Continuation of NLS Discussion Paper 93-16
Part 2 of 3
This version of the paper was split for web delivery.
6.
Characterizing
Estimation
Labor
of our statistical
Market
mdel
Experiences
formally
provides
and Mobility
direct evidence only on the
.
short-teml
relationships
liting
high-wage
employment,
simultaneous
low-wage
Transition
and entrance
and cducatiorrd
pursuits.
the five economic
a period in one status is immediately
status.
These probabilities
employment
long-teml
influences
simulation
specifications
probability
of the transition
employment,
probabilities
nonemployment
indicate the likelihood
that
by a period in the same or in another
One can, however,
in such an employment
a stitightforward
uncover
tie
status and a youth’s
sililulation
exercise.
This
and interaction “effects impl~ed “by particular
and entrance
probabilities
that characterize. a transition
model.
findings,
This collection
long-term
by implementing
the accumulative
To formulate
empirical
high-wage
employment,
indicate how a spell in, say, low-wage
between occupancy
prospects
captures
do not direcdy
of low-wage
the career path of an individual.
relationships
future earnings
followed
statuses
the comprehensive
we carry out several
of simulations
relationships
young men’s working
to summarize
pichlre of economic
simulation
allows US.to -calcula~
among. different
labor-market
summary
Iab.or-rnarket s~tuses
careers after leaving school.
the.cmnulative
analyses
under a variety of scenarios.”
s~tisti”~
illustrating
tie
dqring tile inifial years of
We exploit this iriformation
experiences
6.1
“mobility implied by our
of this population
not only
(i.e., the total
amounts
of time that individuals
characteti
the dynamic
extent of economic
simulation
exercises
each mtegory
respectively.
markets,
amount
for” the four education
race-etinic
of economic
experiences
W bile the measures
about the role of low-paying
information
Sections
,
presented
spells.
Still anotier
investigates
this dynamic
and the links be~een
smhlses,
leave school;
Section
in the five statuses
separately
in
6.2-6.6,
in low-wage
labor
this participation
and the
6.2 examines “=thedistribution
titid combinations
of these
in Section ””6.2 greafly improve our. knowledge
tiese
Of his SOfi provide. Q.nly limi~d
mrly labor-market
feamres
statuses
at the
of individual
how long it takes youths to enter highlabor-market
several
activities.
Section 6.5-
me?s~lres associ?!ed. .Yitll the
to enter high-paying
jobs from
._
the time when: they
they s~rt a low-wage job; and tiley become involved
6-2
experiences.
of this process by looking
along with tie durations
in various
facet. by sun!!llarizing
length of time that it takes individtlals
initially
of who participates
aspect of the picture concerns
given involvenlent
We carry out distinct
in five steps taken up in Sections
6.3 and 6.4 fti in some of the dynamic
of spells in all five labor-market
to
statuses and the
wifi results presented
jobs, cumu]?tiyg. s}!mmaties
number
also
of the role that Iow-.wage employment
about the dynamic process generating
paying employment
,
mobility
of time spent in other labor-market
statuses.
labor-market
but
groups.
To further our understanding
labor-market.
categories,
our characterization
tie extent of this participation,
of cumulative
between
activities),
mobility during tie first 10 years after school.
for the tiree
plays in tie process
~bor-market
patterns of movements
This section develops
.
spend in vtious”
in sonle form
..
of training.
specific
to complete
labor-market
discussing
histories
our simulation
Simulating
play in determining
the information
methods,
fiture
examined
commonly
in our TPM.
generating
entrance
probabilities
activities.
sequences
relevant
career:.
relationships
a simulation
of our empirical
for assigning
variables
model
transition
experiences
and
over tfle
by assigning
the hazard rates Hi(T) (relations
to the predicted
the
values of
the disgrete variable in. the week, under
probabilities
Pr(&i)
(relations
the values of. the discrete
of the working career when formal school ing.. erds;
(relations
among the economic
We carry out such an exercise
uniformly-distributed
in detemlinffig
while a-sp-ell in labor-market
prduce
indicating a “perso-n’s status in each week by comparing
The initial-status
quantities
the long-term
and econometrics,
to “mimic the process “governing. y“ouna.nlen’s
values of independency-drawn
ccmsideration.
6.2, we briefly
of Monte Carlo “tials using “our estimated
of discrete variables
the probabilities
Before
.
used in botfr statistics
Implementing
sequences
initial years of their working
Pr@+i)
employment
the role that
Experiences
needed to describe
involves
relevant
Section 6.6 explores
approach.
Ernployrnent
Simulation
statuses
our characterization,
our first step in the analysis which is tie subject of Section
describe
6.1
Finally,
variables
the transition
(4.3.) and (5. I))) are relevant
activity i is in progress;
(4,5) and (5.3)) are the relevant
6-3
(4. 1) and (5.4)) are the
at tile very beginning
probabilities
for assigning
arid the entrance.
numbers
Pi@,Z) or
.OUtcOl.nes
probabilities
for assignment
when a spell
. .,—
ends.
The analysis
calculates
the predicted
estimates
obtained in the empirical
variables
used in the simulation
the beginning
of a simulation
exercise
(~
status to reflect the simulated
simulated
analysis
include:
labor market history variables
ouflined in tile previous
demographic
that are updated
during
section.
(~
The
that are chosen at
over the sequence
at the beginning
or just completed.
of labor-market
using parameter
of winks;
of each new spell in a
up to the relevant week; and, the lengdl of the
every week over the 10-year period following
one to assign a sequence
covariates
and held constint
experiences
spe~ tilat is in progress
values for these probabilities
Evaluating
the completion
activities
these probabilities
of formal schooling
experienced
in
allows
by a typical individual
fie early years of his life cycle.
Repeating
experiences
the above procedure
for a large number
characterizing
the distribution
of hypothetical
sequences
labor-market
covariates,
racial backgrounds
of simulated
wee~y
experiences.
individuals
traits. 1 Examining
statuses
produces
provides
experienced
attainment
a comprehensive
with suitable adjustments
the same procedure
to simulate
with a v.arie~ of Iabor-i.na.rket histories
sequences
of wee~y
the basis for
the labor-market
and levels of educational
In addition,
one can implement
of individuals
times and recording
of the labor market activities
with a given set of demographic
with different
numerous
by young men
experiences
of men
over various
pictire
of cumulative
in the .v.alues of the
labor-market
as wel! as different
experiences
demographic
traits.
-
1 The findings discussed below are based on random draws of 1,000 hypothetical
for =ch demographic group.
6-4
individuals
6.2
Cumulative
Experience
Summarizing
statuses
during
foundation
economic
the cumulative
amount
of time men occupy various
the initial years of their labor-market
for constructing
mobility.
experiences
“provides a valuable
picture
of low-wage
employment
6.2-LQ
and 6.2-M
a comprehensive
Tables 6. 1-LQ, 6; l-M,
from the simulations
describing
tie five labor-market
statuses and combinations
the 10-year titervd
definition
of low-wage
correspond.i.ng
the cumulative
after leaving school.
table.
employment
of tiese
labor-market
period tisted in the fust set of columns,
education
results for Whites,
group designated
tables lists tie participation
tie two 5-year horizons
present summary
statistics
activities,
over various
horizons
presented
of
the
in the
of the total number
with results for the entire
of
10-year
for the first 5 years of ‘the period (i.e.,
years 1-5)
(i.e,, years 6:10) in the third set.
Blacks, and Hispanics,
at the left of the corresponding
spend in
“LQ” and “M” indicate
in tie computations
in the second set of”colurnn.s, and for the second Syears
Each cell reports
statuses
The top half of Tables 6.1 report averages
weeks spent in the different
and
amou,nt of weeks that individuals
The designations
assumed
economic
respectively,
The bottom
half of the
rates in the five “statuses over the 10-year period,
along with
mating
up this period.z
row.
for the
These participation
rates reflect the
Z Ii the mse of college grad[]ates, the 6-10 year subperiod of the.1 O-y~r h~rizon is out of
sample for the data set used in estimation. The vast majority of our observations making up the
college sample are in the first five ymrs after Iwving school.
OnIy a fe”w”“of our college”
graduates make it into the 6-10 ymr period and no one makes it past 8 years. We report results
for the 6-10 ya sample mainly for the purpose of making the college sample comparable to the
other dumtion
groups. All the other edueation groups have sufficient data in both the first and
second hdf of the 10-year horizon.
