閤 Price Flexibility and output Stability: An Old Keynesian view OR。 James Tobin The Journal of Economic Perspectives, Vol. 7, No. 1.(Winter, 1993), pp. 45-65 Stable url: http://inks.jstororg/sici?sici0895-3309%28199324%0297%3a1%03c45%03apfaosa%3e2.0.co%03b2-f The Journal of Economic Perspectives is currently published by American Economic Association Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/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 html Each copy of any part of a JSTOR transmission must contain the same copyright notice that ap on the screen or printed page of such transmission STOR 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 Mar1522:15:352007
Price Flexibility and Output Stability: An Old Keynesian View James Tobin The Journal of Economic Perspectives, Vol. 7, No. 1. (Winter, 1993), pp. 45-65. Stable URL: http://links.jstor.org/sici?sici=0895-3309%28199324%297%3A1%3C45%3APFAOSA%3E2.0.CO%3B2-F The Journal of Economic Perspectives is currently published by American Economic Association. Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at 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 22:15:35 2007
Journal of Economic Perspectives-Volume 7, Number 1-Winter 1993-Pages 45-65 Price Flexibility and Output Stability: An Old Keynesian View James Tobin n this symposium I shall play the role in which I was cast, the unrecon structed old Keynesian. Time was when I resisted labels and schools naively hoping that our Aedgling science was outgrowing them. I had, to be sure, been drawn into economics when The General Theory was an exciting revelation for students hungry for explanation and remedy of the great Depression. At the same time, I was uncomfortable with several aspects of Keynes'theory, and I sought to improve what would now be called the microfoundations of his macroeconomic relations The synthesis of neoclassical and Keynesian analysis achieved in the 1950s and 1960s promised a reconciliation of the two traditions, or at least an understanding of the different contexts to which each applies. The hope and he promise were premature, to say the least. In the last 20 years, the dominant trend in macroeconomics has dismissed Keynesian theory. Nevertheless Keynesian models continue to prove useful in empirical applications, forecast- ing and policy analysis. Macro-econometric models are mostly built on Keyne- and practice, are chronic sources of malaise in our discipl a between theory sian frameworks. The gulfs between doctrine and observation, I have benefitted from Gregory Mankiw's"refresher course"in modern macroeconomics(1990). He writes that recent developments--methodological a James Tobin is Sterling Professor Emeritus of Economics, Yale University, New Haven, connecticut
46 Journal of Economic Perspectives new classical, and new Keynesian-are to old macroeconomics as Copernicus was to Ptolemy. It just takes time before Copernican truths can outdo Ptolemaic approximations in practical applications Considering the alternatives, i do not mind being billed as a Keynesian an old Keynesian at that. But old Keynesians come in several varieties, and peak for no one but myself. nor do I defend the literal text of The General Theory. Several generations of economists have criticized, amended, and elabo rated that seminal work. I shall argue for the validity of the major propositions that distinguish Keynesian macroeconomics from old or new classical macroeconomIcs Summary of the Keynesian Case The central proposition of Keynesian economics is commonly described as follows: " According to the Keynesian view, fluctuations in output arise largely from fuctuations in nominal aggregate demand. These fuctuations have real effects because nominal wages and prices are rigid"(Ball, Mankiw, and Romer, 1988, p. 1). On the contrary, I shall argue that Keynesian macroeconomics neither asserts nor requires nominal wage and or price rigidity. It does and require that markets not be instantaneously and continuously cle prices. That is a much less restrictive assumption, and much less contre It leaves plenty of room for flexibility in any commonsense meaning of the rd cr Keynesian models were said to be vulnerable to the charge that"the ucial nominal rigidities were assumed rather than explained, although"it ras clearly in the interests of agents to eliminate the rigidities they were assumed to create.. Thus the 1970s and 1980s saw many economists turn away from Keynesian theories and toward new classical models with flexible wages and prices"(Ball, Mankiw, and Romer, 1988, P. 2). Those market- clearing models have not just flexible prices but perfectly and instantaneously fexible prices, an assumption that is surely more extreme, more arbitrary, and more devoid of foundations in individual rational behavior than the imperfect flexibility of Keynesian models e central Keynesian proposition is not nominal price rigidity but the neous and complete market clearing, output and employment are frequentl onstrained by aggregate demand. In these excess-supply regimes, agents demands are limited by their inability to sell as much as they would like at prevailing prices. Any failure of price adjustments to keep markets cleared opens the door for quantities to determine quantities, for example real national come to determine consumption demand, as described in Keynes' multiplier calculus
