Democracy in Deficit: The Political Legacy of Lord Keynes
What about Full Employment?
Much of our argument in this book may find widespread acceptance. But many who find our diagnosis persuasive may reject our implicit prescription of a return to the old-fashioned norms of fiscal conduct. The fiscal policy clock cannot simply be turned back, and, since the Employment Act of 1946, the United States has been committed to pursuing policies that promote full employment. Those who accept our diagnosis but who balk at our implied remedy are likely to ask, "What about full employment?" The old-fashioned medicine might have been fine for the pre-Keynesian era, but the complex economic and political setting of the modern world may seem to dictate the administration of more potent Keynesian-like elixir. Can we really do other than pursue activist fiscal policies until the economy gets on and stays on its potential growth path? Perhaps then, and only then, we might return to something resembling the rules.
Current Unemployment and the Quandary of Policy
With the passage of the Employment Act of 1946, Congress declared that it was the
policy and responsibility of the Federal Government ... to coordinate and utilize all its plans, functions, and resources for the purpose of creating and maintaining ... conditions under which there will be afforded useful employment for those able, willing, and seeking to work.
The Council of Economic Advisers was created to assist in the implementation of this act. Just what constitutes "full employment" was not defined in the act, and, moreover, it is an inherently unobservable variable.
Despite this necessary imprecision in definition, the Employment Act is exceedingly significant, for with its enactment, the Congress sanctified the principal thrust of the Keynesian analysis. By implication, the free enterprise economy was officially conceived to be unstable; it became the task of government to act as a balance wheel to promote stability and growth. The Employment Act mandated that the government practice compensatory policy so as to promote full employment. Whenever employment declined below that level defined to be "full employment," government seemed legally to be required to undertake expansive budgetary and/or monetary policy.*69
But what is full employment? The Keynesian economists have never been precisely clear on this question, and their acceptable targets have been moved progressively downward through time. In the 1962 Report of the Council of Economic Advisers, full employment was officially defined to be present when there was a 4-percent measured rate of unemployment.*70 This was, however, taken to be only some interim rate, with the long-run objective rate considerably lower. The much-discussed Humphrey-Hawkins bill in 1976 set out a 3-percent target rate, to be attained within four years. Historically, however, unemployment in the United States has only infrequently, and then only temporarily, fallen below 4 percent. During the sixty-three-year period after the creation of the Federal Reserve System, the period 1913-1975, there were only twenty years in which unemployment averaged 4 percent or less. Ten of these years—1918, 1943-1945, 1951-1953, and 1966-1968—occurred during periods when the United States was at war, although the mobilization was not so intense in the latter two periods as it was in the former two. The normal rate of unemployment, though unmeasurable, would appear to lie somewhat above 4 percent. During the fifteen-year period 1946-1960, a period that included both moderate mobilization and moderate recession, as well as predating the full Keynesian conversion of our politicians, unemployment averaged 4.5 percent. Even if we could make the heroic assumption that the institutional-structural features affecting employment have remained invariant over the long periods noted, it would seem clear that the "normal" rate of unemployment lies considerably above 4 percent. And, of course, it would be illegitimate to make such an assumption of invariance through time. Both demographic shifts within the labor force (toward younger and female workers) and policy changes (unemployment compensation, extended minimum wage coverage, increases in welfare payments and retirement support) have had the effect of increasing the level of unemployment that would be consistent with any specified rate of inflation.*71
A national economic policy targeted to achieve, say, a 4-percent rate of unemployment is likely to be inflationary on the one hand and unattainable on the other, especially in the long run, at least in the absence of corrective policies aimed at structural features of labor markets. Efforts to attain such a rate of measured unemployment would probably generate increasing rates of inflation, with little demonstrable effects on employment itself.*72
Since 1964, when we entered the Great Society stage of our national history, we have lived through a period in which federal budget deficits have been increasing rapidly, along with explosive growth in the size of government. In the early years of this period, 1964-1969, unemployment fell steadily, along with continuing inflation. Taken alone, this mid-1960s experience might suggest a changing trade-off between inflation and unemployment. But economic life is not so simple. Inflation may reduce unemployment for a time, but it also distorts the structure of the economy in the process. Such structural distortions may, in turn, require an increase in unemployment before the economy can make the readjustments that are necessary to dissipate the distortions. This day of reckoning can be postponed only through ever-increasing inflation, or through the replacement of the free economy by a command economy, one in which direct controls are imposed. In the years after 1969, unemployment increased along with inflation, and the accelerated unemployment experienced in the recession of 1974-1975 attests to the difficulty of slowing down a rate of inflation once started.