.
6.5
—
.
probability
an individual
experiences
at least one week in tie corresponding
status during
the relevant time perid.
Tables 6.2 present
simulated
distributions
the 10th, 25th, 50fh, 75fh, and 90ti percentiles
of the cumulative
market statuses and a combination
consider
number
assess the extent to which individuals
The results for this secondary
As in previous
respectively,
employment
for the education
group designated
The subsequent
The main conclusions
minimum-wage
results for cumulative
after discussing
Thus,
pichlre
experience
tie LQ restdts
in the findings
to
labor market shfus.
and Hispanics,
In additign
to the percentiles
in the first six sets of rows, the
for the cumulative
weeks associated
focuses on the LQ definition
drawn from our analysis
of a complete
we
with the
respectively.
discussion
definition.
Blacks,
in the row.
which are summarized
the percentiles
fwst and the second 5 years,
combination
and nonemployment
spend time in a “secondary”
fables, each ce~ hsfs results for Whites,
two lower groups present
differences.
The specific
labor-
labor market status are in the last set of rows in the fables,
over the entire 10-year period,
development
of weeks spent in the various
of these activities.
sums the total weeks spent in low-wage
for the
the discussion
experiences
we do not consider
We tien highlighf
implied by the LQ and the M definition.
6-6
the
away from tie
based on the M concept of low-paying
in their entirefy.
employment.
do not change when considering
to avoid diverting
of cumulative
of low-wage
the
employment
fhe relatively
until
subtle
6.2.1
Low-\Vage
Employment
For tie LQ definition,
interesting
patterns
in low-wage
employment,
spend an average
jobs,
The averages
of any education
Specifically,
low.
of slightly
and college graduates
with involvement
reported
in Table 6.1
group spends relatively
little
The time spent
but the averages
for the lowest
the results show that high-school
High-school
graduates
eulployment
spend approximately
(i.e. 75-9&”weeks).
about .75 of a year in low-paying
experience
of
over one year O.e. 67-77 weeks) in low-paying
over the 10-year period.
group experience
Stating
from 5 % to 20% of the 10-year penti.
1.5””to 2 years on average in low-paying
some-coflege
of youths.
is higher than might be expected,
group are surprisingly
employment
member
6.1 -LQ and 6.2-LQ reveal severrd
of Table 6.1 -LQ shows the average number
spend in low-paying
employment,
by college graduates
education
in Tables
the top portion
indicate that the typial
time in low-wage
dropouts
tie findings
in the labor market experiences
weeks that individuals
generally
and Nonemployment
around
from
Those in the
jobs 0 .e. 37:40 week);
.5 of a year on itierags
in low-Paying jobs (~-
25 weeks).
The bottom portion
of Table 6.1 -LQ shows that the Iike!.ihod
wage job at some time during the 10-year period is approximately
school dropouts
and graduates,
and approximately
high-school
4045”%-for
dropouts,
levels of education
75-80%
about 45-50.% for those individuals
those WIIO are college-
participation
in low-paying
and for older workers
~ad.yates..
jobs generally
of holding a low-
with some college,
with
tile exception
diminishes
(where older refers to experiences
6-7
for high-
of
for higher
in the second
half of the 1O-YW period);
f~
for persons
portion
both the average
w“ith higher educational
number
attainments
of weeks and the participation
and for individuals
rates
in the 6-10 year
of””the period after Ieaving school.
In the case of high-school
dropouts,
the participation
~ years
I-5 and 6-10! and tie average
typidly
higher in the’ sec~nd half of the I_O-year wriod
rate remains
about the mme
number of weeks spent in low-paying
afmr departing
jobs is
from fo~~
,
schooling.
declhe
Whereas
h average
year hotion,
low-wage
White high-school
weeks of low-wage
group.
For high-school
higher
reported
high-school
dropouts
in Table 6.2-LQ
tend to experience
for low-wage
jobs is not very high. for the vast
The 90th percentile
graduates
levels of education,
categories.
in the second five years of the 10-
drawn from Table 6. I-L.Q. ““”These findings
spent in low-paying
a slight 3-week
more
in the second 5-year period.
The percentiles
edumtion
tend to experience
participation
bodr Blacks and Hispanic
employment
conclusions
dropouts
Surprisingly,
and beyond,
barely exceeds
low-wage
dropouts
sLipport the
indicate~.~at .&e amount of time
nlajority
of individuals
in evew
4 years for any education
employment
and it is lower for the older workers
high-school
employment
typically
declines
group.
monotonically
for
in the upper educational
experience
less low-wage
.
employment
median
period),
tian their high-school
for dropou~
graduate
“is lower for older workers
the 75th and the 90th percentiles
Looking
counterparts
at the distribution
(i.e.,
over the 10 years.
While me
the second half of the 10-year
are not lower for older. workers..
of experiences
6-8
in nonemployment
reveals that the
smaller
amount of low-wage
arises bemuse
training
dropouts
employment
spend a consi-derable
or high-paying
combind
spent in
education,
low-wage
jobs and nonemployment
in terms of their experiences,
indimtes
that the number
with higher
groups which are roughly
category
of weeks
comparable.
Iow-
groups differ substantially
More &an 50% of
While the typical Black--as
measured
by the
less in this smte in the second five years of the 10-year
there is no improvement
for “the worst 25% of experiences;
weeks, and it is 217 weeks in years 6-10.3
While the experiences
spend
in “the combined state of e + n. over the 10-year period,
and 10% spend more tian nine years.
median \,atues--experiences
for the
resLdts for this
monotonically
that race-ethnic
in
in the 10-year. period than
“especially when considering” Blacks.
Blacks sWnd more than 6.5.-years
years 1-5 is ~2
groups discloses
of the results for the combined
for dropouts
shows that dropouts
“Comparing
decreases
except for the two upper education
wage + nonemployment
horizon,
~oup.
status across education
Careful inspection
graduates
in Table 6.2-LQ
jobs or nonemployment
in the same race-ethnic
labor-market
presented
low-wage + nonemployment
many more weeks in either low-paying
graduates
tian for high-school
number of weeks in nonemployment--not
The percentiles
employment.
labor market status designated
high-school
for dropouts
of Black high-school
dropouts
analysis. predic.rs that many White. and Hispanic dropouls
the 75th percentile
in
._
are especially
dismal,
tile
also spend a great deal of time
3 Of course, the individuals involved in the highest 25% of experiences in y=rs 1-5 are not
neces@ly the same pmple includd in the group with the highest 25% of experiences in Ytirs
6-10.
6-9
in low-wage jobs or out O! work,... About 50% Qf tie Hispanics
10-year period in the combined
s~tus
spend abotit half of tie
of f +n; and 10% of Hispanics
nine years in these labor market statuses,
participate
almost
Whites fare better than Blacks or Hispanics,
and yet 10% of them “spend about 7.5 years or more in low-wage jobs or out of work.
Differentials
.
low-paying
among the race-ethnic
groups
in the total amount of time spent in
jobs or out of work over the 10-year period steadily diminish
considering
higher levels of education,
graduates.
Blacks continue to stand out as the disadvantaged
graduates,
altiough
their position
with the differentials
seen in the case of high-school
of weeks spent in low-paying jobs for, White and .Hispanic
similar over the 10-year horizon.
percentiles
except in the @ils of the distribution
where a differential
lower parts of Table 6.2-LQ
of the 10-year period.
school graduates
category.
6.2.2
Whites and Hispanics
High-Wage
exists favoring
indicates
The patterns
does
The number
graduates
are very
also have similar nonemployment
represented
the Whites.
for race-ethnic
continue to apply when considering
counterparts
dropouts.
high-school
by the 75th and 90ti
Inspection
that most of this differential
observed
for college
group among high-school
relative to their White and Hispanic
not come close to the differentials
experiences,
disappearing
when
of the results in the
occurs in the first ha!f
groups in the case of high-
fiose individ{lals in the some-college
.-
Employment
and Training
The findings of Table&.6 .l-LQand6.2-LQ
6-10
forhigh-wage
ernploymen
tshowthat
the amount
of tiis type of employment
edumtion,
with experiences
either Whites or Hispanics
inversely
monotonically
with higher levels of
similar in the two upper groups.
in the three lower edumtion
linked to education
dropouts.
increases
and widening
Whites and Hispanics
Blacks fare less well than
mtegories,
to an enormous
have similar experiences
margin for high-school
in tie some-college
and in the upper percentiles “associated with lower edumtional
more high-wage
experience
below some college.
lower education
Whites
than Hispanics
levels in part reflects the favorable
of low-wage
For college
graduates,
are “essentially
demographic
groups.
edumtion
which is in sharp contrast
levels
of ~ose
over
some
by this seg!nent of Hispanics.
the same across-”the three
exists,
it favors tile Blacks”at tie
to the order operative
for the lower
groups.
