor.org. http://www.jstor.org Thu Mar 15 22:36:08 2007
MACROECONOMICS: AN APPRAISAL OF THE NON-MARKET-CLEARING PARADIGM Second Thoughts on Keynesian Economics By robert J, barro* My view in the early 1970's of Keynesian, The contracting approach may rationalize on-market-clearing-type models was that some departures of real wages from the the soundness of their theoretical structure marginal product of labor and or the margi hinged on an as yet absent theory of the nal value of worker time, but it does not imply stickiness of wages or prices. The application that levels of employment would differ signif- of contracting theory to macro analysis icantly from the(efficient)values that would seemed promising in this respect. The have been attained under flexible wages presence of employee risk aversion or of Rather than rationalizing the non-market transaction costs associated with marl clearing model as a useful "as if"approach arrangements-which could include elements contracting analysis suggests that-despite of capital that were specific to employment or the possible existence of"sticky"wages-the other aspects of production and exchange continuous market-clearing model may pro- seemed to motivate some long-term, implicit vide a satisfactory framework for the analysis or explicit agreements about wages or prices. of employment and output. Notably, the In particular, a sluggish adjustment of wages approach suggests that such market features to current economic conditions could be as sticky wages or the apparent non-price rationalized by this approach quantity rationing associated Further consideration of the contracting would be of secondary interest in analyses of model suggests that its rationale for sticky business cycles. Since the prevailing wage wages and prices--as far as it goes--does not need not represent the marginal product of xplain the key features of Keynesian analysis labor, the presence of"excess labor supply"at ith regard to the determination of employ- this wage need not signal involuntary unem ment and output. For example, long-term ployment in any economic sense labor agreements do not imply a failure of The conclusions derived from the contract employment to increase when all parties to ing model can be generalized by observing the agreements perceive that they could be that the key assumption of Keynesian analysis made better off by such a change. The so- is the inefficiency of some aspects of private called involuntary unemployment of Keynes- sector activity in comparison to corresponding everyone perceives accurately that the margi- central feature is, of course, the underlying nal product of labor exceeds the marginal basis for the policy activism that typifies value that potential workers place on their Keynesian thinking. In some simpledisequi time -is not compatible with efficient labor librium"macro models, relative private sector agreements. Even in contracts that specify, ex inefficiency is represented by sticky wages or ante, the value of nominal wages over some prices, in contrast to the flexibility of such interval of time, it would be mutually advan- government policy instruments as the money tageous for workers and firms to determine supply, taxes, or expenditures. Technical limi levels of employment in an efficient manner. tations of the private market in the coordina tion of production and exchange--as reflected *University of Rochester. I have benefited from c ge-price stickiness and the associated ments by Herschel Grossman, Bob Hall, and Ben Mccal determination of employment and output ional Science Foundation has supported through a non-price rationing process-are this research remedied through tI ordinate
VOL 69 NO. 2 AN APPRAISAL OF THE NON-MARKET-CLEARING PARADIGM skill of the government--as embodied in these which would imply "insufficient"output on models in the judicious use of monetary and average. This observation would seem impor fiscal policy or in the appropriate direct tant for the design of tax and welfare regulation of prices and quantities programs. However, these types of external Presumably, sticky wages or prices are not effects produced by government intervention intended to be taken literally as the source of (which may or may not be warranted on other problem must reflect some deeper economic the business cycle or for the usual forms of elements, such as imperfect information macro-stabilization policies it the present or future, factor mobility A typical feature of macro analysis is that ts, or some types of significant transaction government intervention into the economy ts.Although of these elements are recommended without bothering to describe probably important in business cycle anal- the supposed externality or private market yses,it is not apparent that they imply rela- failure that underlies the call for policy activ tive efficiency of the government over the ism. As a recent example of this tendency, private sector in handling such economic consider the proposal for a tax-based incomes disturbances as oil crises, harvest failures, or policy(TIP), which involves a tax penalty for even autonomous changes in liquidity prefer- price or wage changes above some amount ence or the perceived marginal product of and a reward for changes below this amount capital-if such shifts occur on a significant I honestly have no idea what sort of private scale. For example, uncertainty and mobility market failure or externality is supposed to costs seem to imply that the allocation of ratic this sort of government interfer sources is a difficult problem--not that the ence with the price-setting process.