Chapter 12

Unemployment and Inflation

I. Unemployment and Inflation: Is There a Trade-Off?

just as in Chapters 14 (Monetary Policy) and 15 (Fiscal Policy),

this chapter continues our analysis of how macroeconomic
policy works and how it can best be used

in Section 12.1, we study unemployment and inflation together

these two variables are sometimes referred to as the “twin

evils”—they represent the most important macroeconomic
problems

studying unemployment and inflation together gives us the

Phillips curve relationship

after that, we separately examine unemployment (Section 12.2)

and inflation (Section 12.3)

A. The Original Phillips Curve

conventional wisdom leads the public to believe that there is a

trade-off between inflation and unemployment

1
1. The “Phillips curve” is the relation derived by A.W. Phillips
in 1958 that shows the negative (inverse) relationship between
unemployment and inflation

technically, though, this is incorrect

Phillips’s actual study examined the rate of nominal wage

growth in British industries

later economists changed the variable “nominal wage growth” to

“inflation”

2. Phillips’s data showed that many countries in the 1950s and

1960s (including the U.S.) seemed to have this negative
relationship between the two variables

3. This relationship suggested that policymakers faced a menu of

options; they could choose the combination of unemployment
and inflation they most desired

“you can pick where you want to be on the curve”

2
4. The original relationship fell apart in the next three decades

in the 1970s, there was both high inflation and high

unemployment

this situation is called “stagflation”, and it is totally inconsistent

with the Phillips curve

B. The Expectations-Augmented Phillips Curve

1. The original relationship broke down so quickly because the

public’s way of forming expectations changed

in the original model, there is an implicit assumption that πe = 0

but once inflation started to become persistently positive, the

public changed its expectations

2. Even before the original Phillips curve began to break down

in the 1970s, a new theory argued that the cyclical
unemployment rate (u - u) depends only on unanticipated
inflation (π – πe)

this idea was published by Friedman and Phelps

NOTE: The terms “Friedman-Phelps Phillips curve”,

“expectations-augmented Phillips curve”, and “modified Phillips
curve” all mean the same thing!

3
this implies that the relationship between inflation and the
unemployment rate is not stable; it can fluctuate

so now we talk about the relationship between the rate of

unemployment and the change in the inflation rate

3. The expectations-augmented Phillips curve is written as

π = π e
− h (u − u )

where h is the slope of the curve and (u − u) is cyclical

unemployment

in words: the change in the rate of inflation depends on the

difference between the actual unemployment rate and the natural
rate of unemployment

the “natural rate of unemployment” is the rate which maintains a

constant rate of inflation

this means Δ π = 0, NOT π = 0

so we can infer the following relationships:

(1) when π = π e
→u= u
this means we are at a point on the curve

(2) when π < π e

→u> u
4
this means the economy is in a recession

(3) when π > π e

→u< u
this means the economy is in an expansion

C. Shifting the Phillips curve

1. The Phillips curve shows the relationship between

unemployment (u) and inflation (π) for a given expected rate of
inflation (πe) and natural rate of unemployment ( u )

2. If you ↑ πe or ↑ u , this shifts the Phillips curve up and to the

right

5
3. Supply Shocks (SRAS) and the Phillips curve

adverse supply shocks increase both πe and u , so the Phillips

curve will shift up and to the right

this is partly why the original relationship started to fall apart in

the 1970s

there were the oil crises of 1973/74 and 1979/80, and the
Phillips curve will always be unstable in periods with many
supply shocks

plus, demographic changes were increasing the U.S. natural rate

of unemployment

plus, monetary policy was very expansionary and this led to

high and volatile inflation

D. Macroeconomic Policy and the Phillips curve

1. Can the Phillips curve be exploited? Is it really a policy

menu? Can politicians choose the desired combination of
unemployment and inflation?