The results in Tables 6.1 -LQ and 6.2-LQ
both low- and high-wage
employment
holding. at low- and high-wage
across educational
attainment s..
LQ reveal that about one-third
during
by Hispanics
which indicates
employment
To the extent that a differential
Whites have
of education
margin attained
employment,
the .~pical experiences
category
in tie bottom Percentiles
substitution” of low-paying. jobs for high-wage
lower percentiles,
attiimnents.
in the lower percentiles
The favorable. position of Wlliks
in the upper percentiles
with the differential
relating
to the nu!nber of weeks spent in
show that the phenomenon
positions
is not “uncommon
The. participati~ri
of iidjy(~uj)s
of simultaneous
and do=
the same week at some time during the 10-year period,
6-11
not differ nluch
rates (n the lower portions
will hold a low-wage
job
of.Table
6.1-
and a higi~-w?ge job
and the averages
in the
.
upper part of this table indicate that individuals
type of employment
tie up~r
indicate that only 10% of any education
Furthermore,
or more in this category
of employment.
this type of employment
across race-ethnic
Regardless
.
job holding
here
is no systematic
reveal that individuals
participating
tites” reported
from any education
in training
ye average
number
patterns in the amount
hold after exarntilng
of training
rehlrning
6.2.3
Differences
.
reported
indicates
for average
Based
of weeks spent
of the Table 6.1 -LQ
including
These
conclusions
Retill
apprenticeships,
after dropping
is no “more likely
and there is no systematic
in Table 6.2-LQ.
varies considerably
in Findings
A casual comparison
definitions
formal schooling,
to formal schooling
No doubt, this composition
in
group have around a 50-60””% chance of
reported
analysis refers to a varie~” of activities,
while working,
in the lower portion
across education. groups.
the percentiles
pattern
It ranges from 18 to
at some Jime over tie 10-year P*rig~. , Trqinipg
to occur in the first 5 years after leaving
6 months
groups.
of the educatiorr group considered,
The participation
in his
is not rare,
group experience
h training over the 10-year period after leaving school is not small.
59 weeks.
months
I-4
While simultaneous
over a 10-year horizon.
percentiles
spend approximately
continue
that training
on-the-job
across the education
i:. this
training
out, and vow~ional
training.
grollps.
on LQ and M Definitions
of tie results in Tables 6.1 and 6.2 for the. LQ and the M
a remarkable
cumulative
similarity
in the summary
weeks reported
6-12
smtistics.
to
The nLnnbers
in Table 6. -LQ are quite close to those
presented
in Table 6. 1-M, and the participation
tables are vi~a~y
A detafied
in the findings
summafied
examination
implied
of Tables 6.2-LQ and 6.2-M
by the WO definitions
groups.
counterp-
based on the LQ definition,
surprisingly,
the ordering
Most percentiles
low-wage
employment,
the two definitions.
wage + nonemployment
ktween
but these differences
which primarily
spells
as mtich as 20 weeks.
time spent in high-wage
is shorter
divergence.
for
Not
jobs.
using the M definition.
are
being only about-15
weeks.
but there
the M and the LQ percentiles.
in training over the 10 years differs
patterns across either education
identical,
for low-wage
tends to be longer using the M definition,
of the differences
time spent in nonemployment
can be
based on the LQ and M definitions
with the maximum
iri “both” employment
amount of time predicted
employment
with a few moving
in this type of employment
ranking
that tie differences
increase by only a few weeks over their
is “reversed .yhen considering
by only a few winks,
is no systematic
indicates
of the percentiles
Once zgain, in most instances the percentiles
The experience
of these
of time spent in low-wage jobs using the M definition
all education
experience
of low-wage
A comparison
reveals sfighdy longer amounts
different
in the bottom
identid.
quite succ”incdy.
Cumulative
rates reported
for tl]e two concepts
The
of
are “rnafginal, and there is no cilange in the
or race-edlnic
groups.
The predictions
of the amount
of
based on the LQ- and the M defifi[tion” are essentially
reflects the fact that this. state is defined
Cumulative
experience
is predicted
ii the combined
the same way under
smte of low-
to be higher using the M definition,
6-13.
with the margins
:
comparable
to observe
to those noted above for purely low-wage
in comparing
only the lower percentiles
percentiles
and below).
the cumulative
are different
lower education
education
6.2.4
groups,
in low-wage+
in tie case of high-school
All of the percentiles
groups under the M definition,
.
experience
employment.
An important
nonemployment
dropouts
fi.e.,
point
is that
the 50th
are higher for tie three upper education
but the lower ~rcentiles
and the upper percentiles
are the most affected
for the
are the most affected for tbe higher
levels.
Summary
of Rndings
for Cumulative
Experiences
The results in Tables 6.1 and 6.2 lead to the following
conclusions:
●
Young men spend relatively fittle time in low-wage employment
in the first
10-year period after leaving school. H[gh-school
graduates
experience
the
largest amount of involvement
in [ow-wage jobs, followed by high-schoo[
dropouts,
with persons possessing educational
attainments
abote high school
~i!e
high-school d~opouts and
having considerably
less experience.
graduates have a 757.-80% chance of holding a low-wage job at sometime
during the 10-year period, they average a modest 1-2 years of low-wage
employment
in this period; the majority spends less than one year; and only
10% spend more than 4 years in low-paying jobs.
The participation
rate of
college graduates
in .tbe low-wage sector in the first 5 years ~fter_:chool
is a
surprisingly
high 30Y0, but the average amount of time in low-wage jobs for
this group is only about 3 months over this period.
●
me combined state of low-paying e~lployment
and nonemployment
provides a more relevant notion of hsbor-inarket
stialrss than time spent in
low-paying employment
atone for judging. the prospects of individuals
during the initial stages of their working careers.
The larger .amognt of
low-wage employment
for high-school graduate:
relatiye_ to_d~opouts reflects
the fact that a significant
segment of the dropout population
spends a
considerable
amount of time in nonemployment--not
in training or high-
6-14
paying employment.
wage+ nonemployment
for older workers.
●
The extent
decreases
of participation
in the state of lowsharply for higher levels of edttcation
and
Black high-school dropouts spend an exceptionally
large amount of time in
the combined state of low-paying
employment
and nonemployment;
the
prospects for some mite
and Mspanic
dropouts are not much better.
More than 50% of the Black spend more than 6.5 years in low-paying jobs
or out of work during the first 10 years after school; and 1070 spend more
than 9 years.
About 50% of Hispanics
spend around half of the 10-year
period in low-wage employment
or out of work; and 10% participate
abnost
g years in these statuses.
Mite
dropouts
fare better than Blacks and
Hispanics,
and yet 10% of mites
spend about 7.5 years or more in lowwage jobs or out of work.
. Not surprisingly,
the amount of high-wage .employrnent
is greater for higher
levels of edumtion,
and the amount of nonemployrnent
is smaller.
The
average amount of time spent out of work in the 10 years after school for
high-school
dropouts ranges from 2.5 years for mites
to almost 5 years for
Blacks.
These averages fall to Ie.ss. than a year for those individuals
with
educational
attainments
above high school.
. Simultaneous
job-holding
of 1OW- and high-wage positions is not uncommon
and does not differ much across education
groups, ”’but only 107.. of any
group experience 6 months or more in this category of employment.
Any\vhere fron~157G-30%
of any education
group hold a low-wage and a
high-wage job during the same week at some time during the 10 years
following school.
. The average number of
following school falls in
education group having
form of trainingat
least
weeks spent in training during the 10-year
the range of 18-59 weeks, with individuals
around a 5070-6070 chance of participating
ante in the period.
period
from any
in some
. For all categories of labor-market
experiences,
rac.e+thnic differentials
steadily diminish with higher Ievels of education.