The combination of substantial unemployment and inflation creates a quandary for economic policy. The Employment Act of 1946 seems to commit the nation to public policies designed to promote full employment. In the simplistic Keynesian theory of economic process (and the Employment Act implied an acceptance of this Keynesian theory), total employment varies directly with the volume of total spending. The mere presence of unemployment provides a signal for expansionary policies that increase aggregate or total spending in the economy. Inflation, by contrast, can be alleviated only through contractionary policies that reduce aggregate spending. Unemployment calls for expansion, while inflation calls for contraction. This is the quandary, pure and simple.
The Keynesian Theory of Employment
It will be useful again to summarize the basic Keynesian theory of economic policy that allegedly supports the politically dominant policy paradigm. If some such extraordinary and exogenous force or event, a collapse of the banking-monetary structure, a revaluation of the national currency at some disequilibrium level, or a major physical catastrophe, has generated a reduction in the aggregate demand for goods and services in the economy, a reduction that has dramatically modified business expectations, output and employment will have been reduced, possibly along with prices and wages, although the latter response may lag behind the former. In this setting, an explicit program of expanding aggregate spending in the economy through fiscal measures (the efficacy of monetary policy may be temporarily reduced by the existence of excess liquidity) will modify business expectations and will succeed in expanding the total volume of spending relative to total labor costs (and, to employers, this will amount to a reduction in real wages as a share of product prices). As a result, output and employment will expand, possibly along with some increase in prices, although the latter may again lag behind the former.
This summarizes our understanding of the theory of macroeconomic policy, as presented by Keynes himself, with the one exception that he falsely proclaimed it to be a general theory, presumably applicable to economic environments in which no extraordinary event has occurred at all, environments that are not remotely akin to those of the depressed 1930s. It seems quite likely that Keynes, always willing to change his mind, would have quickly abandoned these claims to generality had he lived into the years following World War II. But his followers, the Keynesians who became his disciples charged with spreading the gospel, made a simplistic extension of the basic model to economic environments for which the whole "theory" is clearly inapplicable.
If, instead of some extraordinary and exogenous event that has literally plunged the economy into a depressed state, financially and psychologically, endogenous structural features of the market (including governmentally enforced regulations and restrictions) have generated a level of unemployment (say, 5, 6, 7, or even 8 percent) that is deemed "unacceptable" against some arbitrarily chosen standard (with, say, a 3-percent or a 4-percent target), the policy norm derived from the Keynesian model may not be at all appropriate. In such a setting, the Keynesians would have us apply essentially the same expansionary tools as those applied in the extraordinarily depressed economy. Aggregate spending "should" be expanded so as to increase the level of employment. And, as we have noted, the Keynesians did succeed in convincing the politicians to this effect.
