^^^«%.
^'
Digitized by the Internet Archive
in
2011 with funding from
Boston Library Consortium IVIember Libraries
http://www.archive.org/details/observationsconjOOkrue
',tm&J
working paper
department
of economics
OBSERVATIONS AND CONJECTURES ON THE
U.S. EMPLOYMENT MIRACLE
Alan Krueger
Jorn-Steffen Pischke
No.
97-16
August 1997
,
massachusetts
institute of
technology
50 memorial drive
Cambridge, mass. 02139
WORKING PAPER
DEPARTMENT
OF ECONOMICS
OBSERVATIONS AND CONJECTURES ON THE
U.S. EMPLOYMENT MIRACLE
Alan Knieger
Jorn-Steffen Pischke
No.
97-16
August, 1997
MASSACHUSEHS
INSTITUTE OF
TECHNOLOGY
50 MEMORIAL DRIVE
CAMBRIDGE, MASS. 021329
MASSACHUSETTS INSTITUTE
"'OGY
4])
LlBMMMItS
JAN 19 1998
—
Observations and Conjectures
on the U.S. Employment Miracle
Alan B. Krueger
Jorn-Steffen Pischke
Abstract
This paper has three goals;
first
place U.S. job growth
to
perspective by exploring cross-country differences
growth. This section finds that the U.S. has
especially
female workers
countries.
The
market has
markets are
rigidities
into
employment
flexible
rigid.
wages and employment
The second goal
of the
hypothesis. Although greater
employment
in
in
many European
from relative
wage
at
wage
to
absorb added worl<ers
a greater rate than most
phenomenon
practices,
paper
is
its
is
whereas European labor
inflexibility
population, the slow growth
absorb so
skill
in
groups
alone. Furthermore, a great deal of labor
at
a constant
real
wage
leads to the third goal: to speculate on other explanations
to successfully
markets
probably contributes to the
countries appears too uniform across
market adjustment seems to take place
managed
that the U.S. labor
to evaluate the labor
flexibility
creating jobs for
international
employment and population
in
managed
leading explanation for this
U.S.'s comparative success
to result
—
in
many new
in
the U.S. This
why
the U.S. has
entrants to the labor market.
We
conjecture that product market constraints contribute to the slow growth of
employment
in
many
countries.
Observations and Conjectures on the U.S. Employment Miracle
Alan B. Krueger
Princeton University and
NBER
and
Jorn-Steffen Pischke
Massachusetts
Institute of
Technology and
NBER
June 4, 1997
August 14, 1997
First Draft:
This Draft:
*We
thank Roope Uusitalo for excellent research assistance, and Lars Calmfors, David
Card, John Dunlop, Jim Heckman, Bertil Holmlund, Joachim Moller, Sherwin Rosen, and
Cecilia Rouse for helpful comments. Naturally, the authors are responsible for all views
expressed. This paper expands the analysis in an earlier paper by Alan Krueger entitled
"The U.S. Employment Miracle in Perspective." The data from the "Qualifications and
Careers" survey was obtained from the German Zentralarchiv fur Empirische
Sozialforschung at the University of Koln (ZA).
The data were collected by the
Bundesinstitut fur Berufsbildung and the Institut fur Arbeitsmarkt und Berufsforschung.
Neither the producers of the data nor the ZA bear any responsibility for the analysis or
Interpretation of the data in this paper.
Compared
to
most
of
Westem
Between 1985 and 1995,
impressive.
U.S. and by an average of just
growth
Europe, employment growth
5.7%
for
in
Germany, France and the
employment growth has caused the unemployment
April
in
it
1997
much
is
of
for the first time since 1973.
employment
A
why
the
Strong
U.S.
fall
Australia. With these statistics
the recent U.S. experience
below
5%
in
8%
as background,
frequently characterized as an
is
miracle.
employment growth has been
remarkably steady since the Great Depression, growing by
intervals since 1940,
and by more than 15%
at least
in all
10%
in
each
of the
but five of these intervals.
U.S. employment growth has been aided by comparatively fast population growth.
example, between 1985 and 1995 the working-age population
11.4%, while
the
Employment
U.K.''
rate in the U.S. to
longer term perspective reveals that U.S.
58 ten-year
in
in
Meanwhile, the unemployment rate exceeds
Western Europe, Canada and
not hard to see
the U.S. has been
example, employment grew by 16.6%
the private sector has been especially impressive
in
in
it
increased by an average of 6.8%
in
in
For
the U.S. increased by
Germany, France and the U.K.
combined.
This paper has three goals.
international perspective.
is
A
First,
to
put the
Nonetheless, the U.S. has
managed
to
employment growth
in
employment growth
in
the U.S.
working-age population
in
the U.S.
large share of the stronger
attributable to comparatively faster growth of the
U.S.
absorb added workers ~ especially female
^These figures, and subsequent employment figures in the paper, are based on BLS
data which put foreign countries on a similar basis as the U.S. OECD employment
figures are generally similar. One major difference, however, is that the BLS data pertain
to
West Germany
after unification.
3
workers
--
into
employment
economy created so many
that the U.S. labor
market
is
flexibility
jobs for our growing population?
market
probably contributes to the U.S.'s comparative success
of the labor market.
some
facts that challenge the
skill
groups
to result
wage
in
explanations
the U.S.
why
from relative
seems
Furthermore, a great deal of labor market adjustment
managed
to successfully
of the
wage
its
of this feature
many European
wage
inflexibility.
to take place at
This leads to the third goal of the paper:
the U.S. has
in
is
flexibility
creating jobs for
in
employment
in
The
greater
paramount importance
Most importantly, the slow growth
countries appears too uniform across
real
Although
hypothesis.
rigidities
rigid.
The second goal
wages.
well as relative
has the U.S.
The leading explanation
and European labor markets are
wages as
to evaluate the labor
population, there are
How
a greater rate than most countries.
is flexible,
applies to the level of real
paper
at
a constant
To speculate on
other
absorb so many new entrants
to the labor market.
1.
An Analysis
The
of
Employment Growth
analysis
makes use
in
of the
10 Countries
Bureau
of
Labor
(BLS) international
Statistics's
estimates of employment and population. These data use U.S. concepts of employment
and working-age population
for
country estimates reported to the
employment concepts.
schooling age
age
limit,
in
10 major countries.
OECD
for
breaks
in
The minimum working age
the country (14-16).
so adjustments were
made
The
U.S.
Briefly,
series
is
the
and
BLS
for
adjusts
official
major deviations
typically set at the
in
compulsory
employment data do not have an upper
to the data for countries
which limited employment
4
to
workers below a certain age. Population pertains
than the minimum working age.
Table
1
long-term trends
for
more
presents regression estimates of the log of
Over
term and individual country dummies.
exceeded the
Germany's
log of the
rest of the
working-age population
model the U.S. trend term
total
1
the
falls to
added
first
Germany accounts
The models
same
in all
to the regression
.40 percentage points
country
in
much, but not
column
at different rates.
employment.
variables,
1
all,
of
its
In
column
model. Interestingly,
above the world
(2),
not be the case.
column
to
(3),
2,
the
in this
trend; thus,
for
by faster
slower population growth
effect of population
in
growth to be the
For example, various economic and
different countries to translate population
In
year
lower trend-growth of employment.
and 2 constrain the
may
This
may cause
dummy
According to column
the U.S.^
for
countries.
social rigidities
growth
in
linear
of
percentage point per year, while
about 60 percent of the faster U.S. employment growth can be accounted
population growth
column
U.S. employment growth
sample by about 0.7 points per year.
is
10
for
international data.)
employment on a
this long period,
sample by approximately
trailed the rest of the
BLS
German employment,
U.S. and
in
details of the
older
is
The data are annual, from 1959 through 1995
major countries.^ (See Godbout, 1993
To estimate
to the civilian population that
the population variable
is
growth
into job
interacted with 10
allow for differential effects of population growth on
The countries can be grouped
into three
groups: countries where a
^For Australia, comparable employment data are only available for 1964-95, and for
the Netherlands they are only available for 1973-95.
