Milton Friedman Prize Lecture

278

Economic Sciences 1976

rather than the separate effects of separate causes. The data are not inconsistent

with the stronger statement that, in all industrialized countries, higher rates

of inflation have some effects that, at least for a time, make for higher unem-

ployment. The rest of this paper is devoted to a preliminary exploration of

what some of these effects may be.

b. A 

tentative 

hypothesis

I conjecture that a modest elaboration of the natural-rate hypothesis is all

that is required to account for a positive relation between inflation and

unemployment, though of course such a positive relation may also occur for

other reasons. Just as the natural-rate hypothesis explains a negatively sloped

Phillips curve over short periods as a temporary phenomenon that will dis-

appear as economic agents adjust their expectations to reality, so a positively

sloped Phillips curve over somewhat longer periods may occur as a transitional

phenomenon that will disappear as economic agents adjust not only their

expectations but their institutional and political arrangements to a new

reality. When this is achieved, I believe that - as the  natural - rate hypothesis

suggests - the rate of unemployment will be largely independent of the average

rate of inflation, though the efficiency of utilization of resources may not be.

High inflation need not mean either abnormally high or abnormally low

unemployment. However, the institutional and political arrangements that

accompany it, either as relics of earlier history or as products of the inflation

itself, are likely to prove antithetical to the most productive use of employed

resources - a special case of the  distinction between the state of employment

and the productivity of an economy referred to earlier.

Experience in many Latin American countries that have adjusted to

chronically high inflation rates - experience that  has been analyzed most

perceptively by some of my colleagues, particularly Arnold Harberger and

Larry Sjaastad [(12), (25)] - is consistent,  I believe, with this view.

In the version of the natural-rate hypothesis summarized in Figure 2, the

vertical curve is for alternative rates of fully anticipated inflation. Whatever

that rate - be it negative, zero or positive - it can be  built into every decision

if it is fully anticipated. At an anticipated 20 percent per year inflation, for

example, long-term wage contracts would provide for a wage in each year that

would rise relative to the zero-inflation wage by just 20 percent per year;

long-term loans would bear an interest rate 20 percentage points higher than

the zero-inflation rate, or a principal that would be raised by 20 percent a year;

and so on - in short, the  equivalent of a full indexing of all contracts. The

high rate of inflation would have some real effects, by altering desired cash

balances, for example, but it need not alter the efficiency of labor markets, or

the length or terms of labor contracts, and hence, it need not change the natural

rate of unemployment.

This analysis implicitly supposes, first, that inflation is steady or at least

no more variable at a high rate than at a low - otherwise,  it is unlikely that

inflation would be as fully anticipated at high as at low rates of inflation;

second, that the inflation is, or can be, open, with all prices free to adjust to

---

M. Friedman

279

the higher rate, so that relative price adjustments are the same with a 20

percent inflation as with a zero inflation; third, really a variant of the second

point, that there are no obstacles to indexing of contracts.

Ultimately, if inflation at an average rate of 20 percent per year were to

prevail for many decades, these requirements could come fairly close to being

met, which is why I am inclined to retain the long-long-run vertical Phillips

curve. But when a country initially moves to higher rates of inflation, these

requirements will be systematically departed from. And such a transitional

period may well extend over decades.

Consider, in particular, the U.S. and the U.K. For two centuries before

World War II for the U.K., and a century and a half for the U.S., prices

varied about a roughly constant level, showing substantial increases in time of

war, then postwar declines to roughly prewar levels. The concept of a “normal”

price level was deeply imbedded in the financial and other institutions of the

two countries and in the habits and attitudes of their citizens.

In the immediate post-World War II period, prior experience was widely

expected to recur. The fact was postwar inflation superimposed on wartime

inflation; yet the expectation in both the U.S. and the U.K. was deflation.

It took a long time for the fear of postwar deflation to dissipate - if it still

has-and still longer before expectations started to adjust to the fundamental

change in the monetary system. That adjustment is still far from complete

[Klein (16)].

Indeed, we do not know what a complete adjustment will consist of. We

cannot know now whether the industrialized countries will return to the pre-

World War II pattern of a long-term stable price level, or will move toward

the Latin American pattern of chronically high inflation rates - with every

now and then an acute outbreak of super- or hyperinflation, as occurred

recently in Chile and Argentina [Harberger (11)] -or will undergo more

radical economic and political change leading to a still different resolution of

the present ambiguous situation.

This uncertainty - or more precisely,

 

the circumstances producing this

uncertainty - leads to systematic  departures from the conditions required for

a vertical Phillips curve.

The most fundamental departure is that a high inflation rate is not likely to

be steady during the transition decades. Rather, the higher the rate, the more

variable it is likely to be. That has been empirically true of differences among

countries in the past several decades [Jaffe and Kleiman (14); Logue and Wil-

lett (17)]. It is also highly plausible on theoretical grounds - both about

actual inflation and, even more clearly, the anticipations of economic agents

with respect to inflation. Governments have not produced high inflation as a

deliberate announced policy but as a consequence of other policies - in

particular, policies of full employment and welfare state policies raising

government spending. They all proclaim their adherence to the goal of stable

prices. They do so in response to their constituents, who may welcome many

of the side effects of inflation, but are still wedded to the concept of stable money.

A burst of inflation produces strong pressure to counter it. Policy goes from one

---