(1) classicals say “no” because if people are rational, they will
change their expectations in response to political actions

6
as the public’s expectations adjust, the unemployment rate
returns to its natural level quickly

so long as the public is characterized by forward-looking

“rational expectations” (in contrast to backward-looking
“adaptive expectations”), unemployment can change from its
natural level only for a very brief time

why? because people will catch on to policy games and they will
anticipate policy changes; they cannot be systematically fooled

(2) Keynesians say “yes, temporarily”

if prices, wages, and expected future prices are sticky, then the
actual unemployment rate may differ from the natural rate for
some time

2. The Lucas Critique

you cannot expect historical relationships between variables to

continue to hold if there has been a major policy change

when the rules of the game change, behavior changes

the Phillips curve fell apart as soon as policymakers tried to

exploit it

Lucas argued that both economic theory and empirical analysis

are necessary to evaluate policy changes

E. The Long-Run Phillips curve

7
1. In the long-run both Keynesians and classical agree that u =
u
2. Thus, the long-run Phillips curve is vertical, since when π =
π e, u = u

3. Classicals and Keynesians also agree that money (and the

growth rate of the nominal money supply) is neutral in the long-
run

so, even though expansionary monetary policy may reduce

unemployment temporarily (think about what’s happening to the
LM curve!), policymakers may want to do this if they can create
an economic boom right before elections

II. Unemployment

A. The Costs of Unemployment

1. Loss in output from idle resources

8
workers lose income

society has to pay for unemployment benefits and the

government loses tax revenue

2. Okun’s Law (from Chapter 3, Section 3.6)

Arthur Okun, chair of the Council of Economic Advisors during

the 1960s, posited a relationship between output (relative to full-
employment output) and cyclical unemployment

Okun’s Law states that each percentage point of cyclical

unemployment is associated with a loss equal to 2 percent of
full-employment output

algebraically, the formula is

( Y − Y) / Y = 2 (u − u )

the coefficient in Okun’s Law is 2 (and not just 1) because

several other things happen when cyclical unemployment rises

for example: the labor force falls, hours of work per worker
decline, and the average productivity of labor declines

if that’s true, then the result is a 2 percent reduction in output for

every 1 percentage point increase in the actual unemployment
rate

9
a possible numerical example might be if full-employment
output is $ 7.5 trillion, then each percentage point of
unemployment sustained for one year would cost the economy $
150 billion

3. Personal and Psychological Cost

these costs affect not just workers, but also their families

this is especially true for those who experience long spells of

unemployment

4. Offsetting Benefits

(1) unemployment leads to increased job search and learning

new skills

hopefully, this frictional unemployment will lead to increased

future output

(2) unemployed workers have increased leisure time

but is this really worth being out of work?

B. The long-term behavior of the unemployment rate

1. The changing natural rate

from the 1950s – 1960s, and from the 1990s – 2008: around 5 –
5½%

10
in the 1970s and 1980s: over 6 %

it has been falling due to demographic forces; the proportion of

the labor force aged 16 – 24 years old fell from 25 % in 1980 to
16 % in 1998

some economists think the natural rate of unemployment is 4.5

% or even lower

the natural rate has also been falling because of ↑ efficiency in

the Labor Market and better matching of workers and jobs (this
reduces frictional and structural unemployment)

2. Measuring the natural rate of unemployment

economists disagree about how to measure the natural rate of

unemployment

because of this, the Congressional Budget Office (CBO) has

revised its measure of the natural rate several times

economists have tried to calculate the natural rate, but it cannot

be measured precisely with current econometric methods (the
confidence interval is too large)

policymakers need a measure of the natural rate of

unemployment to use the unemployment rate for setting policy

but since we don’t know what the natural rate is, that implies
policy decisions are probably wrong

11
or, at the very least, U.S. policy is based on the wrong
information

III. Inflation

A. The Costs of Inflation

1. Perfectly-Anticipated Inflation

this is one extreme

if all prices and wages (and therefore, incomes) keep up with

inflation, then there are no adverse effects

but even under the best conditions, technically we would still

have two costs to worry about

(1) “shoe-leather costs”: ↑ π → ↑ i → a higher opportunity cost

of holding currency

this is often called the “inflation tax”

as a result, people ↓ money balances by making more trips, more

often, to the bank (and they wear out their shoes)

if inflation were lower, then people could be doing other things

instead (like working more or enjoying more leisure)