^Godbout (1993) also finds that faster population growth
stronger employment growth in the U.S.
is
a major reason for the
5
proportionate increase
in
employment
(U.S.,
in
population
is
associated with more
tinan
a proportionate increase
Canada, the Netherlands); countries where employment growth
is
proportionate to population growth (Australia, Japan, and Sweden); and countries where
employment growth
Italy
and the
Figure
is
less than proportionate with population growth (France,
U.K.).
1
graphically illustrates the relationship
population growth, over five-year intervals.
that population growth accounts for
rises
A
somewhat more than
is
between employment growth and
regression through these points indicates
56 percent
Moreover, the slope of the relationship
employment
of the variability in
Thus,
1.27.
same
The
men and women
set of regression
models
across countries.
for
U.S. superior trend growth
employment growth.
over five-year intervals,
proportionately with population growth.
The aggregate employment data mask important
treatment of
Germany,
differences
Columns 4-6
of
in
the labor market
Table
1
men, and columns 7-9 present estimates
in
employment
is
women
greater for
present the
for
women.''
than for men.
Moreover, columns 6 and 9 indicate that the relationship between population growth and
employment growth
growth
in
varies by gender within several countries.
female population hardly translates
the coefficient on population
International differences
in
is
into greater
In
Japan and Germany,
employment.
over three times as great for
In
the Netherlands,
women
as
for
men.^
the relationship between population growth and employment
"These regressions are based on unpublished data provided by the BLS.
Connie Sorrentino for providing these data.
We
thank
^The Netherlands has an extremely high rate of part-time employment among women,
which may explain the high coefficient for women (see Godbout, 1993).
6
growth are much greater
We
for
women
OECD
have also used
estimate the aggregate models
estimate of the model
coefficient
on
in
employment data
Table
in
column
log population that
restricted to the 10 countries
used
than they are for men.
a
is
in
Interestingly, the
1.
OECD
with
(2)
little
Table
less than
1,
a wider set of countries to
for
data
for
employment
even
if
rate.
This
employment
is
rises at the
growth roughly
supply of workers creates
The
the
a country
to the estimates in
to work.
Because the
working
is (slightly)
--
in
of
of
Table
of
first
order approximation,
proportion to the workers'
--
but
it
does suggest
in
above the
expect
1
of
include
would be working, but well within a standard
OECD
^We
are grateful to Sherwin
to test
According
former East Germans
German Socioeconomic Panel (GSOEP), we
Germans worked
Table
47%
full
full-time
and 6.3% worked
and part-time workers, the
fraction of the marginal population that the
^The
that the
those workers.
we would
1,
of the
former East
figures
data are
because there may be unemployment
as population
many
for
OECD
to provide jobs for a marginal population increment.^
column 3
45.2%
rate
the
and West Germany provides a unique opportunity
Using the 1995 wave
calculate that
time.°
demand
re-unification of East
ability of
same
if
the coefficient rises to 1.16 with a standard
Says Law
not quite
24 countries produce a
However,
1.
error of .04.^ Thus, for the industrialized world as a whole, to a
countries tend to absorb population
unweighted regression
part-
fraction
model predicts
error of the point estimate.
For
men
the
data are for 1963-94.
Rosen
for
suggesting
this test to us.
®ln calculating these figures, we try to implement the
concepts that are used in the BLS international data.
same sample
restrictions
and
7
model predicts
it
predicts
that
73%
23%. The
46.1%, respectively
women,
II.
actual
--
but roughly
of the marginal population
employment
slightly
would be employed, and
rates for former East
lower than predicted for
for
women
Germans were 65.6% and
men and
higher than predicted for
the right ballpark.
in
Evaluation of Labor Market Riqidities:
Employment and Population Growth by
Skill
Level
The most common explanation
its
growing population
countries.
To some,
is
this
for the U.S.'s
that the U.S.
explanation
is
has a more
tautological.
there would be no involuntary unemployment.
account
is
for involuntary
not the source of
manifests
to
itself in
greater success
unemployment. An
If
flexible labor
wages are
is
providing jobs for
market than other
perfectly flexible, then
Thus, downwardly
alternative view
employment problems,
in
that labor
rigid
wages must
market
but inflexibility elsewhere
in
slower job growth. For example, restrictions on bringing
inflexibility
the
economy
new
products
market or on starting new enterprises could reduce job growth.
Another potential form
system.
A number
of
Unemployment Insurance
benefit availability
--
of labor
market
"rigidity"
involves the social insurance
microeconometric studies have found that more generous
(Ul) benefits
-- in
terms of both benefit levels and duration
of
are associated with longer durations of unemployment spells (see
Solon, 1985; Katz and Meyer, 1989; Meyer, 1990; Meyer, 1995; and Hunt, 1995).^
^Of course, this distortionary effect of Ul and other social insurance programs must
be weighed against their desirable features, such as consumption smoothing during times
of economic hardship.
8
These studies do
unemployment
not prove that generous
however, because non-UI recipients
new
(e.g.,
of the
more robust
findings
in
Forslund and Krueger, 1996).
More
(e.g.,
see Layard,
jobs
fill
connection
in this
cross-country studies
associated with higher unemployment
may
entrants)
those on the Ul program (see Levine, 1993). But
one
benefits reduce employment,
is
it
is
left
noteworthy that
that UI benefit generosity
is
and Jackman, 1991 and
Nickell
generally, social welfare benefits
workers' reservation wages, rendering a floor on market
available by
wages
similar
may
increase
effect to
in
a
minimum wage.^°
In
general, however,
main source
we
find that the
Europe and Canada's employment problems
of
conveyed by the widespread consensus
which
some people argue
job growth. But
just
evidence that labor market
it.
This
is
in
not to
support of
deny
that
by reforming
its
labor market
will
are the
not as strong as
is
and the stridency with
this view,
some
we suspect that any country that seeks
is
rigidities
labor market rigidities restrict
to solve
its
employment problems
be disappointed, because more seems
to
be
at
work.
In
its
downwardly
simplest form, the labor market
inflexible in
Europe because
of
rigidities
wage
wages, unions, insiders or welfare programs.
view
floors,
Wage
is
as follows:
Wages
established by legal
are
minimum
floors are less of a constraint in the
U.S. because unions represent only one-tenth of private sector workers, the real value
of the
minimum wage
is
quite low ($4.75 per hour
in
1997), and the safety net
is
weaker.
^°See Calmfors and Forslund (1991) for a theoretical model of how social welfare
programs can affect union wage setting, and related empirical evidence for Sweden.
9
The next
link
quite plausibly
--
experienced a decline
demand
in
demand shock has manifested
and lower employment
concluded
"that
in
enough
--
is
to
for less skilled
itself in
lower
assume
workers
wages
in
industrialized countries
the 1980s and 1990s. This
for less skilled
As Krugman (1994)
Europe.
all
workers
many
notes,
the U.S.
in
observers have
growing U.S. inequality and growing European unemployment are
different sides of the
Beyond the
same
coin."