When this falsely applied theory of policy is appended to an apparent legislative mandate for the achievement of "full employment," a mandate that was, itself, a product of depression mentality, the policy prescriptions become straightforward. When unemployment exists, for any reason, the stream of spending must be increased. Conversely, any policy that reduces aggregate spending must increase unemployment. Unfortunately, this states all too perfectly the macroeconomic policy paradigm of modern democracy. Any politicians who want to appear responsive to the needs of the unemployed must support expansionary fiscal measures.*73 (The parade of presidential hopefuls in 1976 who mouthed the simplest of Keynesian propositions should, in itself, offer substantive proof for the central argument of this book.) Only some misanthropic capitalist or his lackey would suggest that the unemployment observed in the 1970s may not be much reduced by further increases in total spending, that such reductions that could be achieved might be short-lived, and that these could be secured only at the expense of accelerated inflation. Herein lies what may properly be considered our national economic-political tragedy, one that finds its origins in ideas that were both imperfectly understood and inappropriately applied.*74 Having entered the realm of political discourse, however, these ideas cannot be readily exorcised by the empirical findings of the academic economists.
The Inflation-Unemployment Trade-off
In its early textbook formulations, although not in Keynes' own work, the Keynesian theory of macroeconomic management posited a categorical distinction between the conditions under which expansionary policies would be inflationary and the conditions under which they would generate noninflationary increases in employment and output. Up to a certain level of employment, expansionary policies would elicit increases in employment and output, but without increasing prices. Beyond that level of employment, expansionary policies would increase prices, but without increasing employment and output.
It was always recognized that this view was but a simplified representation, though one that captured adequately the central features of the phenomena under examination. By the late 1950s, this simplistic view gave way to the widespread realization that acceptably full employment and stability in the value of money might be inconsistent. In an economy with strong labor unions and/or governmental wage floors, both full employment and a stable price level were not likely to be achieved; some trade-off was necessary. If unemployment was to be reduced to tolerable levels, increasing prices might have to be tolerated. The empirical basis of this trade-off was the so-called "Phillips curve," and the inflation-unemployment relationship became the focal point of almost all discussions of macroeconomic policy after the late 1950s.*75 Inflation and unemployment became matters of political choice, and a politician who assigned weight to stability in the purchasing power of money was automatically branded by popular opinion as someone who favored higher rates of unemployment, as someone who would deliberately create food-stamp lines.
Like the simplistic Keynesian theory of employment, the inflation-unemployment trade-off possessed a striking attention-arresting power. This trade-off came to dominate the images people formed as to the nature of reality, and democratic politics conformed to this image. When this is combined with the view that inflation is a problem not nearly so severe as unemployment, the inflationary bias becomes transparent. What decent politician could countenance greater unemployment simply to attain price-level stability?
By the late 1960s, the foundations of the inflation-unemployment trade-off began to erode, in the minds of academicians, though not in the minds of citizens and politicians. The Phillips curve, it came to be realized, described only a short-run, not a long-run, trade-off.*76 Expansions in aggregate demand, accompanied by some inflation, could reduce unemployment in the short run, but only because the inflationary effects were not fully anticipated. Once the predictable effects of inflation on real wages came to be understood, permanent structural features of the economy would reassert themselves. As expectations came to be adjusted to inflation, unemployment would rise to roughly the level determined by these structural features. The price level would, of course, be higher because of the expanded monetary spending, but there would be no permanent increase in employment. A permanent increase in prices would be the cost of, at best, a temporary and short-lived reduction in unemployment.
The idea that the inflation-unemployment trade-off can be exploited only in the short run, not in the long run, embodies the notion of a "natural rate" of employment. This natural rate of unemployment corresponds to full employment in the classical, non-Keynesian analytical framework.*77 Unemployment can be reduced below this natural rate only by a continually accelerating inflation. If, for instance, a 4-percent rate of unemployment is accepted as an objective of macroeconomic policy, and if the natural rate of unemployment lies above 4 percent, say in the vicinity of 5 percent, pursuit of the standard policies of Keynesian management will produce a continually accelerating inflation in pursuit of this unattainable objective.
The Inflation-Unemployment Spiral
This has come increasingly to be recognized by economists, and much of the earlier academic support for expansionary fiscal-monetary policies based on the alleged Phillips-curve trade-off has disappeared. Politically, however, the arguments proceed unchecked, and it seems highly unlikely that elected politicians can adhere to the discipline that would be required to escape from the inflationary spiral. The experience with the recession of 1974-1975 offered a test.