James Tobin 47 For this reason, Keynesian macroeconomics alleges that capitalist societies are vulnerable to very costly economy-wide market failures. Individuals would be willing to supply more labor and other resources in return for the goods and services the employment of those resources would enable them to consume now or in the future, but they cannot implement this willingness in market transac tions. As the quotation from Ball, Mankiw, and Romer suggests, many contem porary theorists cannot believe any theory that implies socially irrational market failures. They suspect that individual irrationalities are lurking somewhere in the theory. In continuously price-cleared competitive markets, they know, individually rational behavior implies collectively rational outcomes. But this theorem does not apply if markets and price-setting institutions do not produce perfectly flexible competitive prices. Individual rationality does not necessarily create the institutions that would guarantee"invisible hand"results. Keynes as not questioning the rationality of individual economic agents; he was arguing that their behavior would yield optimal results if and only if they as citizens organized the necessary collective institutions and government policies In the same spirit though in different contexts, some modern theoretical research has shown that welfare-improving policies may be designed even when asymmetries of information and incompleteness of markets prevent the achievement of global optima Ball, Mankiw, Romer and others style themselves as New Keynesians Their program is to develop improved microeconomic foundations for imper fectly fexible prices. In the process, they hope to illuminate the paradox that individually rational or near-rational behavior can result in significant collective market failures. These are certainly laudable objectives. In the end, I suspect the program will not change the essential substance of Keynesian macroeco- nomics. But it will make Keynes more palatable to theorists In Keynesian business cycle theory, the shocks generating fluctuations are generally shifts in real aggregate demand for goods and services, notably in capital investment. Keynes would be appalled to see his cycle model described as one in which "fluctuations in output arise largely from fluctuations in nominal aggregate demand"(Ball, Mankiw, and Romer 1988, p. 2). The difference is important. The impact on real purchases of a one-time one percent shock to aggregate nominal spending will be eroded if and as nominal prices increase in response, and eliminated once prices have risen by the same one percent as nominal spending did. But suppose it is real demand that initially rises one percent. At the prevailing prices nominal spending will rise one percent too. But if and as prices rise in response the one percent real demand shock becomes an ever larger amount of nominal spending. Its impact is not mechanically eroded by the price response; if it is absorbed, the process subtle and indirect The big issue between Keynes and his"old classical"opponents was the efficacy of the economy's natural market adjustment mechanisms in restoring full employment equilibrium, once a negative real demand shock had pushed
48 Journal of Economic Perspective the economy off that equilibrium. Keynes and Keynesians said those mecha nisms were weak, possibly nonexistent or perverse, and needed help from government policy. That is still the major question of macroeconomic theory and policy, even though new classical economists finesse it by assuming that the economy can never be pushed out of equilibrium even for a moment. Keynes classical contemporaries and predecessors would never have drawn real-world lessons from theories based on such an assumption. Their successors strain redulity when their models imply that markets are cleared and joblessness oluntary when measured unemployment is 10 percent as truly as when it is 5 percent. Keynesian theory of nominal wage stickiness does not deserve the disdain with which it is commonly regarded. It is not dependent on"money illusion But Keynes certainly would have done better to assume imperfect or monopolistic competition throughout the economy, in both product and labor markets. In markets of these kinds, nominal prices are decision variables for sellers or buyers or are determined by negotiations between them. They therefore move only at discrete intervals. Despite considerable effort over the years to give macroeconomics improved microfoundations along these lines there is plenty of scope for the"New Keynesian"program of theoretical and empirical research on this topic In the absence of perfect flexibility, does greater fexibility of nominal prices strengthen the equilibrating mechanisms, or does it weaken them? Keynes doubted that the problems of involuntary unemployment and underuti lized capacity would be mitigated by greater flexibility of nominal wages and prices. On the whole, he favored stable nominal wages. Critics of Keynesian macroeconomics forget this strand of the argument when they assume that without absolute"rigidity"aggregate demand could never be deficient. Fortu nately, this