This
is
a simple and appealing hypothesis.
unemployment
prediction of rising
in
Europe and growing inequality
in
the U.S., however, this explanation predicts that the decline
in
employment
would be disproportionately concentrated among the groups
of
workers whose wages
have
fallen
most
evidence on
in
the U.S.; that
this prediction
is,
among
in
an
rise in
unemployment
highly educated workers
rise
unemployment.
was
substantial,
Furthermore,
countries.
As evidence,
between 1971-82 and 1983-90 the
workers to that
rates
the
for high
Nickell
ratio of
in
(or rise in
in
far,
the
a number of studies
unemployment)
countries with
weak
is
overall
employment growth.
the
EC Europe
(with the exception of
and roughly proportional
U.S.
rate of unskilled relative to skilled
many European
of
Europe
papers, Nickell and Bell (1995, 1996) find that between the
Italy) for
unemployment
employment
countries with strong overall
in
influential set of
1970s and 1980s, the
in
in
the low-wage end of the labor market
employment growth than
In
and young workers. So
has been quite mixed. The findings
challenge the hypothesis that the decline
more concentrated
less-skilled
in
an
experienced
workers that was
and
Bell cite
educated workers rose from 3.8
to 4.2 in
increase
in
the
fairly similar to that of
OECD
unemployment
to the overall
data showing that
rate for low
educated
Germany, from
1.7 to
10
2.0
in
facts,
Spain, and from 3.9 to 4.7
however,
is
that the overall
ratio of the
unemployment
increase
the
in
number
a greater overall rise
unemployment
in
the
p.
307) conclude:
in
unemployment
rate rose less in the U.S.,
unemployed than
unemployment.
rate
difficulty in interpreting tliese
would
rates of unskilled to skilled workers
of unskilled
in
One
the United States.
in
In
skilled
still
unemployed
other words, whether
so a constant
imply a greater
in
countries with
we measure changes
Nonetheless, Nickell and Bell (1996,
levels or logs matters.
We
have analyzed the broad-brush hypothesis that European
unemployment has risen dramatically relative to that in the United States
because there has been a substantial shift in demand against the unskilled
in both places but relative wages are rigid in Europe and flexible in the
United States.
OECD
The
unemployment
workers
in
We
conclude that
sought to
find
a
this
hypothesis
relationship
work as
work
is
follows: "there
less prevalent
is little
solid
have achieved
among
youth and low-educated
Their results indicated statistically insignificant
and often wrong-signed correlations (see OECD, 1996;
this
inadequate.
between wage compression and
rates or employment-population rates
a cross-section of countries.
is
evidence
to
p. 75).
The
OECD
summarizes
suggest that countries where low-paid
this at the cost of
higher unemployment rates and
lower employment rates for the more vulnerable groups
in
the labour market, such as
youth and women."
Card, Kramarz and Lemieux (1996) examine trends
in
employment-to-population
11
rates
in
Canada and France,
the U.S.,
the population
find that
each country
in
low-wage groups
in
into
and the
rate
education-by-age
Between 1979 and 1989 they
cells.
the U.S. suffered large real earnings declines, as to be
expected. Additionally, they find a
employment
They group
using comparable micro data sets."
initial
weak
positive relationship
The
level of earnings.
between the change
decline
in
employment
in
the
rates for
lower-paid groups of American workers has been widely interpreted as a labor supply
response, but
we
note
it
possible that despite their steep decline
is
above the full-employment equilibrium
been a constraint as
while the results for
well.
level for
these workers so
For France, the results
employment are
growth across age-education
cells in
for
France
is
demand might have
wage growth
surprisingly similar.
In
unrelated to
wages remained
are quite different,
contrast to the U.S.,
initial
wage
levels,
a result of a labor market institutions that prevent wage adjustments. This
But changes
the
initial
in
the employment-to-population rate
wage
level
- employment
wage
rates
in
is
wage
probably
as expected.
France are only weakly related
to
have declined by about the same amount
regardless of the
initial
that rigid relative
wages have reduced employment
level of the cell.
This pattern
is
inconsistent with the view
among
especially
the low-skilled
in
France.
We
trends by
1
979 and
extend Card, Kramarz and Lemieux by comparing
skill
1
groups
in
Germany and
991 outgoing rotation group
the U.S.
files of
the
For the U.S.
wage and employment
we pooled
CPS. For each
period,
data from the
we
constructed
"See Edin, Harkman and Holmlund (1996) for related evidence for Sweden.
Especially for adults, their results are similar to Card, Kramarz and Lemieux's findings for
France.
12
age-by-education groups using three education levels (high school or less,
college or more)
60).
college,
and seven age groups (25-29, 30-34, 35-39, 40-44, 45-49, 50-54, 55-
For each group,
population rate (using
we
calculate the
CPS
mean
log hourly
wage and
the employment-to-
weights).
Unfortunately, no large data set similar to the
Germany.
some
CPS
is
We therefore constructed employment-to-population
"Mikrozensus" published annually by the
Statistical Office.''^
publicly available for
rates using data from the
This forces us to use the
three education groups available there, which correspond to the school-leaving degrees
of the various tracks of
secondary school. Classifications by post-secondary education
are also available, but the categories other than completed apprenticeship and university
degree are so small as
to
render them useless for our purposes.
secondary school degrees should work
well for this
predictors of final educational attainment,
skills,
Nevertheless, the
exercise since they are good
and wages.
We
picked the three
education groups for the U.S. so that they are of roughly comparable size to the
groups.
The German data are
the cell for the highest
employment
who have
rates
skill
available for the
group
and wages
in
For Germany,
five-year
age groups.
the age bracket 25-29, because
for this cell
yet to complete university
20 age-by-education
same
we were
German
We dropped
worried that
might be contaminated by including students
and be working
part time.
This provides a
total of
cells.
we
calculated the average log
wage
for
each
cell
using the 1979
^^These are taken from Statistisches Bundesamt, Fachserie 1: Bevolkerung und
Reihe 4.1.2: Beruf, Ausbildung, und Arbeitsbedingungen der
EnA/erbstatigen (Wiesbaden).
Enwerbstatigkeit,
13
and 1991
"Qualifications
sample (about 25,000 observations)
we
Therefore,
is
limited to
civilians
who
encompasses Germany's
still
mean
to avoid
large rise
any impact
Lemieux found
long log
The
figures reveal a pattern that
for the U.S.
wage growth and
and France.
the change
against the 1980-81 average log
wage
cells
wage
First,
in
is
The change
in
Using data
of unification, but this
German men
similar to
in
Figures 2 and
what Card, Kramarz and
the employment-population rate for each cell
Real wages declined
between 1979-80 and 1990-91, and grew
profile in the U.S.
wage. The sample
consider the U.S. Figure 2 plots decade-
for the cell.
consistent with the well-documented rise
a large
unemployment.
in
Results are presented graphically for American and
3, respectively.
log hourly
is
employed persons.
are employed outside of apprenticeships.
1970s and early 1990s allows us
for the late
period
that consists exclusively of
only use this data set to calculate the
German
The QaC
and Careers Survey" (QaC) surveys.
in
for the highest
the education
lowest
for the
ones.
This
is
premium and age-earnings
the employment rate, on the other hand,
is
only weakly
correlated with the base wage.
The
the
wage
results
on wages
structure did not
slight tightening of the
Institutionally rigid
and the
initial
relationship
for
in
Figure 3 are noisier, but
change very much between 1979 and 1991.
wage
structure
wages may
wage
German men
seemed
well explain the
level of the cell.
between the change
in
to
it
is
clear that
If
anything, a
have taken place during
this period.
weak
correlation
between wage growth
Turning to employment changes, there
is
no
the employment-to-population rate for the age-
education cells and the base wage, despite the fact that wages appear
rigid.