The differences
become
negligible for college graduates.
. me evidence supports the familiar concept of life-cycle wage growth, which
implies more employment
and higher wages as individuals
age and
Not only is there generally more
accumulate
labor-market
experience.
employment
in the second 5 years of the 10-year period following school,
6-15
but a greater
Tbwe effects
6.3
Numbers
amount of this employment
takes place in high-paying
diminish with higher levels of education.
of Episodes
The above findings
market activities
on the cumulative
tell us who the primary
they te~ us litfle about the process
participants
h the low-wage
staying for short periods
long periods?
jobs.
amount of time spent in the various labor-
participants
involved
in producing
sector there because
at the number
address the remainder
labor markets,
these experiences.
but
Are
they enter the sector many times,
each time, or do tiey ente_r the sector only a few times and stay
we address
In the next step of our analysis,
section by looking
are in low-wage
part of tiis question
of spells in each of the five labor-market
of this question
in this
sratuses.
in the next section where w.e desctibe
We
spell
durations.
Tables 6.3-LQ
and 6.3-M
report percentiles
for the number
the five labor-market
statuses.
results first presented
for the entire 10-year period after men leave school followed by
measures
These tables are organized
of spells that occur in
for tile two 5-year periods
tables are constructed
describing
the number
making
analogously
up this horizon.
The three divisions
in the same way except for one important
of spells in the second 5-year period.
to Tables 6.2 with
difference
Specifically,
counrs spells hat begin in the second 5-year period and does not include
progress at the s~rt of this period.
employment
For instance,
spells in the second 5-year period
6-16
consider
tie number
for college graduates.
in these
in the part
this table only
spells in
of high-wage
The zero entries in
the table at the 10th and 25ti ~rcentile
a high-wage
population
job, but, instead reflect tie fact that a significant
holds a high-paying
Examination
job throughout
of Table 6.3-LQ
the number of spells experienced
low-wage
do not imply fiat 25% of this group did not hold
spells hdicate
experience
experience
holding
group experiences
both a low-wage
groups.
of high-wage
The representative
Very
employment
during
spells is four over
person from any
horizon,
and no spelIs of
spells than the higher education
dropouts
groups,
with the
to 2 as typical for college
graduates.
A comparison
of the results across
difference
in the distribution-
of “the number
education
group other than high-school
race-ethnic
groups shows no substantive
of spells over the 10-year period for any
dropouts.
Even for this group, the differentials
are quite small and only occur in the tails of the distribution
nonemployment
episodes.
In the light of the differentials
6-17
for
and graduates;
job at the. same time: .The lower education
falling from 5 as typical for high-school
in
only 10% of the high-school
spell over the 10-y=r
and a high-wage
groups have more nonemployment
number
The typical number
one training
The medians
dropouts
more tian 4 entries into low-paying
for all edumtion
groups
zero is typical for college graduates.
many spells in low-wage jobs;
the 10 years after leaving school.
education
labor market activities.
that two spells is typical for high-school
few individuals
the 10:year horizon
most of the 10-year period.
in the various
for those with some college;
or graduates
of this
shows a smooth pattern across education
one is common
dropouts
propordon
for high-wage
and
noted in the previous
discussion,
it is not surprising
that hey
to see that Blacks are the disadvantaged
have fewer high-wage
Inspection
quintile
definition.
spells.
of Table 6.3-M
definition
of low-wage
indicates fiat the patterns
enlployment
Not only is the relationship
education
groups,
the number
group in the sense
are duplicated
of the distributions
of speUs is almost
identicd
observed
for the Iowest-
in the minimum-wage
similar across race-ethnic
for this definition
and
of low-wage
employment.
Summarizing
employment,
the results of Tables 6.3 as hey apply to pafliciPation
in low-wage
the basic findings. are:
. Individuals
typically experience very few entries into low-paying
employment
during the first 10 years after leaving school. Even for the
lowest edumtion
groups, only 50~o of individuals
start low-paying jobs 2
times or more during the 10-year period; only 10%” start such jobs more
than 4 t.bnes.
6.4
tingths
of Episodes
To firther
our understanding
of the dynamicprocess
involved
in labor-market
mobility,
this section “investigates the dl]ration of spells in each of the five labor-market
statuses.
Tables 6.4-LQ
percentiles
statuses
describing
and 6.4-M,
structured
similarly
the length of spells associated
over the 10-year horizon.
S~rting
to Tables 6.3,
list the
with the various. labor-market
wim the results from the Iowest-quintile
.
definition,
Employment
longer
the findings
SIIOWa simple pattern in spell lengths ticross education
spells, regardless
for individuals
of whether they are k Iow:. or high-wage jobs,
with higher levels of education
6-18
groups.
tend to be
with practically. no difference
for
the two upper education
for low-wage
Training
dropouts
we~ under 4-6 months;
spells, most notably
college graduates,
for high-school
of a year, regardless
of educational
attainment.
primarily
well witiin
2 years.
1.5 years.
long periods
tie 90th percentile
Hispanics
Even
For high-school
These findings do
of low-paying
For edumtional
attainments
Blacks spend the longest time in nonemployment
to-only one year for Whites.
Typical
employment.
show up in tf~e lengths of nonemployment
for the upper percentiles.
dropouts,
SWIIS are
short..
for Blacks widens for the lower levels of education
For high-school
and less
in the lengths. of
is surprisingly
spells are finished before
differentials
dropouts
employment
the notion of a “trap” that invokes
The race-ethnic
differential
for higher levels of educational
Most nonemployment
90% of the spells are completed
90.% of the low-wage
jobs
and 10% last longer than 12 months.
of spells in low-wage
spells last less than three quarters
at the extremes,
&roups.
last from about 5 months
12-20 months for high-wage
There is no clear pattern
spells across education
The duration
not support
to around
episodes
spells tend to be shorter
for college graduates.
nonemployment
dropouts,
employment
typical spells last in the range of 6-9 monfis
than 3 months
completed
Typical
jobs for high-school
for college graduates.
attainment;
groups.
spells.
otier
than
The
and the upper percentiles.
for Blacks is. almost three years compared
experience
nonernployment
spells that are about
.
50.% longer than Whites in the high-school
when considering
Race-ethnic
dropout category,
higher levels of educational
differentials
but the differential
shrinks
at~inment..
in the lengths of employment
6.-19
spells are not systematically
finked across educatioti
low- or high-wage
.,
employment
Hispanics.
However,
low-wage
employment
Mely
attainments.
For high-school
are not much different
Blacks and Hispanics
~spanics.
For tie some-college
longer employment
roughly
comparable
For high-wage
Blacks in tie some-college
time, they are also more
categories,
jobs.
Hispanics
the
experience
Whites and Blacks are
jobs in these. two education
categories.
are longer at the extrelne.s for Whit~s than for
with the results reported
definition.
definition
Not surprisingly,
spells are very similar
comparable
or shorter
in Table 6.4-M,
o“femployment
the percentiles
continue
in the
spells in Tables
in the two tables reflecting_~e
dropouts
the main effect of switching
duration
of high-wage
6.4-LQ
to hold for
and. 6.4-M are similar.
Arid
fact that these two
Th~ _rnedians for
For high-school
from the LQ to the M definition:.i.s to lengthen
spells at the extremes,
The same is true for high-school
For college graduates,
the basic
for the training
are defined the same way using the LQ and M definitions.
low- and high-wage
unaltered’.
longer spells in
witi tie Blacks outpacing
and they are roughly
for the Iowest-quintile
he minimum-wage
activities
the durations
these findings
identified
nonemPloYment
in low-wage
category;
Blacks and
group.
Comparing
patterns
employment,
SPIIS in both low- and high-wage
employment,
college-graduate
At the me
and college-graduate
for durations
across Whites,
are more likely to experience
(see the 90th percentiles).
to spend longer spells in high-wage
_dropouts the typical spell length in
with the lengths of low-wage
graduates
spells
and for thqse with some college.
the shift from the L.Q to the M definition
6-20
the
systematically
lengthens
high-wage
SPIIS but leads to no systematic
The conclusions
differentirds
among
broadly
race-ethnic
some-college
hold up when considering
Hispanics
jobs.
low-wage
concerning
tie minimum-wage
spells.
race-etinic
concept.
for the two lower education
The ratings
groups.