Why was the American economy characterized by "stagflation" in 1974 and 1975? Unemployment seemed clearly to be above any plausible "natural rate," while, at the same time, the price level continued to rise. There are two complementary parts of an explanation for this phenomenon. The double-digit inflation set off by the fiscal-monetary excesses of 1972 and 1973 set in motion fears of explosive hyperinflation; this caused the monetary authorities to decelerate the rate of monetary expansion. But, because of built-in expectations of the public concerning the continuance of accelerating rates of inflation, any deceleration in the rate of monetary expansion had an effect essentially the same as an actual reduction in aggregate demand would produce in a setting of monetary stability. The responses of business firms and consumers brought about reductions in output and employment, though upward pressures on prices were maintained, both because of lagged effects and because the deceleration was still within a range that allowed for continued inflation. Unemployment moved upward to levels that were probably well above the sustainable natural rate, given the structural features of the United States' economy.
"Stagflation" could, therefore, be explained in this way, even if we should leave out of account the internal allocative distortions that long-continued inflation itself produced or might have produced. Once we allow for such distortions, any attempt to decelerate the rate of monetary expansion will generate more pronounced output and employment responses, with the magnitude of these being related directly to the length of the inflationary period. It is this second part of the overall explanation of "stagflation" that gives rise to the view that continued inflation, in itself, must be a direct cause of greater unemployment in any future period when any deceleration in the rate of inflation takes place.
Inflation has two effects. One effect is to lower real wages; this was Keynes' thesis, and it is this effect that is recognized by those who argue that, once inflation is fully anticipated, the rate of unemployment cannot be permanently reduced below its natural level. This effect gives the customary result of a short-run trade-off between inflation and unemployment, while any such trade-off is absent in the long run.
The other effect is a possible general disruption of the working of the market economy, reflected in the reallocations of resources that take place in response to the inflation-caused short-run shifts in relative prices, as well as in the reallocations that result in response to the increase in the rate of mistakes owing to the increased difficulty of rational economic calculation. As these allocative distortions become more important, a policy shift away from accelerating inflation both will generate a temporary increase in unemployment above the natural rate and will alter the length of time required to return to the natural rate. If the economy is to "recover" from the Keynesian-inspired inflationary spree, the recovery phase will necessarily involve unemployment above the natural rate, and this phase can be regarded as the "price" of the mistakes of the past.
Suppose, however, that a government necessarily subject to Keynesian pressures, and faced with an observed rise in unemployment above what seems to be the natural rate, does not carry through its anti-inflation commitment and decides instead to stimulate aggregate demand once again. This stimulatory policy unleashes two opposing forces. The temporary fall in the real-wage rate may reduce unemployment somewhat. But the removal of the allocative distortions is postponed and still further distortions are encouraged, which will, in turn, make a subsequent policy of retaining monetary stability more difficult.
If, on the other hand, an anti-inflation policy is carried through, and a rate of unemployment in excess of the natural rate is tolerated for a sufficient time, the economy will eventually "recover" in the sense that the rate of unemployment consistent with basic structural features will be attained. But if the day of reckoning is postponed through periodic injections of Keynesian stimulation, an inflation-unemployment spiral can develop. New doses of inflation may initially stimulate employment, but they will also increase the volume of mistaken, unsustainable economic decisions. The misallocations that stem from such errors will, in turn, generate reallocative unemployment. Should the economy be shocked with still a further dose of Keynesian expansionary medicine rather than being permitted gradually to adjust to its natural equilibrium, some short-run reduction in unemployment may take place once again. But additional mistakes, resource misallocations, will be again injected into the economy, and the rectification of these mistakes will, in turn, generate additional unemployment. In this way, a spiral of inflation and unemployment can result from Keynesian prescriptions. Inflation creates allocative mistakes, and these mistakes cannot be rectified costlessly. There are readjustment costs that simply must be borne before the economy can return to normal.*78 The day of reckoning can only be postponed, not forgiven, and the longer it is postponed, the more frightful the eventual day of reckoning becomes.