issue has been receiving greater attention in the last few years, witl considerable support for Keynes position Macroeconomics with effective demand Constrained The empirical relevance of Keynesian economics is based on its assertion that situations of pervasive excess supply often occur. An advanced capitalist industrial economy is frequently in a state in which most labor and product markets are not clearing at prevailing prices. As a result, workers are involun tarily unemployed and capital capacity is underutilized. The effective constraint on output is the aggregate demand for goods and services; like he effective constraint on employment is the amount of labor required to produce that Keynesian unemployment must be differentiated from both frictional and lassical unemployment. Frictional unemployment occurs because of
Price Flexibility and Output Stability: An Old Keynesian View 49 microeconomic flux Demands and supplies are continually shifting, bringing unemployment and excess capacity in some sectors and contemporaneous labor shortages and capacity bottlenecks elsewhere. The gross aggregates of these frictional excess supplies and excess demands vary together positively over time. In contrast, cyclical excess supplies and demands are negatively corre lated; in economy-wide recessions and depressions, excess-supply markets and sectors predominate, while the reverse is true in inflationary booms. The amount of frictional unemployment depends on the strength of intersectoral shocks and on the mobility of factors of production in responding to them Large and protracted shocks, for example in technology or in supplies and structural unemployment. Neither is remediable by demand expansion aloe a prices of key commodities like energy, convert frictional unemployment A common species of classical unemployment occurs when jobs are limited ecause of excessive real wage rates imposed by governmental or trade union regula tions For individuals who would like to work at or below the wage floor, such unemployment is involuntary. For the workers collectively whose bargain ing strength or political clout established the regulations, the unemployment could be regarded as the voluntary consequence of their exercise of monopoly power Identification of observed unemployment as classical or Keynesian is some times difficult. In either case unemployment might be observed to be associated with real wages above their full employment equilibrium values. In the Keynesian case, this could result from perfect competition among producing firms; they would be paying workers the high marginal products associated with low employment. The big difference between the two cases is that in the Keynesian case, but not in the classical case, real wages would decline on their own and output and employment would increase in response to expand demand. In the classical case removal of the regulations would be essential There are several variations on the classical unemployment theme. One case is queuing for a high-wage job. An artificially high wage in a particular sector could draw workers from employment elsewhere to wait and hope. This model was originally designed to explain the heavy unemployment in the urban centers of developing countries, where the queuing requires living near the scarce jobs, far from alternative means of subsistence in traditional agricul ture. It fits less well in advanced economies where workers can search and apply for better jobs while employed. Another source of voluntary unemploy ment may be unemployment insurance benefits and other transfers that in rease the reservation prices of persons without jobs. However, in the United States, where unemployment is measured by large household surveys con ducted monthly by the Census, persons without jobs will be counted not as unemployed but as"not in labor force unless they report they have been actively searching. Although some misreporting doubtless occurs, it is small,not always in the same direction, and cannot begin to account for the cyclical variability of unemployment rates
50 Journal of Economic Perspectives Agents who are unable to sell as much as they would like at prevailin prices restrict demands in other markets. Unemployed workers cut their consumption. Demand-constrained firms restrict their hiring of labor and their purchases of other inputs. Keynes' insight that quantities actually sold, if maller than sales desired at existing prices, will keep demands in other markets below equilibrium values, was rediscovered and elaborated by self-styled disequilibrium theorists"30 years later(Barro and Grossman, 1971). In old Keynesian economics, multiplier theory formalized the determination of quar tities by quantities. It did not and does not, however, preclude the relevance of other determinants of demand notably prices and interest rates. In this respect it is more general than most of its latter-day extensions in "disequilibrium heory. "In demand-constrained regimes, any agent,s increase in demand-for example, more investment spending by a business firm-has positive externai ties. It will increase the attainable consumption of third parties. In some modern literature, this idea of Keynes is revived and elaborated under the lal strategic complementarity"(Cooper