For the
no longer stand out as having longer spells in low-wage
jobs than Whites or Blacks, but Hispanics
high-wage
in the lengths of low-wage
drawn above for rhe LQ definition
groups are unaltered
group,
shif~
For co~ege graduates,
continue
to experience
the longer spells for
Blacks show up more as having shorter
jobs; Whites and Hispanics experience
more similar durations
in low-paying
employment
than appears using the LQ definition.
Considering
employment
for college graduates,
widen in favor of Hispanics
Blacks when compared
longest
to the LQ definition,
spells in high-wage
Summarizing
participation
the differentials
durations
continuing
in high-wage
and
to_enjoy the
employment.
rhe findings in Tables
in low-paying
with Hispanics
spells in
6.5” in terms of their implication<
jobs, the major findings
for
are as follows:
●
The durations of spelIs in low-wage employment
are surprisingly
short.
Typical spelIs last for less than 5 months for the. high-school dropouts,
and
around 6 months for high-school graduates
and beyond; 90~o of all spells
are completed well wifhin 2 years.
●
Efigh-wage employment
spells last Ionger than those in low-wage jobs, wifh
typi~l spells increasing from 7 months for high-school dropouts to around
12-18 monfhs for those with edu~ations
beyond high school.
_
. The race+thnic
differentials
primarily
show up in the length of
nonernployment
spells (most tiofrd]ly for the longest durations)
and not for
the length of spells in low-wage employment.
Black high-school dropouts
spend the longest time in nonemployment
speIIs with 50% of fheir spelIs
lasting longer than 6 monfhs--more
than double that of mites;
fhe
6-21
.
-—
differential
narrows steadily
for college graduates.
6.5
Durations
To and From
Participation
The results sumrnartied
jobs play in the process
for higher
levels of education
in Low-Wage
mobility;
however,
it vanishes
Jobs
above portray a general picture
of economic
until
of the role low-paying
several
significant
feahires of
>
his role still remain unexplored.
the combined
individuals
re-enter
state of low-wage
the low-wage
issue concerns
high-paying
employment
plus nonemployment
take a. long time to get their first jobs, or because
the likelihood
employment.
measures
jobs.
.
variety of scemrios,
with entering
arid with eienhlally
after the initiation
of a SWII”in low-wage
6.5.1
Until First Jobs
Durations
j.ob.after. finding
fits. into the picture of
these issues by examining
to relate sequences
first jobs,
of labor-market
smtistics
describing
with entering
lo!v-wag:
participating
enter and
starts a low-wage job.
enter a low-wage
This section addresses
designed
continually
A second issue is how
a fourth issue ii how training
Tables 6.5-LQ. and 6,5-M present summary
of duration. associated
hey
once an individual
that individuals
Finally,
mobility out of low-paying
variety of duration
employment
in
occur bemuse
sector wi~l long stays in noriernployment.
long it takes to enter high-wage
tid
One frnporrarit issue is Whether “the long durations
in a specific
a
experiences.
seveml measures
jobs under a.
labor-market
activity
employment.
The top two groups of rows report s~tistics
6-22
corresponding
to the durations
until
A
individuals
enter thek first job, distinguishing
high-wage
employment.
The results
and the three race-ethnic
population
groups.
experiencing
and the remaining
persons
summarize
columns
of low-wage
findings
classifimtion.
who immediately
around
high-wage jobs. .This percentage
of the duration
decline with education,
not systematically
ranked.
other than dropouts
roughly
employment,
Considering
However,
school,
for tiose
definition
dropouts
start employment
in
rises steadily for upper education
level$. reaching
the
Durations
until first employment
except for the two upper edumtion
groups
groups
which are
of”y”oung men ii tliose education
in the 3 uPPer education
that at least 75% are working
Blacks with 12 years of scho.gling,
paying jobs typically
wage jobs.
distributions
find their first jobs shortly after leaviIlg..school.
we can conclude
present
upon leaving
on the lowest quintile
50-60”% of high-school
The vast majority
half of all demographic
of the
immediately.
value of around 85% or above for college graduates,
generally
give the fraction
or
categories
The second set of columns
to the results in Table 6.5-LQ” relying
employment,
for the four edu=tion
enter employment
give the percentiles
who do not enter employment
According
this first job is in low-wage
The first set of columns
each first-job
the fraction of the population
whether
Keeping
grouPs s~fl
groups
in mind that
in
within one year.
those beginning
employment
find. work more rapidly than their co-unteqarts
in 10W-
who start in high-
about.5 % of the group who start in Iow-paying. jobs do Ilot find
6-23
-
work for tie first 5 years.~
Durations
school dropouts,
for Blacks.
or high-wage
.
probabilities
of first jobs in high-wage
indicates
by the end of the 10-year period.
their Black counte~arts;
not working
before 3 years.
differences
for low edu~tional
exist in dumtions
attainments,
Whites and Hispanics
school graduate
do not find employment
and some-college
graduates
ethnic differentials
before 6 years.
than for some-college.
are not ordered
fare significantly
group;
do not
even
better than
The Whites
only around 5 % of them are
groups
narrow for the upper education
Blacks lag behind,
individuals.
in the -e
Indeed, tie
5 % of them are not employed
have the most rapid job acquisition
groups.
within 3.5
until first jobs across race-ethnic
but these differences
generally
is a Iow-
and Iow-wage”etnployment
dropouts
in the dropout
for high-
first employment
only about 5 % have not found jobs before 4 years.
better than the Hispanics
Pronounced
that almost
The Hisparric
fare s~ghdy
high-school
of whether
5% and 10% of this group are not employed
sum to one for Black dropouts
levels.
Regardless
job, around 50% ‘of tie Black dropouts
yearss; between
fact that the
especially
of this length are far more common
For
in the high-
much more noticeably
college
g!advates,
way as for the other groups;
for
race:”
Hispatlics are
4 According to the results for high school graduates with low-paying employment .as first
jobs, drnost 50%” of the Blacks imm.@iately find employment and thp 90th percentile for the
remainder is 252 weeks. Thus, approximately 57. o~the total don ‘t find employment for almost
five years.
S Examini-ng information on first jobs for high school dropouts, less than 30% of the Blacks
find jobs immediately and the 75th percentile for those who don’t immediately find jobs is 170
weeks or gr=ter.
Thus at last .7 x .75 = .53 of the Blacks aren’t in their first job within 170
W-ks.
6-24
the most I*ely
rapidly
to start high-wage
jobs, and tJley acquire
slightly more
than Whites.
Examination
of the summary
wage employment
LQ definition,
increase
reported
but here
in the probability
statistics
in Table 6,5-LQ
isa systematic
For high-school
dropouts,
individuals
in the waiting
high-wage
jobs.
groups,
increase
systematic
rating
for college g~duates,
tiities. for low-paying
of durations
implied by the
there is a sharp
are shorter
to acquire
to. hold:
,..
For
low-wage jobs is
there is no systematic
occurs for the
groups
there is no
jobs.
for shifts in=”durations implied
by the LQ results continue
job rises from
for the three lower education
until first high-paying
.
attainment.
15% to about 20%.
individuals
by the M definition,
pronounced
differences
the basic
exist in
,.,
6 The” main” discrepancy between the percentiles for college-educated
individuals and
durations until first jobs that are low-wage is in the 90th percentile where 340 appears for the
LQ definition and only 1.30 appears for the M definition. AS noted in previous””~iscussion,
we
individuals only for the first few y=rs. after they lmve school.
have rJ~m for college-ducated
Consequently,
the 90th percentile for the LQ definition of 340 extrapolates well outside the
range of otir data. .
.
6-25
_.
for the college-
jobs. 6 Just tie opposite
once again “for college-educated
Even after accounting
supported
is from about
of low-
job for all education
that the first job is a low-wage
the length of time it ties
Here waiting tim~
under the LQ definition;
Most notibly,
to almost 60% for the M definition;
sfightly longer under the LQ definition;
difference
tie basic conclusions
falling for the higher levels of educational
the corresponding
the three lower education
supports
shift in the results.
the probability
around 40 %“in the LQ definition
educated
for first jobs usirig the M definition
that the first job will be a low-wage
groups with the size of the increase
findings
low-wage-jobs
—.
tie durations
until first jobs across race-ethnic
25% of tie Black high-school
jobs,
regardless
dropouts
graduate
for first jobs,
graduate
exceptiona~y
it is in low- or high-paying
and some-college
with Blacks lagging
group;
considering.