Biting the Bullet
A policy of attempting to reduce unemployment through the stimulation of aggregate demand is shortsighted in a situation that is not structurally abnormal (such as the 1930s). Unemployment may be reduced temporarily, but the inflation will exacerbate the maladjustments contained within the economy. There is no costless cure for a maladjusted economy. Reallocations of capital and labor must take place before the economy's structure of production will once again reflect the underlying data to which a free economy adapts. The mistakes that result from people responding to the false signals generated by inflation must be worked out before the economy can return fully to normalcy. Recession is an inherent part of the recovery process; it is the economic analogue to a hangover for a nation that is drunk from Keynesian stimulation.
To attempt to maintain "full employment" is an act of delusion. The readjustments can be postponed, though with ever-increasing difficulty, but they cannot be prevented, at least within the context of a free society. The inflation-unemployment spiral that results from shortsighted efforts at demand stimulation will simply increase the dissonance between people's aspirations and their realizations. As a result, democratic institutions become more fragile. In Britain in the late 1970s, where the policy dilemma has been even more severe than that of the United States, there has been widespread discussion of the prospects of calling in some "leader," empowered to deal effectively with the issues, reflecting a yearning that emerges when people lose their faith in the ability of ordinary democratic process to produce meaningful patterns of economic and social existence.
The Employment Act of 1946, one of our legacies from Lord Keynes, may come to be regarded as one of the more destructive pieces of legislation in our national history. This act pledges the government to do something it cannot possibly do, at least so long as our underlying fiscal and monetary institutions are themselves the primary source of instability. And if fiscal and monetary sources of instability were removed, there should be no need for an Employment Act. The political system is burdened with claims on which it cannot possibly deliver, at least within the context of a nonregimented society. The act has an inflationary bias, a bias that, as Joseph Schumpeter noted with remarkable perceptiveness and frightening prescience, can ultimately topple a liberal, democratic civil order.
So, What about Full Employment?
Our response to the query with which we opened this chapter must necessarily be a seemingly elusive one, for the mind-set within which this question would be posed to us is one that has been molded by the legacy of Lord Keynes. Discussion has been unduly concentrated on the end-result objective of employment, to the neglect of the processes within which end results are produced.*79 Involuntary unemployment is indeed undesirable, but we must try to understand the institutional processes that may have produced an observed result before we act on the end result itself. It is a false vision of reality to infer that the selfsame processes are always at work, that involuntary employment necessarily reflects deficiency in aggregate demand.
In a competitive market system of economic organization, there will, of course, be instances of observed involuntary unemployment. This unemployment will result from such factors as shifts in consumer preferences, the development of new technologies, and errors made by businesses and consumers. Insofar as elements of the economy are noncompetitive, observed instances of involuntary unemployment may be more widespread. Job search becomes more difficult as more areas are closed by restrictions on entry. Nonetheless, all such instances of involuntary unemployment will tend to be, first of all, of short duration, and they would tend to occur fairly evenly over time. The concentration or bunching required to produce what might be considered an economywide recession or depression is a quite different phenomenon, and must be explained by economywide causes. In the Great Depression of the 1930s, the involuntary unemployment did result from a deficiency in aggregate demand, produced largely if not totally by the failure of governments to maintain a stable monetary-fiscal framework.*80 The Keynesian emphasis tended to neglect the more general institutional or framework features, which allow us to make a conceptually distinct separation between the legitimate responsibility of government to maintain a stable monetary-fiscal framework and the sometimes expedient extensions of government activity beyond such limits, extensions that tend, in themselves, to be a source of instability.
Our answer to the query in the chapter title, then, must be a roundabout one. Full employment should not, indeed cannot, be promoted directly through government policies of aggregate demand management. Such policies merely compound past mistakes with present mistakes, thereby making the economy perform ever more poorly. Full employment can be promoted only through a regime in which government conducts its affairs in a manner that avoids injecting new sources of instability into the economy. We shall examine these constitutional principles for stability in the next chapter.