If
demand
14
fell
for less skilled
the lowest
wage
workers,
The summary regressions
in
change
in
wage
in
well as for
little
wage
the log
change
or
pattern of results evident
men. Indeed, contrary
wage
In
in
the
employment
the base period. Separate models were estimated for
in
to the
levels
level than
in
the
wage
rigidity story, for
cell in
in
both
the
the base period.
men and women
employment
rate
of the cross-country
relative
workers" because the distribution of
human
But
this
why
there
Germany and France between
look at
unemployment rates
capital
explanation cannot explain
slope the wrong way, or
A
some
explanation for the rise
of the
in
is
for different
across the workforce
why
Bell),
either
groups
may vary
of
across
the cross-country relationships often
evidently no relationship within countries such as
the decline
in
evidence cited
employment and wage
in
levels.
support of the relative-wage
European unemployment shows
it
to
inflexibility
be less than compelling.
For example, Siebert (1997) writes:
A
the
statistical
wages and
cross-country data between the development of relative
aggregate unemployment rates or
countries.
as
and the
evidence (including Nickell and
Lindbeck (1996) correctly points out that "we should not expect a close
in
women
Germany.
commenting on some
correlation
on
rate
men and women.
Figures 2 and 3 appears to hold for
U.S. displays a steeper relationship between the change
initial
relationship.
Table 2 reinforce these observations. This table
These regressions are weighted by the employment
The general
among
expect to find employment declining most
groups; instead, there appears to be
reports regressions of the
the log
we would
country which institutionally prohibits flexible
wages
at the lower
end can
15
have a low percentage of employment in low-paid jobs.
This is exactly what can be observed. Defining low-paid workers as those
who earn less than two-thirds of the median wage, the percentage of lowpaid workers in total employment varies noticeably with the dispersion of
earnings, from 5.2 percent in Sweden to 25 percent in the U.S. (Belgium
7.2, Netherlands 11.9, Italy 12.5, Germany 13.3, France 13.3, United
Kingdom 19.6; OECD 1996, Table 3.2).
be expected
A problem
with this
to
comparison
wage
one would expect
if
hypothetically that
wage floors
wage
in
some
the
is
that the
observed empirical pattern
had no
floor
raise
effect
countries with higher
relevant question
is
wage
would have a lower
countries that Siebert
of the ladder
lists,
there
is
a
and the incidence
chopped
to just
off,
is
or
fraction of "low-paid jobs".
striking
is
among
But the
were they simply elevated? For the
slight positive correlation
between the employment-
of low-paid work.^^
high school dropouts
among women. One would expect
rates for poorly educated youth to
that
associated with lower employment;
workers without an apprenticeship qualification or higher degree
it
suppose
below the median
Blau and Kahn (1997) find that the employment-to-population rate
higher than
this,
employment. Then one would expect
whether the wage compression
were the bottom rungs
to-population rate
floors
also exactly what
on employment. To see
low-wage workers' earnings
countries, without reducing their
is
in
the U.S.
The
in
in
the U.S. than
Germany
difference
labor market inflexibility to
be higher
among young
is
is
much
especially
cause employment
Germany.
^^The correlation between the employment-population rate using 1994 BLS data and
the fraction of low-paid workers reported by Siebert is 0.37, with a p-value of 0.42.
Belgium is excluded from this calculation because the BLS lacks data on Belgium. The
OECD
(1996) analysis cited above also finds insignificant and often wrong-signed
correlations for a larger set of countries.
16
If
the simple version of
employment growth across
appropriate.
rigidities,
We
throughout the distribution.
market
there
rigidities
is
Europe.
in
may
well
has trouble explaining the pattern
perhaps a more complicated
possibly because of unions.
--
firing restrictions,
do not want
that,
is
it
large effects on
or
inhibit
is
wage
Moreover,
job growth
of the
But, unfortunately,
employment problems
in
not clear theoretically that firing costs should lower average
as expected, more employment security
employment
employment.
much
Bertola, 1992). Analyzing data for ten
cyclical fluctuations in
may
of
dismiss these other forms of labor
to
scant direct evidence that they account for
Furthermore,
model
apply not only to the bottom of the
such as
because they may have
employment (see
found
floors
and top as
structure, but to the middle
rigidities
groups,
skill
wage
For example,
other forms of institutional
wage
employment.
unemployment and
He
OECD
countries, Bertola (1990)
associated with a lower amplitude
is
did not find a clear relationship
firing restrictions,
between
however. Anderson (1993), using
data on the experience rating of the U.S. unemployment insurance system, finds that
firms
hand,
have higher average employment when they face higher
some
layoff costs.
studies do find that job security legislation affects
1994, chapter 6 and Lazear, 1990).
test the effect of
By
their nature,
it
is
the other
employment (see OECD,
hard to identify and empirically
these institutions on employment.
Finally, there
is
the puzzle of
that in the U.S. for decades,
if
why the unemployment
rate
labor market rigidities are so
in
Europe was
critical.
that in the U.S.
If
a more
flexible labor
market
is
well
For example,
5 of the 21 years between 1975 and 1996 did the unemployment rate
exceed
On
in
below
in
only
West Germany
responsible for the recent low
17
unemployment
rate
in
that prevailed in the
Each
matter
much
in
rigid;
the past because there were no adverse
labor market
(2)
demand
shocks;
of these three reconciliations
may have some
merit, but they also
have
For example, which institutions changed, and what evidence links these
limitations.
suffer
There are three possible
the not too distant past?
in
take a long time to take effect after an adverse shock.
(3) rigidities
changes
U.S.
European labor markets have become more
reconciliations: (1)
rigidities didn't
the U.S., what accounts for the relatively high unemployment rates
employment growth? Furthermore,
to
adverse shocks
earlier in the
it
is
hard to imagine that countries did not
post-war period that would have
made
their rigidities
more apparent.
Whether the increase
in
in
wage
Europe represent two sides
suspect that
rising
wage
dispersion
of the
inequality
in
same
in
the U.S. and the rise
coin
the U.S.
is
is
in
an open question.
related to a decline
unskilled workers relative to skilled workers, the evidence that the decline
in
France and Germany
seems
have
to
little
is
concentrated
among
the unskilled
is
us that more must be going on than just labor market
doubt that labor market
rigidities
unemployment
in
in
Although
we
demand
for
employment
surprisingly shallow.
rigidities --
although
are an impediment to job growth
in
It
we
some
countries.
III.
Observations on U.S. Job Growth and Population Growth
How does
population?
the U.S.
economy manage
to provide
so
many
jobs for a growing
Several examples that are described below suggest that the U.S. has
18
managed
change
in
to
absorb population growth
real
fairly inelastic,
than
is
growth
wage
so
it
observed.
is
is
levels.
employment
with relatively
little
Moreover, much evidence suggests that labor
unclear
The
into
why
below suggest
We
do not profess
as much a mystery as a miracle.
labor market equilibrates (or whether
it
market adjustments take place with
little
that U.S.
to
demand
is
wages by more
population injections do not reduce
conflicting observations
systematic
employment
understand
how
the
equilibrates), but recognizing that large labor
wage adjustment
is
an important step
in
understanding the U.S. employment miracle.