6.5.2
experience
for the two lower education
here
of the .ordenng
wage employment,
Percentage
statistics
education
category
“Percentage
for the durations
who enter low-wage
duration
associated
into some form of high-wage
groups
when
After First Jobs
until individuals
6.5-LQ
and
enter some form of high-
desig.nated_.’’Population
group
in tie
spells, and the next column designated
lists non-applicable
The remaining
distribution
in their durations
job, the third group of rows in Tables
are stricfly positive.
jobs d~s
Employment
Event, ” shows the fraction of the race-ethnic
at Zero Weeks”
Inspection
of the race-ethnic
either status h or b. The third column,
Experiencing
in the
group.
Low-Paying
Given entry fo a low-paying
6.5~M lists summary
high-school
behind to a much greater degree in the high-school
is a reversal
To and From
Hispanic
Whites and Hispanics
groups have similar experience
fwst jobs in the college-educated
Wrations
groups;
long waits until first
employment;
fare better than Blacks but not as well as Whites;
high-school
.
of whetier
dropouts
groups
columns
(na) throughout
sinceall
of these spells
in the table report the percentiles
with the ..initiation of low-wage
employment
of the
until entry
job.
of the findings
in Table 6.5-LQ
not vary systematically
across educ?tion
6-26
reveals that mobility out of low-wage
levels, ignoring college graduates. W11O
experience
noticeably
Substantial
differences
classifications
attainment,
with the differentials
Blacks.
50% of individuals
15 monfis;
employment
across race-ethnic
steadily dissipate
being relatively
the higher
small for college-
shortest.
low-paying
employment
While the typical member
or graduates,
are in a high-wage job within
of the race-ethnic
groups--as
Blacks tend to wait the longest and Whites the
of Black high-school
waits; 25% are not in high-paying
employment
dropouts
experiences
before 3.7years;
exceptionally
long
and 10% are not in such
even after 9 years.
The summary
statistics
reported
in Table 6.5-M broadly-support
from the findings based on the LQ definition.
durations
dropouts
similar amounts of time to enter high-paying
in the tails of the distrib.utioo
A small segment
employment
high-school
with the exception
75% hold such jobs within 2-2.5 years; and 90% enter such
by tie median value--waits
employment,
out of low-wage jobs,
Considering.either
entering
within 3-5 years.
measured
drawn
employment
These differentials
indicate a great deal of mobility
of some Iow-educawd
about
dropouts.
jobs than the other groups.
individuals.
Results
around
into high-wage
exist in mobility out of low-wage
for high-school
the education
educated
more rapid movement
until entry into high-paying
employment,
the concltlsions
Table 6.5-M predicts
slighdy
but most of.the. differences
longer
are
inconsequential.
The fourth set of rows in Tables 6.5-LQ
durations
associated
witi entry to low-wage
and 6.5~M prewnts-. findings for the
employment
6-27
measnred
from the time that a
person begins a high-wage
correspond
to those described
dashes reported
.
The statistics
above for durations
jS a Significant
employment,
dropouts
Whereas
75% of college
differentials
above for low-wage
high-school
median entering
Blacks experiencing
groups,
Hispanics
followed
by Blacks,
The findings
high-paying
regardless
.dropou~
to high-wage.
employment,
25% of the
Indeed,
similar durations,
slightly faster.
until participation
as they were for the
the Hispanics
in the upper two education
experience
low-wage enlployment
have the longest dura~ons
tith
low-paying
who start a high-
for the length of spells from tile stati of high-wage
Blacks have tie more favorable experiences
and Hispanic
at the 10-year mark.
graduates
until entry into low-wage jobs are not as pronounced
discussed
The
are back in low-wage jobs within about a year.
The race-ethnic
durations
jobs..
the length of time before participation
wage jgb wait at least five years before entering
employment
from low- to high-wage
the results of Table 6.5-.LQ show that the level of
“factor in determining
occurs in low-wage” employment.
high-school
i: the various.. columns
in tile tables indicate that a spell is right-censored
Once in high-wage
education
job (only status h).
the shortest dLlrations.
groups.
graduates,
Whites
jobs,. followed by Hispanics,
In the some-college
have the longest spells before entrance
White
with Blacks at the
For high-school
in low-wage
and the
and college-graduate
into low-paying
employment,
with Whites having the shortest. durations.
suppofi the general conclusion
employment
of educational
will not experience
attainment.
tilat the majority
low-wage
of people entering
jobs for quite some time,
This is Ilot to say that the possibility
6-28
of ente”iing
low-wage
employment
education
groups.
dropouts
is small in the near future for hose
At the time of the start of a high-wage
take part h low-paying
employment
individuals
within about 3.5 years.
employment
about tie duration
in Table 6.5-M.
to those obtained
timatenal
when one examines
The shift in some ~rcentiles
using the LQ definition,
Between
the durations
Low-Wage
unti an individual
group repom
analogous
from the initiation
Examination
and Training
is quite large when
but the implications
of these shifts is
Experien_ms
into training.
lists summary
enters training after beginning
results for durations
s~tistics
a low-paying
for
job; and the
until entry into low-wage
employment
of training.
of the fi.nding.s in. Table 6.5-LQ
at least 90% of individuals
years.
tie
to tie main emphasis..
Tne
measured
the
of spells from the start of high-wage
The fifti group of rows-in Tables 6.5:LQ” and 6.5-M
skth
10 years if one ignored
. .
untti entry into low-wage jobs are not altered
relevant statistics
6.5.3
Over 25% avoid
of Blacks.
These conclusions
compared
two
job, about 50% of high-school
such jobs for over 8 years, and this number would be above
experiences
in tie bottom
for high-school
dropouts
reveals that
who enter a low-wage. job wait at least a year before going
Blacks wait tie shortest amount of “time; over 50.~ are iri traig.ing within 5
While a minority
than their high-school
of bigll-school
dropout
graduates
counterparts,
(1 O%) move fnto traiting-
the vast majority
6-B
of graduates
more rapidly
take a longer
time
to enter
training.
tian high-school
graduates;
The findings
low-wage
with
some
college
typically
and college graduates
in Table 6.5-LQ
employment
into training,
.
Those
lengthen
show fiat
dramatically
otier 7“5%—oftie high-school
tie
take
less
time to enter training
even less time.
durations
bemeen
gaining
and enw
for higher levels of education.
dropouts
avoid low-paying
into
Given entry
employment
for
well over one year; 50%- avoid such employment
for well over thcee years; and over 25%
are not seen in low-wage jobs for over ! O years.
While Hispanics
tend to have the shortest durations,
the race-ethnic
tie previous
analyses.
participating
in “training ““tiait even” longer before
dropout
counterparts;
Not surprisingly,
groups.
and the race-ethnic
Dumtions
differentials
tie M definition
differences
findings
in the resul~
in Table 6.5-LQ.
wage is systematically
otier
In particular
until low-wage
category,
Iengfien
Hispanics
employment
is
for fi? c~!lege
above do not change when one
there are some systematic
when colmpared to the corresponding
the duration
until training
given entry into low-
for tile Llpper t~vo education
given enq
graduates
disappear.
employment,
~eported in Table 6.5-M
than in
jobs than their
entry into low-wage
described
longer in .Table 6.5-M
hand, the duration
these @o education
“of low-wage
high-school
low-wage
silbstantially
group
are Iess noticeable
In the some-college
virtually
While the main poin~ of the picture
considers
entering
and Blacks the shortest;
for all demographic
graduates,
White and Hispanic
this is not true for Blacks.
have the longest durations,
unlikely
differentials
in the dropout
into tiaining
gro[lps.
is fypically
shor~r
On the
for
groups in Table 6.5-LQ.. Thereis a.@ndenCY
for thesepatternsto
6-30
show also in the two lower education
case, the relationships
mentioned
educational
attainment
and race-ethnic
low-paying
employment.
6.5.4
Summary
of Findings
The essential
groups,
but it is not nearly as systematic.
in the previous
analysis
groups continue
for Durations
across the various
In any
Ieve!s of
to apply under the M definition
To and From
Low-lVage
of
Jobs
findings from Tables 6.5 are as follows:
. Approximately
50~o of high-school
dropouts start their employment
in lowwage jobs.
