It is now over 50 years since Paul Samuelson and Robert Solow published their famous paper “Analytical Aspects of Anti-Inflation Policy,” now remembered mainly for offering the Phillips Curve as a menu of possible combinations of unemployment and inflation, reflecting a trade-off between inflation and unemployment. By accepting a bit more inflation, policy-makers could bring down the rate of unemployment, vice-versa. This view of the world enjoyed a brief heyday in the early 1960s, but, thanks to a succession of bad, and sometimes disastrous, policy choices, and more than a little bad luck, we seemed, by the late 1970s and early 1980s, to be stuck with the worst of both worlds: high inflation and high unemployment. In the meantime, Milton Friedman (and less famously Edmund Phelps) countered Samuelson and Solow with a reinterpretation of the Phillips Curve in which the trade-off between inflation and unemployment was only temporary, inflation bringing down unemployment only when it is unexpected. But once people begin to expect inflation, it is incorporated into wage demands, so that the stimulative effect of inflation wears off, unemployment reverting back to its “natural” level, determined by “real” forces. Except in the short run, monetary policy is useless as a means of reducing unemployment. That theoretical argument, combined with the unpleasant experience of the 1970s and early 1980s, combined with a fairly rapid fall in unemployment after inflation was reduced from 12 to 4% in the 1982 recession, created an enduring consensus that inflation is a bad thing and should not be resorted to as a method of reducing unemployment.
I generally accept the Friedman/Phelps argument (actually widely anticipated by others, including, among others, Mises, Hayek and Hawtrey, before it was made by Friedman and Phelps) though it is subject to many qualifying conditions, for example, workers acquire skills by working, so a temporary increase in employment can have residual positive effects by increasing the skill sets and employability of the work force, so that part of the increase in employment resulting from inflation may turn out to be permanent even after inflation is fully anticipated. But even if one accepts Friedman’s natural-rate hypothesis in its most categorical form, the Friedman argument does not imply that inflation is never an appropriate counter-cyclical tool. Indeed, the logic of Friedman’s argument, properly understood and applied, implies that inflation ought to be increased when the actual rate of unemployment exceeds the natural rate of unemployment.
But first let’s understand why Friedman’s argument implies that it is a bad bargain to reduce the unemployment rate temporarily by raising the rate of inflation. After all, one could ask, why not pocket a temporary increase in output and employment and accept a permanently higher rate of inflation? The cost of the higher rate of inflation is not zero, but it is not necessarily greater than the increased output and employment achieved in the transition. To this there could be two responses, one is that inflation produces distortions of its own that are not sustainable, so that once the inflation is expected, output and employment will not remain at old natural level, but will, at least temporarily, fall below the original level, so the increase in output and employment will be offset by a future decrease in output and employment. This is, in a very general sense, an Austrian type of argument about the distorting effects of inflation requiring some sort of correction before the economy can revert back to its equilibrium path even at a new higher rate of inflation, though it doesn’t have to be formulated in the familiar terms of Austrian business cycle theory.
But that is not the argument against inflation that Friedman made. His argument against inflation was that using inflation to increase output and employment does not really generate an increase in output, income and employment properly measured. The measured increase in output and employment is achieved only because individuals and businesses are misled into increasing output and employment by mistakenly accepting job offers at nominal wages that they would not have accepted had they realized that pries and general would be rising. Had they correctly foreseen the increase in prices and wages, workers would not have accepted job offers as quickly as they did, and if they had searched longer, they would have found that even better job offers were available. More workers are employed, but the increase in employment comes at the expense of mismatches between workers and the jobs that they have accepted. Since the apparent increase in output is illusory, there is little or no benefit from inflation to outweigh the costs of inflation. The implied policy prescription is therefore not to resort to inflation in the first place.
Even if we accept it as valid, this argument works only when the economy is starting from a position of full employment. But if output and employment are below their natural or potential levels, the argument doesn’t work. The reason the argument doesn’t work is that when an economy starts from a position of less than full employment, increases in output and employment are self-reinforcing and cumulative. There is a multiplier effect, because as the great Cambridge economist, Frederick Lavington put it so well, “the inactivity of all is the cause of the inactivity of each.” Thus, the social gain to increasing employment is greater than the private gain, so in a situation of less-than-full employment, tricking workers to accept employment turns out to be socially desirable, because by becoming employed they increase the prospects for others to become employed. When the rate of unemployment is above the natural level, a short-run increase in inflation generates an increase in output and employment that is permanent, and therefore greater than the cost associated with a temporary increase in inflation. As the unemployment rate drops toward the natural level, the optimal level of inflation drops, so there is no reason why the public should anticipate a permanent increase in the rate of inflation. When actual unemployment exceeds the natural rate, inflation, under a strict Friedmanian analysis, clearly pays its own way.
But we are now trapped in a monetary regime in which even a temporary increase in inflation above 2-percent apparently will not be tolerated even though it means perpetuating an unemployment rate of 8 percent that not so long ago would have been considered intolerable. What is utterly amazing is that the intellectual foundation for our new 2-percent-inflation-targeting regime is Friedman’s natural-rate hypothesis, and a straightforward application of Friedman’s hypothesis implies that the inflation rate should be increased whenever the actual unemployment rate exceeds the natural rate. What a holy mess.