Lecture Fourteen
The Short-Run Inflation-Unemployment Tradeoff, and
the Sacrifice Ratio
(Economics 100b; Spring 1996)
Professor of Economics J. Bradford DeLong
601 Evans, University of California at Berkeley
Berkeley, CA 94720
(510) 643-4027 phone (510) 642-6615 fax
delong@econ.berkeley.edu
http://www.j-bradford-delong.net
February 23, 1996
Administration
Accelerating Inflation
The Sacrifice Ratio
"Hysteresis"
Disinflation and the "Sacrifice Ratio"
Anti-Recession Policy, Cyclical Unemployment, and Structural
Unemployment
Administration
Make-up exam to be given on March 15, 2-3 PM
Talk about lecture schedule; next year how much effort should be
devoted to making this a TTh (or a MW) course?
Sidelight
New Federal Reserve nominees; Larry Meyer; Alice Rivlin
Accelerating Inflation
Consider the options that the Phillips curve gives to someone
trying to nudge the economy to full employment with monetary and
fiscal policy. At any single moment, expected inflation and supply
shocks are outside of control. Yet changing aggregate demand alters
output, unemployment, and inflation. Expanding aggregate demand is
close to guaranteed to produce higher output, lower unemployment, and
higher inflation Contracting aggregate demand is all but guaranteed
to raise unemployment and lower inflation.
Almost every time Alan Greenspan appears in front of Congress, he
says something like that this short-run tradeoff is "ephemeral and
unusable". Why? Because the past generation's experience strongly
suggests that any systematic attempt to run a high-pressure
economy--to have a little bit lower unemployment and a little bit
higher inflation--is doomed to failure, because expectations of
inflation will adjust upward. And when they do you have a less
favorable short-run tradeoff.
Thus in the long run, we are back where we were at the start of the
course: chapter 3. Output (and employment) are determined by
"fundamentals": productivity and factor supplies, if we are in a
situation in which all expectations are satisfied.
How long is the long run? Probably more than five years.
The Sacrifice Ratio
In the late 1970s the U.S. economy had an "expected" and an actual
inflation rate somewhere near 8 percent per year (pushed up,
temporarily, at the end of the end of the 1970s by supply shocks).
Suppose--for the sake of argument--that when Paul Volcker was named
to the Federal Resrve, that unemployment was at its "natural rate",
that Volcker sought to reduce inflation and expected inflation from
around 8 percent per year to around 3 percent, and that he had asked
you to tell him what was going to happen as he pursued this
policy.
How do we think about this?
We have the result of our IS-LM apparatus:
(1) Y = Y(r)
If we think of the Federal Reserve as controlling interest rates. We
have our short-run Phillips curve:
(2) p = E(p) + (1/a)(Y-Y*)
But this is not enough; we need to know how expectations of inflation
adjust over time.
One possibility is that people are pretty sophisticated in forming
expectations of inflation. They take a look at the economy as a
whole--including a look at the monetary and fiscal policies in
effect--try to estimate how those policies are going to affect the
economy, and so form their expectations of inflation.
This possibility would suggest that the location of the short-term
Phillips curve is easy to shift--for good or for ill--that if
a central bank or a government tries to expand employment and keep it
above "natural rates" the Phillips curve will shift upward rapidly;
conversely, the mere announcement of a change in policy could
shift the short-run Phillips curve: all you have to do is say "we are
going to be tough on inflation; and you find the Phillips curve
shifting).
There is something to this:
- What Alan Greenspan did to the bond market on Tuesday
- In Mexico the inflation rate has gone from 1/2% per month to
2-3% per month over the past year and a bit
But there is another approach--one that says that this "rational
expectations" vision is hopelessly naive:
People don't trust announcements of policies
People don't trust economic models
Hence, "show me"
This approach implies that reducing inflation will be much more
painful; the only way to push inflation down is to produce enough
unemployment that you can show that inflation has fallen.
How to influence expected inflation?
- Jawboning
- Incomes policies/price controls
- Show me (deep recessions)
Clearly the first two are to be preferred, if they work...
Clearly people are likely to be suspicious of the first two; naivete
of the first; perversity of the second (people react to incomes
policies by raising their expectations of what will happen after the
end of the program): Nixon: these policies will do a lot less damage
as implemented by us, who don't believe in them, than if implemented
by true believers.
Importance of "credibility": Thatcher and the MTFS; Reagan and the
hope for painless disinflation in the early 1980s.
So: once again: Suppose--for the sake of argument--that when Paul
Volcker was named to the Federal Resrve, that unemployment was at its
"natural rate", that Volcker sought to reduce inflation and expected
inflation from around 8 percent per year to around 3 percent, and
that he had asked you to tell him what was going to happen as he
pursued this policy.
Suppose that expected inflation equals last year's inflation:
(3) pt = pt-1 + (1/a)(Yt -
Y*t)
and suppose that you want to accomplish the reduction in inflation
over a period of five years. Then:
(4) p5 = p4 + (1/a)(Y5 -
Y*5)
(5) p5 = p3 + (1/a)(Y5 -
Y*5) + (1/a)(Y4 - Y*4)
Keep substituting in; and eventually we have:
(6) p5 - p0 = (1/a)(Y5 -
Y*5) + (1/a)(Y4 - Y*4) +
(1/a)(Y3 - Y*3) + (1/a)(Y2 -
Y*2) + (1/a)(Y1 - Y*1)
or
Lost Production (relative to potential output Y*) =
a(p5 - p0)
Where a is the inverse slope of the Phillips curve.
Use Okun's Law to express this in terms of unemployment...
Between 1981 and 1985, 9.5 percentage point-years of cyclical
unemployment; inflation down by five percentage points. "Sacrifice
ratio" of 1.9 for unemployment (or 3.8 for output). Reducing
inflation thus very expensive....
Tempted to leave inflation bouncing around in the 5-10 percent range.
But (a) voters hated inflation; (b) fear that permanent
moderate inflation is a contradiction in terms...
"Hysteresis"
An extremely ugly word. Comes from study of magnetism. Attempt by
Larry Summers and Olivier Blanchard to create an intellectual
movement to study peculiar problems of European unemployment.
Brief history of European unemployment; unemployment rises a lot in
recessions; it does not fall much in booms
Natural rate hypothesis...
But recessions can shift the natural rate of unemployment as well;
transformation of cyclical into structural unemployment:
- atrophy of human capital
- labor force attachment
- insiders/outsiders
- endogenous policies
"Hysteresis" raises the sacrifice ratio: makes it infinite, in
fact; most of the attraction of disinflation is that you incur
several bad years for something--lower inflation--of permanent
benefit. But in Europe over the past generation this has not been
true.
The Macroeconomic Policy Debate
Mankiw, chapter 12.
Lecture 14 Equations
(1) Y = Y(r)
(2) p = E(p) + (1/a)(Y-Y*)
(3) pt = pt-1 + (1/a)(Yt -
Y*t)
(4) p5 = p4 + (1/a)(Y5 -
Y*5)
(5) p5 = p3 + (1/a)(Y5 -
Y*5) + (1/a)(Y4 - Y*4)
(6) p5 - p0 = (1/a)(Y5 -
Y*5) + (1/a)(Y4 - Y*4) +
(1/a)(Y3 - Y*3) + (1/a)(Y2 -
Y*2) + (1/a)(Y1 - Y*1)
Lost Production (relative to potential output Y*) = a(p5 -
p0)
>