(However, through I'm sure it h active use of its macro-policy instruments. In under the money demand function that is any event the theoretical case for activism occasionally used to measure the welfare loss would, as in areas like industrial organization from anticipated inflation. It is unclear why and the production of"public goods, "require there is some asymmetry that leads indi as a first step some serious analysis of private duals or firms toward"excessive "rather than market"failure " For example, a frequently insufficient, price changes. Casual arguments e to the oil about external effects from"price leadershi pansion of the money stock or the like do not seem helpful in this respect or the government deficit. I do not see how to Additionally it is unclear whether the TIP construct an economic analysis of private proposal is directed at the costs of the average sector allocation that would imply that the rate of inflation, the uncertainty of inflation ecessary and difficult private adjustments to or to some perceived interplay between (I this type of real--unexpected, but presum- assume, unanticipated) inflation and unem ably perceived-disturbance would be as- ployment. It is also hard to reconcile the plan sisted by an increase in the quantity of with the irresistable link between monetary (I also do not see how this policy expansion and inflation, although the propon- would be called for on grounds of ents may have in mind some subtle pressure here some prices are arbitrarily held fixed to support the TIP plan or other forms of does not seem illuminating. general price controls has not been on the It is not difficult to construct a theoretical same level of economic analysis as that model in which the natural rate of private weak as it may be-which has been provided output is too low, relative to some ideal, to defend government regulation of certain because of external effects. Fo or example, the industries, control of pollution, and so on taxation of market earnings and the existence This lack of a theoretical argument might be of welfare programs for the unemployed drive provisionally acceptable in the light of a wedge between private and social product, supporting empirical evidence on the benefits
AMERICAN ECONOMIC ASSOCIATION MAY /979 of such a policy but the record of previous The formation of expectations is one dimen- price control programs does not seem impres- sion in which the private sector might operate less (or even m nore? effic iciently than the Modern courses in macroeconomics utilize government- other seemingly comparable di price theory to a considerable extent. Howev- mensions would include the range of permiss er, these "micro foundations"are usually ble contracts, coordination of trades, pro- limited to the formulation of sectoral supply duction of information, responsiveness of and demand functions, rather than to the supply and demand to money illusion, and so analysis of"general equilibrium. Despite on. However, it is possible to produce Keyne oddities in some earlier treatments of labor sian policy conclusions in models that incor- supply, the serious problem with non- porate rational expectations, but which market-clearing-type models are not in the contain some other departures from sensible characterization of supply and demand, but behavior, for example, arbitrarily fixed nomi- rather in the neglect of the other branch of nal wages, or money illusion either in supply price theory: namely, supply equals demand. and demand functions or in the form of Supply not equal to demand as a basis for private labor contracts. Thus the nature of the quantity determination in non- market-cl formation of expectations seems to be an models is not on the same analytical level important issue within the general context of as supply equals demand. The latter mechan- the efficiency of private arrangements rela- ism implies that-at least in a direct sense- tive to governmental actions, but it is this the private market manages to exhaust trades general concept of relative efficiency that that are to the perceived mutual advantage of seems to be crucial in evaluations of policy the exchanging parties. On the other hand, by activism mechanically leaving opportunities for mu Monetary control is one area in which a tually desirable trades, the non-market-clear- strong governmental role is generally ac- ng approach makes government policy activ- cepted. I abstract here from important issues sm much too easy to justify. When the that involve the transactions benefits of a arbitrariness of supply unequal to demand is generalized medium of exchange and the replaced by a serious explanation, such resource costs that can be saved by using a imperfect information about exchange oppor- fiat standard for money, rather than a tunities, for the failure of private markets to commodity standard. Even with this abstrac achieve some standard of efficency, the case tion a rationale for government involvement or government intervention becomes much in monetary control can be constructed on less obvious business cycle grounds--in particular, from a Let me now consider some specific issues consideration of the Phillips curve, which I relating to information and macro policy. I view here as a representation of the respon begin with the role of expectations in macro siveness of economic