Examples
of
Employment and Wage Adjustments
Population Shocks
to
Immigration provides a classic population supply shock that would be expected to
influence wages. Yet careful analyses of four large episodes of immigration
boatlift to
Miami; the repatriation of Algerian French workers to southern France; the
migration of workers from Angola and
colonies;
the Mariel
--
Mozambique
and the recent mass migration
of
to Portugal after Portugal lost
from the former Soviet Union to Israel
reach the surprising conclusion that large immigrant flows did not depress the
natives (see Card, 1990; Hunt (1992);
Carrington and
De
The swings
in
in
Ireland,
have not been accompanied by
labor supply brought about by the
another example of large employment
shifts with
-- all
wages
of
Lima, 1996; and Friedberg
(1996)). Olson (1996) also notes that large population declines
as was the case
its
due
real
to outmigration,
wage
such
increases.
baby boom and baby bust provide
seemingly
little
consequence
for
wages.
19
The
early
work on the absorption
baby boom cohort
of the
large cohort of workers experienced
in
the U.S. found that this
modest wage reductions when they entered the
labor
market, but faster earnings growth thereafter which diminished their losses (see, for
example, Welch, 1979).
Subsequent research by Berger (1985) found
reduction for the baby cohort tended to be larger and
in
the experience-earnings profile
call for
the entry of the smaller baby bust cohort
decline
in
earnings
earnings profile to
for
young workers;
become much
the late 1980s
in
If
Card and Lemieux, 1996). Moreover,
up
this pattern
Notably,
was accompanied by a sharp
cohort size were
for the
wage
But recent trends
phenomenon has caused
this
expected the cross-sectional experience-earnings
bid
persistent.
a re-examination of this conclusion.
steeper.
because wages would have been
more
that the
profile to
critical,
become
the experience-
one would have
flatter in this
period
newly entering baby bust cohorts (see
seems
to
be appearing worldwide (see
Blanchflower and Freeman, 1997).
Another example concerns female employment
earnings gap between
1994, the fraction of
to
56.2%
(while the
male earnings
relative
ratio
men and women has
women age 20 and
employment
rate for
the U.S., which has grown while the
narrowed. For example, between 1979 and
over
who were employed
men
declined by 3.9 points), and the female-to-
increased from .63 to
wage and employment
in
.73.^"
trends to a
One
demand
increased from 47.7%
might be tempted to attribute these
shift in
favor of
women,
but there
is
^''These statistics are from Report on the American Workforce U.S. Department of
Labor, Washington, D.C. 1995. The earnings data are for women age 25 and over, and
.
pertain to
1980 and 1994.
20
reason to doubt
higher
is
wage
this explanation; in tine
groups, and
women
women were
that, in the past,
It
is
well
relative
known
Table 3 presents
that
discriminated against
rate
real
in
An
women
wages are only
in
and a time trend
toward
alternative explanation
A
the labor market.
well as related
of
to liave siiifted
changes
in
decline
in
social norms,
to rise.
mildly pro-cyclical
recent evidence that reinforces this point.
regressions of the annual change
unemployment
women, as
wage and employment
aggregate
some
demand appears
tend to earn less than men.
labor market discrimination against
caused both the
1980s,
the log of the real
wage on
for several different
measures
The
the
in
the U.S.''^
table reports
change
of real
in
the
wages and
over various time periods. These results suggest that the weak responsiveness of real
wages
to the state of the
business cycle has
Although the aggregate wage data
may be
become even weaker
biased by changes
in
since the 1970s.^^
the composition of the
workforce over the cycle, the employment cost index (EC!) data attempt to adjust
for
composition changes by holding the job mix constant. Notably, the extent of pro-cyclical
wage movements
are even smaller with the ECl compensation measure.^^
^^The literature on the cyclicality of real wages began with Dunlop (1938). For a
thorough survey, see Abraham and Haltiwanger (1995).
^^The price index used to deflate these series is the CPI-W.
Abraham and
Haltiwanger report that wages tend to be more procyclical when they are deflated by this
price index than by the Producer Price Index (PPI).
^^See Solon, Barsky and Parker (1994) for evidence from longitudinal data that
in the composition of the workforce over the business cycle masks greater
changes
procyclical
movements
in real
wages.
21
The preceding observations
is
highly elastic
direct
a
shift in
be interpreted as evidence
population can be employed with
evidence on employment demand
emphasized
elasticities,
that
little
employment demand
change
however, suggest a
For example, the center of the range of Marshallian
of elasticity.
labor
--
migint
Hamermesh's (1993) review
in
demand
relatively inelastic,
described above resulted
in
it
of the literature is
seems anomalous
in
wages. The
fairly
demand
elasticities
-.50.''^
Thus, with
around
that the large population
of large intertemporal labor
supply responses. This too
microeconometric evidence (see Killingsworth, 1983
IV.
swings
employment gains without reducing wages.
Others, of course, interpret the weakly pro-cyclical tendency of real
evidence
low level
for
is
wages as
contradicted by
much
a review).
Theoretical Explanations
A
great deal of labor market adjustment
systematic real
demand
shifts,
wage movements.
while
in
others
it
is
wage changes appear only weakly
While labor market
rigidities
In
related to
seems
to take place without
cases, the adjustment
is
associated with
shifts; yet in either
employment movements. How can
part of the explanation for inflexible
to explain
population changes at seemingly fixed wages.
comes from
the U.S.
associated with supply
may be
themselves they are hard pressed
some
in
why employment
We
scenario
this
be?
wages, by
often rises or falls with
speculate below that recognizing the
employment effects of minimum wage
increases, at least in the U.S. A variety of evidence has found only small or undetectable
movements in employment associated with minimum wage increases (see Card and
Krueger, 1995, and Brown, Gilroy and Kohen, 1982).
^^Further evidence
the very small
22
role of certain product
observed empirical
market
restrictions in addition to
economy
course,
is
as population
A problem
is
doubled the capital stock and other inputs
An
proportion to population; namely, "entrepreneurial talent."
bring about an equal proportionate increase
in
number
the
of
is
that
may
rise
may
roughly
influx of population
people with
not
skills
in
may
capable
becoming an employer.
This point can be
made
clear
in
a modified version of Lucas's (1978) model of
employment and management choice.
In
Lucas's original model, a fixed population
chooses between supplying labor as an employee
equally productive
talent
or on
Consequently,
that
if
all
wages.
First
We
In
All
rigid in
is
a distribution
of entrepreneurial
Lucas's model, there are no restrictions on entry into
output
resources are
wages are more
Europe.
or as an entrepreneur. All people are
they work as laborers, while there
across people. ^^
management
in
wage
with this interpretation, of
But one often-overlooked form of inputs
necessarily double.
of
can help explain the
population at a fixed real
displays constant returns to scale.
that
rigidity
facts.
One reason why employment may grow with
the
wage
fully
is
sold at a constant price, taken as the numeraire.
employed.
Europe and
add these two constraints
that
A number
new
to the
of
commenters have noted
entrepreneurs are more constrained
model.
consider the effects of a constraint on wages, such as a
wage
floor.
As
in
the standard model, firms would choose to hire less labor to equate the marginal product
^^The idea of modelling a distribution of entrepreneurial talent goes back to at least
J.-B. Say (see Palmer, 1997, Chapter 4). The Lucas model can also be thought of as a
special case of the Roy model.
23
wage
of labor with the binding
uniformly across the labor force.
expected wage
managers
now be
--
queue
because
of
that
demand
a wage
(MGI)
in
growth
to
a recent study of
in
than one, the
and
this will
In
-
managers would
workers
in
the process.
a series of cross-country
regulations,
rather than labor market rigidities
productivity
MGI
France,
distributed
draw marginal
emphasizes product market
employment and
Germany and
for labor is less
to release their
zoning laws, and restrictions on employers
as the main barriers
is
floor,
consider the employer side of the market.