(2) “menu costs”: when prices rise because of inflation, there

are costs involved in changing prices

12
technology may mitigate this somewhat

2. Unanticipated Inflation (π − π e)

the main cost here is that actual realized real returns on assets
differ from their expected real returns

the actual “r” differs from the expected “r” by π e

–π

this affects investments, wages, and salaries

it is basically a transfer of wealth

(1) from lenders to borrowers when π > π e

(2) from borrowers to lenders when π < π e

NOTE: To the economy as a whole, on net, there is no change

in wealth. The effect is neutral.

but since people want to avoid the risks of unanticipated

inflation, a second cost is the resources spent on forecasting
inflation

to avoid these costs, some investments and government

contracts are “indexed”; meaning they are linked with the level
of inflation

examples of indexed contracts are adjustable-rate mortgages,

inflation-indexed bonds, and COLAs

13
3. Hyperinflation

hyperinflation is a very high, sustained inflation (for example,

50 percent or more per month)

Hungary (August, 1945) had inflation of 19,800 % per month

but then, between August 1945 and July 1946, in Hungary the
purchasing power of a unit of money fell by a factor of 4
octillion!

(that’s 4,000,000,000,000,000,000,000,000,000)

Bolivia had annual rates of inflation of 1281 % (1984), 11,750

% (1985), and 276 % (1986)

a major cost of hyperinflation is shoe-leather costs, as people

minimize cash balances

people spend many resources getting rid of money as fast as

possible

tax collections fall, as people pay taxes with money whose value
has declined sharply

even worse, the invisible hand falls apart; prices become

worthless signals

this causes markets to become inefficient

14
B. Fighting Inflation: The role of Inflationary Expectations

1. Disinflation: a reduction in the rate of inflation

be careful, “disinflation” is not the same as “deflation” (π < 0)

disinflations usually lead to recessions

problem: an unexpected ↓ π leads to ↑ u along the Phillips curve

remember that when π < π e, u > u in the expectations-

augmented Phillips curve

2. The costs of disinflation (specifically, ↑ u) could be reduced if

π e fell at the same time actual π fell

but somehow we have to convince the public to decrease their

expectations of future inflation

3. The Sacrifice Ratio

if the public does not expect (or believe) that tight monetary and
fiscal policies will actually be used to reduce inflation, then
there will be a reduction in output and employment

these costs must be considered against the benefits of lower

inflation

the variable that measures these societal costs is the sacrifice

ratio
15
defined as the number of percentage points of output lost in
reducing inflation by one percentage point

Example: In the U.S. in the early 1980s, π ↓ by 8.83 points and

output ↓ by 16.18 % of GDP. Thus, [16.18 / 8.83] = 1.832.

Interpretation: For every percentage point we ↓ π , the country

lost 1.832 % of a year’s potential GDP

NOTE: The sacrifice ratio can also be interpreted as how much

cyclical unemployment we accept to lower inflation. In this
definition, the sacrifice ratio is (1 / h)—the slope of the Phillips
curve.

4. Wage and Price Controls

price controls lead to shortages and inefficiency; once the

controls are lifted (if ever), prices will rise again

the outcome and efficacy of these controls depend on what

happens with fiscal and monetary policy (does the public
actually believe that the government is really trying to fix
things?)

5. Credibility and Reputation

the key determinant of the costs of disinflation is how quickly

π e adjusts

this depends on government credibility

16
if people believe the government—and if the government carries
through with its stated policies—then π e should drop rapidly

how does a central bank gain credibility?

(1) establish a reputation for following through on its promises,

even if it’s costly in the short-run

(2) make the central bank follow a rule that is enforced by some
outside agency

(3) have a strong and independent central bank that is committed

to low inflation

(4) appoint a “tough” or “hawkish” central banker

(5) change central bankers’ incentives (fire them if they do not

achieve the stated inflationary goals)

17