activity to unanticipated nodels. Thanks especially to the work of movements in money and the absolute price Robert lucas, we have a much better idea of level. The potential for confusion between the significance of well-informed private ex- absolute and relative price changes in a mone- pectations in macro analysis. This ary economy, which is a possible basis for the cance arises in at least three areas: positive Phillips curve, seems to justify some public analyses of the effects of monetary and other control over the quantity of nominal money hocks on economic activity, analyses of the A major empirical finding is the central role of government polices; and evaluations role that monetary shocks have, in fact and carrying out of econometric estimation. played in business cycles. The greater year Nonetheless, I agree with the view that to-year stability of the underlying monetary rational vS. nonrational expectations is not mechanism in the United States since World per se the key division between Keynesian and War II-which involves a substantial in- non-Keynesian models and, accordingly, is crease in government regulation--has led to a not the essential basis for a division between smaller amplitude of business fluctuations in activist and nonactivist policy conclusions. comparison to those of either the interwar
VOL 69 NO. 2 AN APPRAISAL OF THE NON-MARKET-CLEARING PARADIGM period or the pre-World War I era. However, tions are unclear to me. However it might be it also seems that the important change worth remarking, from the standpoint of the monetary structure has been a reduction in adjustment costs implied by a shift in mone- the short-run variance of money, rather than tary structure, that an analysis of expecta a move toward an activist feedback policy by tional changes associated with a structural the monetary authority. Specifically, the ten- shift cannot usefully be separated from an dency to increase the money stock at a higher analysis of the changes in the underlying rate in response to a recession-which is a variables that led to the shift in structure. In pattern that appears to originate in the Ful particular, there is no reason to believe that Employment Act period following World expectations about money growth and infla War 1l-does not seem to have contributed to tion are either more or less flexible than the enhanced economic performance underlying structure that generates the actual The effects of the shift to greater year- values of money growth and inflation. There- to-year stability in money indicates the poten- fore, one would not predict that a shift to a tial real effects of changes in the underlying new monetary environment--such as those monetary institutions. The evaluation of this that occurred in the past concerning the particular change (i. e, a comparison of the monetary role of gold or the Federal gold standard money process before World Reserve-would involve an adjustment period War I with today's fiat standard managed by in which expectations lagged behind the the Federal Reserve)involves a complicated changes in"reality. "A shift in the United tradeoff. In my view the benefits of greater States to an institution that delivered a lower year-to-year stability in todays output and average growth rate of money would not employment have been bought with two imply a transition period of unusually high major costs. The first of these concerns the unemployment monetary education during the early years of The interpretation of the Great Depre the Federal Reserve, a process that is doubt- is a key matter dividing policy activitists from less still continuing. The monetary errors of nonactivist. The activist view is that the the interwar period-which seem to be much Great Depression was a symptom of an inher more extreme than those that would have ently unstable private economy that exper arisen under the pre-World War I regime- ienced large gyrations in output and which can be credited with much of the costs of the tolerated prolonged periods of higl Great Depression, as well as with those of the ployment. Governmental activism 1937-38 contraction. Although the gold stan- middle and late 1930's--notably, the high dard was not"ideal, "it did require less levels of public expenditures--are thought to knowledge by the government or anyone else have been helpful in restoring some measure about how the aggregate economy worked. of economic prosperity The second cost of the monetary change is the The alternative view is that the great hronic inflation of the present environment. Depre is in large part a product of Further attained by moving to a monetary institution monetary policy of the Federal Reser pt governmental mistakes; specifically, the ine that first, and most importantly, delivered Further, the governmental interventions asso- even greater year-to-year stability in the ciated with the New Deal, including the money supply, for example, by constraining volume of public expenditures and direct the monetary authority to achieve an approx- price regulations, retarded the recovery of the imately constant growth rate for M,, and economy, which was nevertheless rapid after second, that attained an average monetary 1933. The sharp increase in reserve require growth rate below the roughly 6-7 percent ments was primarily responsible for the 1937- annual rate for M, that is implied by the 38 recession present structure. The precise design of the I do not presently see a fully satisfactory new governmental institution and, more " equilibrium"or"disequilibrium"story of mportantly, the political-economic process the Great Depression. With respect to one that leads to changes in this or other institu- type of equilibrium model that has been