Institute
unemployment
the least talented
In particular,
causing them
for work,
McKinsey Global
example,
assume
the elasticity of
into the for-hire workforce.
Now
restrictive
If
of laborers will rise
enticed to
studies,
We
floor.
in
many
sectors.
(1997) concludes,
"In all
For
the
sectors studied, the main barriers to higher productivity have been product market
regulations that
and global competition
limit fair
constraint on employers,
permitted to
become an
as does the wage
floor.
suppose
that
If
number
the
no more than a proportion
The
entrepreneur.
of
restriction
would be the
when a
--
is
first
In equilibrium,
like to
of
new
firms
-- is
curtailed
least
their
own
firms.
is
countries,
some
potential
managerial talent (and smallest
for labor is
more
for labor
inelastic
adjustment
by regulation.
there are unemployed resources
open up
of the population
restricted, then
a binding constraint on entrepreneurs because one margin
namely the formation
would
is
become employees. The demand
to
of capturing the
a probably varies across
employers
employers become employees. Employers with the
firms)
As a simple way
...."
in this
model, and
An exogenous increase
in
some workers
population would be
24
associated with a proportionate increase
the proportion of the population that
If,
however, the
grow
in
employment and unemployment, as long as
permitted to
is
become an entrepreneur
on entrepreneurs constrains the number
restriction
population.
rises,
Intuitively,
and employment
rise
will
but
entrepreneurs to
unemployment
less than proportionately with population, then the
the population
of
if
the
number
cannot
of firms
rate
will rise
as
than proportionately with
less
the constraint on the employer-side of the market
intense with population growth
constant.
is
becomes more
rise in proportion with
population.
More
generally, the
demand curve
for labor will
become more
constraints on capital or other factors of production that
This result
demand
labor
a manifestation of Marshall's
for labor declines
Appendix).
have the
is
if
in
will
there are
demand
law of derived
is
--
the elasticity of
inelastic (see Hicks, 1932,
the capital market or on entrepreneurs can obviously
making the supply
demand curve
if
the supply of these factors.
limit
the supply of capital to a firm
Certain constraints
effect of
third
inelastic
of
be constrained
these resources
inelastic.
relative to a situation
in
In this situation,
which capital
is
the
elastically
supplied.
Figure 4 illustrates the effect of a
curve.
The unconstrained
respectively.
effective
The wage
labor
demand curve
cause the demand curve
to
floor
and constraints on the labor demand
demand and supply curves
fixed at w-bar.
is
demand curve (denoted
the effective
wage
The
are given by
become more
and SS,
constraint on entrepreneurs
causes the
We
have drawn
D'D') to rotate clockwise, as depicted.
quite steep, to
DD
emphasize
inelastic.
that product
market constraints
Notice that simply lowering the
wage
25
floor will
have a modest impact on employment because the demand constraint causes
the effective
demand
the effective
demand
(as
become more
curve to
curve to the
right
if
inelastic.
An increase
population
will shift
the population increase relaxes the constraint
employment
the model with heterogeneous entrepreneurs), causing
in
in
to rise but not
wages.
This model also
for
we
example,
constraint
is
may
help
suspect that the wage constraint
demand curve
it
demand
closer to the unconstrained
firms are less constrained
Germany
cross-country comparisons.
in
is
to shift out
much more
country's difficulty
in
in
(1
level.^°
The
demand
curve.
in
of the
for
enough
capital,
labor
is
so effective labor demand
20|
is
demand
we
up
start
expect the effective
By
contrast,
in
when
population expands.
of output
is
in
fixed
is
constrained
Figure
4.
In
in
the
the Barro
above the equilibrium
is
therefore determined by the product
hired to
meet the constrained product market
of labor
workers because they cannot
model. Allowing
the U.S. and the
features as the model
demanded
Germany,
Furthermore, because
example, the product price
quantity of output
Just
same
to
businesses, which might contribute to the
which the quantity
in
demand. Although the marginal product
hire additional
new
providing additional jobs
971 ) model,
curve.
in
proportion with population growth.
difficult to start
Notice also that a model
and Grossman
lower
the U.S., as population grows
about
aggregate would have many
is
Compared
exceeds the wage, employers do not
sell their output.
Barro and
Grossman
perfectly inelastic with respect to the
for capital-labor substitution will introduce
some
wage
ignore
in their
elasticity to the effective
For a review of non-Walrasian models, see Benassy (1986).
26
demand
the
but
curve,
unconstrained case because there
cause the demand curve
if
model, however,
this
the equilibrium level
The
is
is
no scale
effect.
if
will
still
be less than
An increase
demand
in
is
reason
for the
An
unsatisfactory
product price to be fixed above
is
to
emphasize
that rigidities
may
arise from areas
outside the labor market, but have an important impact on labor demand.
number
may be
partially relaxed
companies
start-up
of
as the population grows.
or capital
many
starting
countries.
new
constrain
wage
The U.S.
is
employment by making
floors
With
less restrictive than
enterprises and bringing
make
little
new products
effective
demand
effective labor supply elastic
direct
Moreover,
Restrictions on the
market imperfections or product market
regulations or restrictive zoning rules or restrictive store hours
In
demand
unspecified.
point of this discussion
these restrictions
the
the product market rises with
rise to satisfy the larger population.
that the
in
population would
in
population increases because of immigration, the
goods and services may also
aspect of
demand
to shift to the right
the population. For example,
for
labor
of
elasticity
in
evidence on the importance
many
may depress employment
countries
to market.^^
when
comes
to
barriers
may
same way
that
These
relatively inelastic, the
it
sticky-wage models.
of product
market constraints,
at this
stage our theoretical predictions should be regarded largely as conjectures (hence our
title).
The impact
^^Even
the
of
product market constraints on the
however,
demand
for labor
should be a
a reduction in
entrepreneurs' personal income tax rates induces them to hire more workers. A plausible
interpretation of this finding is that entrepreneurs are liquidity constrained, and finance
in
U.S.,
Carroll
additional expansions with their tax windfalls.
et
al.
(1996)
find
that
27
priority for future
environments
and
Identifying cross-country differences
research.
for specific industries
role of trade associations
would be one route
for future
--
--
and
work
in
product market
including the competitive structure of the market
linking
in this
these differences to employment growth
area to take.
28
V.
Conclusion
The
U.S. has been very successful at providing jobs for our growing population.
A great deal
of
employment adjustment seems
movement. The empirical observations and
demand-side constraints
are important.
(e.g., restrictions
demand
the
demand
greater
in
and
The
for labor.
many European
little
relax
for labor while
inclination to
programs) and
in
its
some
of the
demand-side
population
become an
is likely
to
in
constraints,
wages remain
supply increases and
increase the pool of
entrepreneur, thereby increasing
constraints on the demand-side of the market
countries than
wage
paper suggest
in this
social insurance
real
on entrepreneurs and product market regimes)
roughly constant. For example, an increase
individuals with the talent
and
floors
A growing population may help
thereby increasing the
take place with relatively
theoretical conjectures
wage
that both supply-side constraints (e.g.,
to
the U.S., and
may
may be much
prevent employment from
rising with population.
Much
policy discussion
of the market.
