Lecture Thirteen
Inflation, Unemployment, and the Phillips Curve
(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 21, 1996
Models of Aggregate Supply
What Do We Think of These Models?
From Aggregate Supply to the Phillips Curve
Disinflation and the "Sacrifice Ratio"
Anti-Recession Policy, Cyclical Unemployment, and Structural
Unemployment
Models of Aggregate Supply
The goal is, ultimately, to get to an aggregate supply equation of
the form:
(1) Y = Y* + a(P - E(P))
Where "E" denotes "expectation"--that when the actual price level P
is below what people had expected it to be, output (and employment)
is depressed.
Now why do shocks to spending show up as shifts in prices and
rather than as shifts in prices alone? The first part of Greg
Mankiw's chapter 11 gives four possible answers:
Sticky Wages
The first is that nominal wages are sticky. Adjust with a
substantial lag. Typical labor contract has wages fixed (for a
while), and employment variable (employment at will).
Why structure labor contracts this way? Why not make employment
status fixed and wages variable? Incentive compatibility: employer
loses something if he or she fires workers for no good reason.
Employer loses nothing if reduces wages of workers for no good
reason. Hence employers' self-interest cannot be relied on to check
desire to reduce wages.
Once you have a sticky nominal wage, then firm profit maximization
says that the higher the price level, the higher employment and
output--and we have arrived at equation (1).
Worker Misperceptions
Workers do not really "know" the general price level. They estimate
their real wage by dividing their nominal wage by what the
think the price level is. A positive boost to prices and wages
raises workers' perceived (although not their actual) real wage;
increases the labor force; raises employment.
Imperfect Information
Everyone confuses changes in relative prices with changes in the
absolute price level. If relative prices change, the right response
is to increase or decrease production. This confusion leads
production and employment to be correlated with unanticipated shifts
in the price level.
Sticky Prices/Small Menu Costs
Changing prices is somewhat expensive, and is done relatively rarely.
Firms want to keep their customers happy by not changing their prices
too much (or too far), and by not having to tell customers that their
goods are out of stock.
Staggering of Wages and Prices
Aggregate Demand Externalities (under monopolistic
competition)/Recessions as coordination failures
What Do We Think of These Models?
Well, they are all at work out there in the real world. Prices
are sticky; information about which shifts in demand are
shifts in relative demand and which are shifts in aggregate demand
is imperfect; workers do misperceive the general price
level; wages are sticky and employment--bodies and hours--is
variable.
A mistake to think that there is the model of aggregate supply
which is the truth. Complicated world. Lots of things going on...
Nevertheless, note that worker misperceptions, imperfect
information, and sticky prices all imply that people are--or
should be--unhappy with booms. Worker misperceptions imply
that, after a prolonged boom, people are unhappy because they look
back at the real wages they actually received and think "they weren't
worth the extra hours (or the extra jobs)". Imperfect
information implies that businesses look back after a boom and
say "we shouldn't have produced all that stuff". Sticky prices
implies that businesses during a boom are saying "we shouldn't
be producing so much; our marginal costs are too high".
Even the sticky wages model is uneasy with the implication
that people are happy in a boom. Aggregate demand
externalities and recessions as coordination failures are
the only possible models that conform to reality...
Greg calls this an "active area of research"; and it is the area
where he spends most of his time (and also where economists' models
are perhaps least satisfactory).
From Aggregate Supply to the Phillips Curve
Recall equation (1):
(1) Y = Y* + a(P - E(P))
and let's divide everything by a and move the price level P to
the left-hand side:
(2) P = E(P) + (1/a)(Y - Y*)
Now subtract last year's price level from both sides:
(3) P - P-1 = E(P) - P-1 + (1/a)(Y -
Y*)
and notice that:
(4) Inflation = Expected Inflation + (1/a)(Y - Y*)
or, from Okun's Law:
(5) Inflation = Expected Inflation - B(u - u*) + e
Supply shocks; oil price increases; price controls and their
relaxation.
1961-1970: Expectations of inflation low; as unemployment
falls from 1961-1970, the inflation rate rises as we produce a "high
pressure economy"
1971-1975: Expectations of inflation rising rapidly as people
note the loss of the economy's commitment to stable prices; oil
shocks; outward shift of the "Beveridge Curve" as labor market seems
to work less well...
1976-1980: Expectations of inflation more-or-less stable at
7-8% per year or so; as unemployment falls, inflation rises;
1961-1970, but with less favorable inflation expectations.
1981-1984: Another supply shock; fear that expectations are
moving in a very unfavorable direction; Volcker disinflation--not
only raise unemployment high enough to push inflation down, but keep
unemployment high enough long enough to convince people that
inflation will remain low.
1985-1995: Expectations of inflation low (but a bit higher
than in 1960s). As unemployment falls and rises, inflation rises and
falls; some sign of further adjustment of expectations/inward
movement of Beveridge Curve in recent years.
Disinflation and the "Sacrifice Ratio"
How to influence expected inflation?
- Jawboning
- Incomes policies/price controls
- Show me (deep recessions)
Clearly the first two are to be preferred...
Clearly people are likely to be suspicious of the first two...
Importance of "credibility": Thatcher; Reagan...
Anti-Recession Policy, Cyclical Unemployment, and Structural
Unemployment
>