AMERICAN ECONOMIC ASSOCIATION MAY /979 proposed, I do not know of an empirical the form of price controls during the Korean documentation of a major downward shift in War and from 1971 to 1973 did not seem to factor supply during the early 1930,s in have such important output effects any case I rega perceived prices. Actually, the depression hypotheses derived from alternative theories perience does not stand out in this respect, of business cycles as a matter of continuing ince this type of factor-supply story has not priority. Examples of testable hypotheses that been satisfactorily documented empirically are implied by some " equilibrium"macro even for the more mild fluctuations of the models with incomplete information are: 1) post-World War II period. On the other hand, only the unanticipated part of movements in the disequilibrium type of model, which relies money affects"real"variables like output and a nontheory of price rigidities, does not real interest rates; 2) the anticipated part of seem to have more impressive empirical current money movements affects the price level contemporaneously on a one-to-one relates business tary disturbances, the contraction from 1930 prices, a positive effect on anticipated real 1933 seems to be well in line with other rates of return, and an ambiguous effect on experiences. The unprecedented monetary nominal interest rates; 3)an increase in the collapse over this period accords quantita- variance of money would reduce the sensitiv ively with the drastic decline in economic ity of output to a given size money shock and activity. The magnitude of the monetary raise the dispersion of relative prices; and 4) contraction can be appreciated by noting that changes in consumption or in federal tax rates the reduction in(annual average)M,, 1929- are, as a first-order approximation, unpre- 33, averaged 7.3 percent per year. The only dictable from lagged data. The equilibrium ince the Civil wa that show a decline in the money stock are effects of monetary and real shocks, although much milder: 1875-79, 0.8 percent per year this aspect of the theories is less developed for M2; and 1892-96: 0.4 percent per year for The approach is also consistent with real M2(A decline of 10.2 percent in M, occurred effects of government spending, which would 1920-21, but the 1920-24 period shows no depend especially on the substitutability net change in the money stock) etween public and private expenditures in A somewhat greater puzzle is posed by the utility and production functions, and with ecovery periods 1933-36 and 1938-41. real, relative price-type effects of changes in These periods may have exhibited a slower taxes, unemployment compensation, and the ecovery rate than would have been antici- like. aggregate demand effects of shifts pated, although the average annual growth between taxes and debt issue would arise, at rates of real GNP from 1933 to 1936 of 10.3 most, when these shifts were unanticipated percent and from 1938 to 1941 of 10.4 per- Testable hypotheses from simple Keyne- cent(which would be reduced somew hat if sian models would seem to include: 1) he final date were 1940) are exceptional for increases in money imply increases in output peacetime periods. In fact, these growth rates and decreases in nominal and real interest are about the same as that (10.5 percent per rates, all of which persist over a substantial year)that appears in the reported statistics period. Price responses depend on the initial for the World War II expansion 1941-44. It state of excess demand (whatever that possible that e recove means), but generally the rate of change of 1933 to 1941 involve the retarding influence prices is raised above what it otherwise would f massive governmental interventions into have been. However, the short-run effect on the price-setting process. However, this expla- prices is weak. The role of price and money nation must be regarded as highly tentative, expectations is not stressed in simple models especially since governmental interventions in of this type. 2)Increases in the government
VOL 69 NO. 2 AN APPRAISAL OF THE NON-MARKET-CLEARING PARADIGM deficit, produced by higher expenditures or stickiness, the cyclical behavior of real wages reduced taxes with the money stock held is indeterminate fixed, imply increases in output-which are I will not attempt at this time to present a likely to be multiplicative-and increases in detailed appraisal of the state of current interest rates. These effects persist over a empirical evidence-however, I think that the substantial period. Prices rise at a faster rate major doubts about Keynesian, non-market than otherwise, but the initial price response clearing-type analysis that are prevalent is weak. The nature of government expendi- today are primarily a reflection of perceive tures and expectations about future taxes, empirical inadequacies of the theory, espe prices, etc, are not stressed in simple models cially in an inflationary environment. I expect of this type. 3)Real wages move countercycli- that the final verdict on the usefulness of cally in models that assume only wage sticki- Keynesian economics will also come primarily ness. In models that also assume some price from empirical analysis