A
has focused only on the supply side
neglected concern
market regulations and
institutions
and become more
is
may
that restrictions
distort labor
flexibility
would generate more wage dispersion but possibly
Just as sharpening one blade of a scissor
blades are
be less
dull,
reducing
effective than
is
wages
will
scenario, providing additional
not greatly
little
additional
wage
employment.
enhance performance
without relaxing constraints on the
commonly expected.
rigidities)
demand, causing the labor demand
shift in
In this
wage
on entrepreneurs and product
curve to
inelastic.
(e.g.,
if
both
demand curve may
29
References
Abraham, Katharine and John Haltiwanger. 1995. "Real Wages and the Business Cycle."
Journal of Economic Literature 33, no. 3, pp. 1215-64.
Anderson, Patricia. 1993. "Linear Adjustment Costs and Seasonal Labor Demand:
Evidence from Retail Trade Firms," Quarterly Journal of Economics 108, p. 1015-1042.
.
and Herschel Grossman. "A General Disequilibrium Model
Employment." American Economic Review 61, pp. 82-93.
Barro, Robert
of
Income and
,
1986. Macroeconomics:
Benassy, Jean-Pascal.
Approach
New
.
York:
Bertola, Guiseppe. 1990. "Job Security,
Review
.
An
Introduction to the Non-Walrasian
The Academic Press.
Employment, and Wages," European Economic
34, p. 851-879.
Berger, Mark.
1985.
"The Effect of Cohort Size on Earnings Growth: A Reexamination
Economv ,93. no. 3 (June), pp. 561-573.
of the Evidence," Journal of Political
Turnover Costs and Average Labor Demand," Journal
Labor Economics 10, 389-411.
Bertola, Guiseppe. 1992. "Labor
of
,
Blanchflower, David and Richard Freeman.
OECD," mimeo., presented
Advanced Countries.
the
in
at
1997. "Analysis of Youth Labor Markets
in
NBER Conference on Unemployment and Employment
and Lawrence Kahn. 1997. "Gender and Youth Employment Outcomes:
The US and West Germany, 1984-91," NBER Working Paper No. 6078, Cambridge, MA.
Blau, Francine,
Brown, Charles, Curtis Gilroy and Andrew Kohen. 1983. "Time Series Evidence on the
Effect of the Minimum Wage on Youth Employment and Unemployment." Journal of
Human Resources
18, pp. 3-31.
Calmfors, Lars and Anders Forslund. 1991.
Market
Policies:
"Real- Wage Determination
The Swedish Experience," The Economic Journal
1
01
,
and Labour
(September), pp.
1130-48.
Card, David, Francis Kramarz and
Structure of
France."
Thomas Lemieux.
Wages and Employment: A Comparison
NBER
1996.
"Changes
in
of the United States,
Working Paper No. 5487, Cambridge, MA.
the Relative
Canada, and
30
Card, David and Thomas Lemieux. 1996. "Adapting to Circumstances: The Evolution
of Work, School, and Living Arrangements Among North American Youth," mimec,
Princeton University.
Card, David and Alan Krueger. 1995. Myth and Measurement:
the Minimum Wage Princeton, NJ: Princeton University Press.
The New Economics
of
.
"The Impact of the Mariel
Card, David. 1990.
Industrial
and Labor Relations Review 43,
Africa
De
on the Miami Labor Market."
no. 2 (January), pp. 245-47.
,
Carrington, William and Pedro
Boatlift
1996. "The Impact of 1970s Repatriates from
and Labor Relations Review 49, no.2
Lima.
on the Portugese Labor Market."
Industrial
,
(January), pp. 330-347.
Douglas Holtz-Eakin, Mark Rider, and Harvey Rosen.
Carroll, Robert,
Taxes and Entrepreneurs' Use of Labor," Princeton
Paper No. 373 (December), Princeton, NJ.
Dunlop, John.
pp. 413-34.
1
Edin, Per-Anders,
Wage
Movement
938. "The
Inequality
in
and
Real and
New
1996.
.
191,
"Unemployment and
1996. "An Evaluation of the Swedish Active Labor
and Received Wisdom," forthcoming
R. Topel, (eds.).
Working
Money Wages," Economic Journal
Anders Harkman and Bertil Holmlund.
Sweden," mimeo, Uppsala University.
Forslund, Anders and Alan Krueger.
Market Policy:
of
1996. "Income
Industrial Relations Section
The Welfare State
in
in
R.
Freeman,
B.
Swedenborg,
Transition (Chicago: University of Chicago
Press).
Friedberg, Rachel.
1996.
mimeo.. Providence,
Rl,
"The Impact
Brown
of
Mass
Migration on the Israeli Labor Market,"
University.
Godbout, Todd. 1993. "Employment Change and Sectoral Distribution
1970-90". Monthly Labor Review (October), pp. 3-20.
Hamermesh,
Daniel. 1993. Labor
Hicks, John.
1932.
Hunt, Jennifer.
Labor Market."
The Theory
Demand
of
in
10 Countries,
(Princeton, NJ: Princeton University Press).
Wages (New
York: Macmillan Press).
1992. "The Impact of the 1962 Repatriates from Algeria on the French
Industrial and Labor Relations Review 45, no. 3 (April), pp. 556-72.
1995. "The Effect of Unemployment Compensation on Unemployment
Germany." Journal of Labor Economics 13, no. 1, pp. 88-120.
Hunt, Jennifer.
Duration
in
,
31
Lawrence and Bruce Meyer. 1990. "Unemployment Insurance, Recall Expectations
and Unemployment Outcomes." Quarterly Journal of Economics 104, pp. 973-1002.
Katz,
.
Killingsworth, Mark.
1983.
Labor Supply Cambridge, England: Cambridge University
.
Press.
Krugman, Paul. 1994. "Past and Prospective Causes of High Unemployment," Reducing
Unemployment: Current Issues and Policy Options A Symposium Sponsored by the
Federal Reserve Bank of Kansas City, August 25-27, Jackson Hole, Wyoming, 1994, pp.
.
49-80.
Stephen Nickell, and Richard Jackman. 1991. Unemployment:
Macroeconomic Performance and the Labor Market (Oxford: Oxford University Press).
Richard,
Layard,
Lazear, Edward. 1990. "Job Security Provisions and Employment," Quarterly Journal of
Economics 105,
Levine,
pp. 699-726.
1993.
Phillip.
Unemployed."
Industrial
"Spillover Effects Between the Insured and Uninsured
and Labor Relations Review 47, no. 1, (October), pp. 73-86.
.
Lindbeck, Assar. 1996. "The West European Employment Problem." Weltwirtschaftliches
Archiv (Review of World Economics), pp. 609-637.
Lucas, Robert.
Economics
"On The Size Distribution
2 (Autumn), pp. 508-523.
1978.
9. no.
of
Business Firms,"
Bell Journal of
Meyer, Bruce. 1989. "A Quasi Experimental Approach to the Effects of Unemployment
Insurance," NBER Working Paper No. 3159, Cambridge, MA.
Meyer, Bruce D. 1995. "Lessons from the U.S. Unemployment Insurance Experiments,"
Journal of Economic Literature 33, March 1995, pp. 91-131.
.
Stephen and Brian
Nickell,
Unemployment Across
the
Bell.
1995.
"The Collapse
OECD." Oxford Review
of
in
Demand
Economic
for the Unskilled
and
Policy 11, pp. 40-62.
1996. "Changes in the Distribution of Wages and Unemployment in OECD
Countries." American Economic Review, Papers and Proceedings 82 (2), pp. 302-308.
.
Olson, Mancur. 1996. "Big Bills Left on the Sidewalk: Why Some Nations are Rich, and
Others Poor." Journal of Economic Perspectives 10, no. 2, pp. 3-24.