and John, 1988) Liquidity constraints are an important but extreme form of effective de- mand constraint. Some wage earners, no doubt, depend on each week,s wages to buy the goods for that week,s consumption. But Keynes' principle does not depend on such short horizons for consumption-smoothing. Expectations of future spells of unemployment, enhanced by present and recent experience, can limit the current consumption and durables purchases even of long horizon households. Liquidity constraints and prospective effective demand constraints can also limit business investment. Common observation suggests that households and businesses, and governments too, differ ly in their horizons, i. e. the length of the future period over which expected resources are egarded as potentially available for spending today. These hor rizons. more over, doubtless change over time with circumstances and behavior. The multipliers relating change in aggregate demand to demand shocks from policies or other events, are not as large as they were thought to be when the concept was first introduced and estimated in the 1930s. One reason is a substantial structural change in democratic capitalist economies. Governments are much larger relative to private sectors than before World War Il, and their fiscal institutions are"built-in stabilizers. " Their expenditures are quite unre- sponsive to current business conditions, while their revenues (net of transfers to the private sector) are cyclically sensitive and thus moderate swings in private incomes. A second reason is that economists have come to recognize that thanks to accommodating capital markets as well as to their own foresight, most economic agents have horizons longer than one year. While this consideration implies that multipliers of transient shocks are lower than for permanent changes, it by no means implies that they are zero Both consumption and investment appear to be sensitive to contemporaneous and recent incomes. For most agents capital markets are far from perfect; in particular future and current labor incomes are not fungible. Moreover
James Tobin 51 expectations of economic futures, individual, national, and global, are inft enced by current events, perhaps to an irrational extent. As Keynes explicitly observed, his theory refers to economies with incom- plete markets. In his day futures markets were rare, and contingent futures markets even rarer. They are still scarce. As Keynes explained, decisions not to spend now are not coupled with any definite orders for future or contingent deliveries. Typically they result in accumulations of assets that can be spent on anything at any future time. The multiplier effects of lower current spending propensities are not offset by specific and firm expectations of higher future eman Business Cycles as Demand Fluctuations According to Keynesian macroeconomics, business cycles are fluctuations in aggregate effective demand, carrying output and employment in their wake. They do not reflect movements in market-clearing supply-equals-demand rla Supplies of labor and other factors of production move fairly smoothly from year to year and from cycle to cycle. So does economy-wide factor productivity, largely reflecting technological progress. Equilibrium output and employment cannot be as variable as actual cyclical observations. In the neoclas- sical neo-Keynesian synthesis, trend growth is supply-determined; markets are cleared; supply truly creates its own demand. In cyclical departures from trend demand evokes its own supply. Keynesian short-run macroeconomics does not pretend to apply to problems of long- run growth and development Equilibrium cycle theories(Plosser, 1989) are unconvincing. They rely on incredible volatili hnology, retrogressive as well as progressiv rely on extreme intertemporal substitutions among work, leisure, and con- sumption. Or they contrive informational asymmetries and misperceptions that seem easy to correct. For example, a few years ago a popular theory attributed business cycles to confusions by suppliers of products and labor between increases in their own real prices, on the one hand, and economy-wide infla tion, on the other. Evidently businesses and households were assumed to ignore the food of current statistics on prices and money supplies I am using the word equilibrium to mean Walrasian market-clearing by prices, as is the current usage of both new classical macroeconomists and isequilibrium theorists. Keynes used it otherwise, to refer to a position of rest. That is why he referred to outcomes with involuntary unemployment as equilibria on a par with full employment, and why he termed his the general"in the title of his book. The basic issue is not semantic. It is whether situations of general excess supply can and do exist for significant periods of ime, whether or not they are called equilibria