Notes for this chapter
The concluding clause of the paragraph in which the citation in the text appears goes, "and to promote maximum employment, production, and purchasing power." This wording suggests that the commitment to full employment is not absolute, and the discussion of the trade-off between unemployment and inflation that surfaced in the late 1950s was a response to this perceived ambiguity. What this meant in practice was that full employment was only a relatively absolute absolute.
Economic Report of the President, 1962, p. 46. "Okun's Law," which was developed to give a quantitative estimate of the loss from unemployment, is based on this 4-percent measure, for its measure of the loss from unemployment is
For a brief general discussion, see Michael L. Wachter, "Some Problems in Wage Stabilization," American Economic Review 66 (May 1976): 65-66. For a specific analysis of unemployment compensation, see Martin Feldstein, "Unemployment Compensation: Adverse Incentives and Distribution Anomalies," National Tax Journal 17 (June 1974): 231-244.
Michael L. Wachter estimated that the rate of unemployment consistent with price-level stability was 5.5 percent in 1975, which he admitted to be a minimal figure. See Wachter, pp. 66-67.
Werner Zohlnhöfer has developed a similar point in much more general terms. He argues that a politician or party in opposition to the ruling coalition can never enhance chances of success by lending support to anti-inflationary measures. See Werner Zohlhhöfer, "Eine politische Theorie der schleichenden Inflation," in Schriften des Vereins für Socialpolitik, Gesellschaft für Wirtschafts- und Sozialwissenschaft, N. F. Band 85/1 (1975), pp. 533-555.
By common argument, the situation in Great Britain in the late 1970s was even worse than that in the United States. For a set of assessments in 1976, see John Flemming et al., Catch 76, Occasional Paper no. 47 (London: Institute of Economic Affairs, 1976). Also, see Peter Jay, Employment, Inflation, and Politics, Occasional Paper no. 46 (London: Institute of Economic Affairs, 1976).
As noted earlier, the basic paper is A. W. Phillips, "The Relation between Unemployment and the Rate of Change of Money Wages in the United Kingdom, 1861-1957," Economica 25 (November 1958): 283-299. In this article, the trade-off was between wage inflation and unemployment. This relation was subsequently transformed into the more familiar relation between price inflation and unemployment. The American discussion was inaugurated by the famous Samuelson-Solow paper. See Paul A. Samuelson and Robert M. Solow, "Analytical Aspects of Anti-Inflation Policy," American Economic Review 50 (May 1960): 177-194.
A survey of alternative perspectives toward the Phillips curve is presented in Thomas M. Humphrey, "Changing Views of the Phillips Curve," Federal Reserve Bank of Richmond, Monthly Review 59 (July 1973): 2-13.
See, for instance, Milton Friedman, "The Role of Monetary Policy," American Economic Review 58 (May 1968): 1-17.
A clear, early presentation of this theme can be found in Friedrich A. Hayek, Prices and Production, 2d ed. (London: Routledge and Kegan Paul, 1935). A general exposition is contained in Gottfried Haberler, Prosperity and Depression, 5th ed. (London: Allen & Unwin, 1964), pp. 29-72. A recent, formal development of this possibility occurs in Robert E. Lucas, Jr., "An Equilibrium Model of the Business Cycle," Journal of Political Economy 83 (December 1975): 1135-1137.
Many of the strictures against end-state norms of justice raised by Robert Nozick might readily be translated into strictures against the policy-oriented discussion of "full employment" in the post-Keynesian period. See Robert Nozick, Anarchy, State, and Utopia (New York: Basic Books, 1974).
See the careful and thorough documentation of this point in Milton Friedman and Anna J. Schwartz, A Monetary History of the United States, 1867-1960 (Princeton: Princeton University Press, 1963), pp. 299-419.
End of Notes
Return to top