Organization for Economic Co-operation and Development. 1996. Employment Outlook
(July), Paris, France.
32
1994.
.
The
Adjustment Potential
L
Palmer,
OECD
of
Jobs Study
the Labor Market.
1997. J.-B. Say:
Evidence and Explanations.
.
Part
II:
The
Paris, France.
An Economist
in
Troubled Times (Princeton, NJ: Princeton
University Press).
"Labor Market Rigidities -- At the Root of Unemployment
Forthcoming, Journal of Economic Perspectives
Siebert, Horst.
Europe."
1997.
in
.
1985.
Solon, Gary.
"Work Incentive Effects
of
Taxing Unemployment Benefits,"
Econometrica 53, pp. 295-306.
,
Solon, Gary, Robert Barsky, and Jonathan Parker. 1994. "Measuring the Cyclicality of
Real Wages: How Important Is Composition Bias?" Quarterly Journal of Economics 109.
February, pp. 1-26.
1979. "Effects of Cohort Size on Earnings: The Baby Boom Babies'
Financial Bust." Journal of Political Economy 87, no.5, October, S65-S98.
Welch,
Finis.
Table
1:
Determinants of Cross-Country Employment Growth, 1959
-
95
Dependent Variable: Log Employment
-
Log
1.29
Population
(4)
(3)
(2)
(1)
Women
Men
All
~
-
(5)
1.21
(6)
(7)
-
"
1.60
.95
-.52
-.10
.43
-.98
-.66
1.88
(+ 100)
(.05)
(.04)
(.09)
(.04)
(.04)
(.08)
(.07)
Yearx
1.01
.40
.81
.27
(.13)
(.06)
(.12)
(.05)
-.73
-.18
-.47
-.09
(.13)
(.06)
(.12)
(.05)
Year
U.S. (+
(9)
-
(.07)
(.03)
(.03)
(8)
1.09
(.19)
.07
.49
(.09)
(.17)
.41
(.12)
100)
Yearx
Germany
-1.27
-.37
(.20)
(.13)
Yes
Yes
(-^100)
Yes
Country
Yes
Yes
Yes
Yes
Yes
Yes
Dummies
Log Population
Interacted with:
1.28
U.S.
Canada
Australia
Japan
1.17
(.06)
(.05)
(.11)
1.25
1.10
1.69
(.05)
(.04)
(.08)
.97
.90
(.05)
(.05)
Germany
Italy
Netherlands
R2
(.13)
.62
.60
.86
(.98)
(.09)
(.20)
.47
.73
.23
(.14)
(.10)
(.31)
.40
.68
.51
(.12)
(.10)
(.22)
Notes:
Sample
size
.999
is
(.18)
1.69
.80
(.15)
(.15)
.69
.60
(.23)
(.18)
.999
351 for Columns
2.70
.79
(.09)
1.05
.994
.49
(.06)
(.10)
U.K.
(.09)
(.07)
1.27
Sweden
1.32
1.27
.90
France
1.58
.995
1
-
.999
(.28)
1.43
(.46)
.986
.999
3 and 343 for Columns 4
-
9.
.995
.999
Table
2:
Regression of Change
in
Log Wage or Employment-Population Rate on Base Wage, United
States and Germany
Women
Men
Independent
A mean
A empi.
A mean
A empl.
variables
In(wage)
rate
In(wage)
rate
1.
ln(1
979 wage)
.393
.075
.309
.092
(.038)
(.018)
(.056)
(.038)
2.
ln(1
979-80 wage)
U.S. (20 groups)
-.140
(.084)
Germany
(20 groups)
.008
-.104
-.069
(.018)
(.066)
(.053)
Table
3: Cyclicality of
Real Wages, Various
Wage
Change in
Unemployment Rate
Employment Cost
A. Series:
1981:Q1
-
1997:Q1
Series and Time Periods
Trend
n
-.07
-.09
65
(.12)
(.03)
Index; Quarterly Data
Production/Nonsupervisory Workers (CES); Annual Data
B. Series:
1948-1996
1948-1980
-.44
-.09
(.20)
(.02)
-.46
-.14
(.25)
(.03)
-.19
-.00
(.21)
(.04)
49
33
"
1981
-
1996
C. Series:
1965:M1
1981:M1
-
-
Production/Nonsupervisory Workers (CES); Monthly Data
1997:M3
1997:M3
1970:M1 -1979:M12
1980:M1
1990:M1
-
-
16
1989:M12
1997:M3
-.66
-.07
(.08)
(.01)
-.15
-.00
(.06)
(.01)
-1.12
-.48
(.14)
(.06)
-.24
.16
(.13)
(.06)
-.05
.28
-
Continued
387
195
120
120
87
Table
D.
3:
Continued
Hourly Earnings by Decile from Outgoing Rotation Groups of Current Population Survey, 1973-96
Percentile
10th
50th
90th
Change in
Unemplovment Rate
Min.
Wage
Increase
Trend
n
23
-.95
.034
.09
(.48)
(.014)
(.10)
.04
-.009
-.06
(.24)
(.007)
(.05)
-.36
-.006
-.04
(.26)
(.008)
(.05)
Notes: Reported coefficients are from regressions of the annual change
23
23
in
the log real
wage
(deflated by
CPI-W) on the annual change in the unemployment rate (divided by 100), a linear time trend (year divided by
100), and an intercept. In panel D, a variable indicating whether the minimum wage increased that year is
also included; this variable equals one if the minimum wage increased on January 1st of the year, and the
fraction of the year in which the higher minimum wage was in effect for years in which the minimum
increased after January 1st. The dependent variable used in Panel D was provided by Jared Bernstein; all
other wage data are from the BLS.
,
;
Figure
Change in Log Employment
.2
vs.
1
Change in Log Population, 1964-95
2
2
2
'^:^^^^
?
c
cu
E
o
2
1
.1
-
B
10
10
Be
1
^^
jl
1
1^
jn^
10
2
2
/I
O
D1
C
-
fO
%^^R)"6'
c
fD
O)
I
in
-.1
-
i
.05
i
.15
.1
5-Yean Change in Log Population
Country codes: U.S.=1; Canada=2; Australia=3 Japan=4
France=5; Germany=6; Italy=7; Netherlands=8 Sweden=9;
UK=10.
;
;
Source: BLS International Comparisons Program.
en
CNJ
QJ
-M
fD
CO
CZ
d
UJ
_CNJ
CNJ
QJ
o
cn
cn
cn
cn
CU
CD
c
I
C^
o
OvJ
QJ
CJ1
cn
QJ
Q-
"a
en
ro
fD
o
-CM
-rH
Ll
cn
QJ
c
QJ
fD
QJ
CD
CD
ru
fD
cn
cn
cn
cn
o
QJ
cn
QJ
cz
CD
cn
c:
fD
JZ
u
o
00
D
sa^BH ;uaiuAoiduJ3 'sbBbm ui aBueqo
MIT LIBRARIES
3 9080 01444 4050
-m
cu
CD
CD
cr
o:
en
CD
B
o
C)
en
CD
en
CNJ
en
c
CD
CL en
I
on
_co
^^
CJ
CD
O
CD
cm
o
zi
CD
c:
ro
•r-l
j_
„
cz
CD
en
CD
:^
cji
c:
ID
C\J
CD
CD
(3n
o
C
-rH
CD
CD
£Z
CD
CD
(JJ
CD
en
CD
rsj
o
D
sa^ey
;u3LuAoiclaJ3
'saSeM ut aBuEL|3
c=
CD
CD
OS
O
O
Q
03
72
n
Date Due