This percentage
drops steadily for higher education Ievets,
falhng to below 2070 for college graduates.
. Durations
until first employment
generally dectine with education,
except
for the two upper education groups which are roughly similar.
For
education
groups other than high-school
dropouts,
the vast majority of
young men find their first job shortly after leaving .school, with at least 75%
working within 1 year. Long durations
until first employment
occur for
some segments of the hig.h-schooI dropout and high-schooI graduate
populations.
. About 2.570 of BIack high-school
graduates
starting their first employment
in low-paying jobs do not begin working within 5 years after leaving school.
Regardless
ofwhether
first enlployment
is in Iow-orhigh-wage
jobs, about
SO%of Black high-school
dropouk
donotfind
employment
within 3.5
years, and between S%-1070 are note.mployed
before 6 years. About 570 of
Hispanic high-school
dropouts
d-onotfind
jobs before 4yearsafter1eaving
school; and about S% of the~ites
dropouts are not workksg before3
years.
mite
and Hlspa[iic dropouts \vhoenter
lo\v-paying jobs as their
first employment
acquire these jobs more rapidly than their counterparts
who enter high-paying
jobs as their first employment;
this is not true for
Blacks.
. mile
pronounced
difference
exist in the durations
race+thnic
groups for Iow+du&~tionaI
attainments,
marginal
at higher levels of education.
6-31
until first jobs across
the differences are only
Concefing
findings
the durations
support the foflowing
from low-wage
employment
to high-paying
jobs,
the
conclusions:
. There is a great deal of mobifity out of low-wage jobs with the exception of
some Iow+ducated
Blacks.
About 5070 of high-school dropouts
or
graduates
entering low-paying employment
find a high-wage job within
about 15 months; 9070 find such employment
within 3-5 years.
Wile
tie
typid
members of race+thnic
groups wait s~ar
amounts of time to enter
high-paying
employment,
a component
of Black high-school dropouts
experiences exceptionally
long waits; 2570 are not in high-paying
employment
before 3.5 years, and 107o are not h such employment
even
after 9 years.
. Mobifity out of low-wage jobs does not vary systematidy
across edumtion
levels, except for college graduates
who experience noticeably more rapid
movement into high-wage jobs than any other group.
. Substantial
different=
exist in mobihty out of low-wage employment
across
raethnic
groups at lower levels of educationd- attainment,
but these
differentials
steadily dissipate for the higher eduation
groups and become
inconsequential
for college graduates.
The findings on rhe durations
leads to the following
from high-paying
employment
to low-wage
jobs
conclusions:
. Once in high-wage employment,
individrrrds with higher education
are fikely
to stay out of low wage employment
for much longer than less+ducated
men.
~ereas
7570 of the college graduates who start a high-wage job
wait at least 5 years before entering low-wage employment,
25% of highschool dropouts are back in low-wage jobs within about a year.
. The findings generally support the yiew that the major~ty of people entering
high-paying
employment
will not experience low-wage jobs for qu”ite some
time, with only the members of the two Iower education groups having
much of a chance at all in the near future.
Measured from the beginning
of a high-wage job, about 50~o of high-school dropouts take part in Iowpaying employment
within about 3.5 years; over 25% avoid such jobs for
over 8 years.
6-32
6.6
Linking
Work
~lstori~
An important
employment
us to ascertain
experiences.
reported
activities
tie results reported
tie importance
of any potential
end up spending a signifimnt
and nonemployment
unfavorably
dissipate
early labor-market
in explaining
experiences
low-wage
experiences
pky in
above provide ody fimitcd insights
in akcring
subsequent
6.2 tirough
compounding
employment
6.5 do not aUow
effects of labor-market
this section addresses
questions
such as:
amount of time in low-wage employment
that continue to
for a long time into the future, or does the influence
early labor market experiences
rather rapidly?
How important
tie race-ethnic
differentials
of
are the different
found in the
analyss?
Tables 6.6-1 tirough
cumulative
the findings
following
dropouts;
various
hypothetical
6.6-4 summarizes
of low wages.
for tie som.e,college
results for college graduates.
6-33
the
statuses during
2-year work experiences.
Table 6.6-2 gives analogous
Table 6.6-3 lis~ statistics
statistics describing
spend in five labor-market
in tiese tibIes rely on tie “LQ” definition
for high-school
graduates;
6.6-4 present a variety of summa@
number of weeks that individuals
tie ti~ree years immediately
results
betw=n
because they stafi fife with a bad set of experiences
experiences
previous
tie relationship
in Sections
To M in fils part of the picture,
Do individuals
influence
The findings
of early kbor-market
For ins~ce,
governing
Activitiw
mobility is the role that prior labor-market
activities.
into tie importance
experiences.
Labor-Market
aspect of the process
and economic
future employment
and Fu@re
Table 6.6-1 reports
findings for high-school
mtcgory;
AU
and, finally, Table
The rows in the tables consider five
—.
Nor-market
in ststis
activities:
t; high-wage+
low-wage,
boti,
which sums cumulative
statuses h and b;
tinirrg,
which represents
the 3-yesr period;
nonemployment,
Iow-wage+nonemployment,
.
which refers to cumulative
weeks over 3-y~r
hotion
weeks over 3-year horizon
cumulative
weeks classified
in
in activity e in
which signifies tie totrd weeks spent in status n; and
which totals the number
of weeks in statuses
f and n during
the three yws.
=ch
tible considers
ex~riencti
by an individual
designation
of tiese
six scenarios
regarding
the kbor-market
activities
in tie fust two years of the career after leaving school.
2-year work histories and their cotiesponding
definitions
are:
Base Case:
75% of 2-year period spent employed
Weeks of employment equal in the fust and second year
No training in 2-year period
AU employment in two years in low-wage jobs
More
Base Case, except only 25% of 2-year period spent employed
n:
Remnt
n:
Base Case, except all employment
none in the second
The
occurs” in the first year and
Training:
Base Case, except individual
in the 2-year period
Early
Base Case, except low-wage experience only in the first year and
all employment in second year in high-paying jobs
f:
No e:
These tables report
succeeding
participates
Base Case, except no low-wage
period
statistics
summarizing
the experiences
experience
ii” 10 weeks of traifiing
at all in tie 2-year
in the three years directiy
the 2-year work histories specified above, with the 2-year period assumed
6-34
to
end in an exit from a one-w~k
fich
CCUin tkese tables presents
tie columns
fit
focusing on different
set of coknnns
market activities
number
SWII in nonemployment.
resrrlts for Wkites,
aspects of tie distributions
gives the fraction
of tie population
at any time in the 3-year horizon.
of weeks spent in the various activities
cohunns
Bbc&,
md Hispanics,
of cumulative
that pardcipates
report tie 10th, 25tk, 50th, 75th, and 90tb percentiles
weeks.
The
in the five labor-
The second set presents
over the three years.
witi
the average
The remaining
for the simulated
distributions.
6.6.1 Relationships
Betieen
Comparing
the education
a prominent
previous
tiduccs
predictor
experience
Whereas
in total employment
especially
only marginal
constint
a lowering
Employment
prewrrted in tie tables for each of
experience
of total employment
irrespective
experience
of whetier
over the 3-year period (i.e.,
groups, movement
is
tie
shifts in total employment.
increase
in
from the “base
Holding
in the first two YWS, the primary
the total
consequence
more recent O.e., moving from “recent n“ to
“base case”) is to shift future employment
contrast,
md Future
shifting from tie “base ~se”. to “more n“
for the lower education
of past work experience
of making this labor-market
scenarios
future employment,
is recent or not.
a sharp decrease
Experience
tiat tie total amount of previous labor market
of individuals’
case” to “recent n“ produces
amount
Work
tie first three Iabor-market
groups revds
nonemployment),
Previous
slighfly from low- to high-wage
jobs.
in the first two years indtlces a decrease
6-35”
In sharp
in
—
employment
levek
over tie subsequent
of ducstion.