52 Journal of Economic Perspective Some passages of The General Theory can be read to assert that involuntary unemployment is much more than a temporary cyclical phenomenon, that it is in the absence of remedial policies a chronic defect of capitalism. This was a atural enough view in the 1930s. In Alvin Hansen's American Keynesianism (e. g, Hansen, 1938)secular stagnation was a central proposition. Formally, however, the analysis of The General Theory is limited to a time period short enough that the changes in capital stock resulting from non-zero investment can be ignored Postwar Keynesians, for the most part, have not regarded protracted depression as a likely outcome. Chronic inflationary gaps could also occur, and alternations between excess-supply and excess-demand regimes were highly probable. Keynesian macroeconomics is two-sided. Deviations on both sides of Walrasian market-clearing can occur, though not necessarily with symmetrica symptoms. Excess demand in aggregate is mainly an"inflationary gap, gener ating unfilled orders and repressed or open inflation, rather than significant extra output and employment. Macroeconomic stabilization requires two-sided countercyclical demand management. In any case, habitual application of Keynesian remedies reinforces what- ever natural mechanisms tend to return the economy to its full employment growth path. Expectations that those remedies will be used contribute to the stability of that equilibrium path The Efficacy of Classical Adjustment Mechanisms: Interest Rates uppose that shocks to current real demands for goods and services create, at existing prices and wages, excess supplies of labor and capital services. What are the variables whose changes would avert or eliminate macroeconomic disequilibrium? The leading candidates are current prices, which include both wages of labor as well as prices of products, and interest rates, which involve future as well as current prices. In what follows, I shall set forth Keynesian skepticism regarding the efficacy of these classical adjustment mechanisms If these mechanisms respond instantaneously to shocks, no actual discrep ancy between demand and supply will occur or be observed. The shocks will be wholly absorbed in the market-clearing variables. This is the assumption of equilibrium business cycle theory and of the"real business cycles"approach. It this assumption that, among other things, enables new classical macro- economists to dismiss out of hand real aggregate demand shocks and to react In Tobin(1955), stagnation is one possibility, the stable solution of a non-linear model whose
Price Flexibility and Output Stability An Old Keynesian view with incredulity when Keynesians mention them However if these adjustments do not occur instantaneously but take real time, then Keynesian situations of excess supply do occur. They occur even if prices and interest rates are falling at the same time. The consequence is that the quantity adjustments of the multiplier process start working counter to the possible equilibrating effects of interest rate and price reductions In standard Walrasian/Arrow-Debreu theory, perfect flexibility of all wages and prices, present and future, would maintain full employment equilibrium Short of that, an old question of macroeconomic theory is whether, given current nominal wages and prices, changes in future money wages and prices that is, in nominal interest rates-could do the job In old classical macroeconomics, interest rates are the equilibrators of both capital markets and goods markets. Their adjustment is crucial to the Say's Law story, which dismisses as vulgar superficiality notions that an economy could suffer from shortfalls in demand for commodities in aggregate Market interest rates keep investment equal to saving at their full-employment levels-and therefore keep aggregate demand equal to full employment output-even if nominal product prices and wages stay put. Indeed classical doctrine is that the real equilibrium of the economy is independent of nominal prices, as if it were the outcome of moneyless frictionless multilateral Walrasian barter Can interest rates do the job? The Keynesian insight is that the institution ally fixed nominal interest rate on currency, generally zero, limits the adjust- ment of nominal interest rates on non-money assets and imparts to them some stickiness even when they are above zero. As a result, after an aggregate demand shock they may not fall automatically to levels low enough to induce sufficient investment to absorb full employment saving. As a result, aggregate demand-consumption plus investment-will fall short of full employment supply The case for significant non-zero interest elasticity of money demand is simply that the opportunity costs of holding money fall as the interest rates available on non-money substitutes decline. As those rates approach the interest paid on money itself, zero at the lowest, the opportunity costs vanish. The interest rate on money sets the floor for other nominal market interest rates The familiar specific money demand models-transactions costs, risk aversion regressive interest rate expectations-all depend on the fixed nominal interest The interest-elasticity of money demand is a key parameter in macroeco- nomic theory. Three cases can be distinguished. One is a classical extreme, often associated with the quantity theory of money: the elasticity is zero. At the other extreme is the Keynesian liquidity trap market interest rates are so close to the foor that people are on the margin indifferent between money and Dudley Dillard (1988)calls this the"barter illusion"of classical economics