4-month
three years, with dre effect diminishing
A one-year
rise in employment
incr-
in experience
over dse subsequent
3-morrth increase for high-school
graduates;
h the fist
with higher
No’ years impties:
three years for high-school
about ,a 2-month
a
dropouts;
a
rise for those witi some
.,
,
college;
and only a one-month
dramatic
for those individuals
education
group.
educational
periti,
m~es
titde difference
of the 90th percentiles
of nonemployment
increase
in experience
or training over the following
other than college graduates.
For those experiencing
a shift of experience
The effects are even more
the highest amourrK_ of nonemployment
reveals that a one-year
month rise in employment
3-yw
for college graduates.
experienchg
Comparison
attainments
increase
in any
across
lads
to a 4-5
3-year period for all groups
tie most nonemployment
over the
from early in the first two years to”later in this period
in the toml amount
of nonemployment
experienced
over the future
3-year horizon.
6.6.2
Effects
of Training
and Composition
of Past Employment
on Future
Work
Activities
The comparison
training
horizon,
of the “training”
impties only marginal
groups,
with the base case indicates
in employment
and it does very little to shift employment
For the lower two education
increases
increases
scenario
training
over the subsequent
from low-wage
hat
higher
3-year
to high-wage jobs.
during the 2-year work history signals slight
in the amount of training done over the future three ymrs of the working
6-36
~eer.
No doubt,
reflec~
our defition
tils positive influence of past tining
of training wtich
as we~ as pure educationrd
training has essentially
y=
incorporates
on fiture
vocational
training
training
For tke upper two education
activities.
no effect on any labor-market
activities
in part
while employed
groups,
incr-d
during the subsequent
3-
horixon.
Comparing
tables ~.e.,
“wly
past labor-market
graduates
the base case with &e !ast two labor m~ket
#” and “no f”) helps determine
experience
and beyond,
paying employment
switching
primarily
wage jobs .ti subsequent
college graduates.
of past labor-market
implies m increase
months during
from low-wage
previous
tie role of aftcnng
to high-wage
labor-market
While this substitution
tie composition
in the future,
from low- to high-
substim.tion
of high-wage
for lowfor
dropouts,
a shift
high-wage
employment
amounting
to approximately
This increase in employment
the Iargmt amount of nonemployment
of
For high-school
is even larger for high-school
in average employment
the group who experience
fismd in *e
wi~ tie effect almost nonexis~nt
expene”nce from Iow-.wage to purely
the 3-year horizon.
jobs.
experience
implies a near one.to-one
time spent employed,
s~narios
dso
i-2
is not larger for those in
(i.e.,
those at the 90tfr
percentile).
6.6.3
Summary
of Findings
for Work
Histories
The results in Table 6.6 suggest that much of the differences
race-etinic
groups
work histories
in the earlier discussion
on future ex-periences,
refleck
across the
a c.ompo~mding effect of Previous
rather than an unalterable
6-37.
observed
disparity
inherent
in the
groups.
years afir
leaving school for M individuals
resulting
Tables
6.6 assume tie wme work history at the end of the fit
race-ethnic
differences
in labor-market
ethnic groups are much sm~er
LQ and 6.2-LQ
.
experiences
shows that tie difference
either of the 5-Y=
the means or the percentiles
for the combined
sta~s
10%). 7 Comparing
the differentials
either of the 5-year periods
when compared
between
to the differentials
considered,
observed
presented
for rhe different
time spans.
differences
observed
the race-ethnic
across
the findings
Tables 6.1-
hotion
when
(compafig
either
reveals differentials
that Vpically
even at the extremes
of the
in Table 6.2-LQ
for nonemployment
dropouts
h
are enormous
given any reasonable
For the three upper education
categories
dropous
Whites arsd Blacks in Table 6.6-1
in Tables 6.6-1,
for these groups are even less pronounced
In summary
Enmining
nonemployment
White and Black high-school
adjustment
findings
of low-wage+
status_ of low-wage +nonemployment
Similarly,
and tie
between White and B1ack high-school
fa~ weti below 6 montis for any scenario
distribution.
6.2.
periods making up the 10-y=
above
attainment,
over tie next three years across ram-
than those noted in Sections
is more than a year for the combined
considering
of a given edumtiod
two
groups,
in Tables 6.6 compared
than for high-school
the
to earlier
dropouts.
from Tables 6.6 allow us to draw the following
conclusions:
●
The main factor predicting the amount of future employment is the total
amount of past labor-market experience, irrespective of whether this past
7 For ymrs 1-5, the 90th percentile differs only 29 weeks between White and Black high
school dropouts, however this reflects the fact that a large fraction “of the Blacks are truncated
at 260 weeks which represents the entire 5-year period.
6-38
experience occurred
eruplo~ent.
recently or whether it occurred
me fink between previous work experience
diminishes with higher Ievek of education.
in low- or high-wage
and future employment
A large component of the race+thnic
differentials observed in the amount
of emplo~ent
and the fraction of this employment spent in high-wage jobs
reflects the accumulation of unfavorable experiences early in the work
career that have a compountlng
effect on subsequent labor-market
activitiw.
While it may be tempting to interpret
approptite
to do so.
suggest an attempt
amdysis,
Some of our statement
that tie empiri~
in character.
market history applicable
labor-market
into low-wage
activities
labor~market
experiences
throughout
from a nonstrrrctud
causal relationships.
in the @tire;
labor markets,
it predicts
as well as the conditions
finked to continued
is not ,a basis for infernng
Prediction,
however,
findings
do indicate how early labor-market
the dynamic
cannot formally
scenarios
relationships
model
labor-
model
the routes followed by these individuals
low-wage
and they identifi
statistical
with this history will spend in alternative
and it firther
While our empirical
tils study are
Given a particular
these markets and to return participation;
markets.
in our
up to some year after leaving school, our empiriml
the lengths of time that individuals
it is not
tie results in this section may
results presented
They are derived
that does not provide a basis for inferring
predicts
summarizing
to make a causal link between
but we recognize
purely descriptive
the above findings in cauml context,
predicts
6-39
in
the routes of escape from
caumfity.
be given a causal interpretation,
link to subsequent
associated
participation
employment
with participation
they
activities,
in the low-wage
sector.
An elaborate
behaviod
interpretation
of our findings
most of the questions cited in tie previous dismssion.
reader beheves
exercised
hterprebg
that a behaviod
in using our em”piritil
tie esdmates
Ii*
is required
is not needed to answer
For those questions where a
for m answer,
tie same care must be
findings to arrive at an msswer as one wou[d apply in
of multiple regression
6-40
models
h a behavioral
context.
7. Conclusion
The findings of this study offer many insights into the role played by Iow-wage
jobs in tie overaU employment
a useful source for addressing
subsequent
discussion
pic~re
of youtis,
seveml impo~t
undertakes
and a synthesis
questions
such a syntiesis
of Wese resul~
about tiis role.
in an effort to answer
provides
The
the foI1owirrg
questions:
.
Among young men, who participates
. How much does earnings
less stable for workers
@ Is participation
in low-wage
Disputes
h low-wage
employment
over tie answers
over government
a port-of-entry
minimum
wages.
markets?
or persistent?
into high-paying
jobs?
are often at the heart of many public debates
program
presumes
earnings
from tiese jobs contribute
examples
The main rationale
that participation
against minimum
to income?
in labor markets.
One of the most prominent
arguments
in low-wage
sectors tempo~
to these questions
intervention
sectors?
from low-wage jobs contribute
. h employment
.
training
in low-wage
concerns
for raising minimum-wage
in low-wage jobs is relatively
a..significant
wage legislation,
wages in such legislation,
policies dealing
rely on the supposition
7-1
rates as an”anti-poverty
permanent
fraction to household
or for the inclusion
with the setting of
income.
and that
The
of sub-minimum
that the hourly wages earned on
.,.
in a career are not nearly as important for new kbor-force
jobs wly
oppo~~ties
mobfity
for ~ese en~~
~ acqu~
me work exwrience
in tie labor market.
Answering
the above questions
information
opposhg
7.1
for establishing
the vatidi~
parties in minimum-wage
hong
debates.
Ymrng Men, Mo
All demographic
school graduates
hem
in hw-Wage
groups have some experience.
individurds
employment
graduates
followed
for high-school
by high-school
graduates
High-school
dropouts
in Iow-wagejobs;
dropouts,
during
in tie
high 30%.
with college-educated
The larger amount of low-wage
spends a considerable
during tie 10-year period.
their employment
and high-
rate of college graduates
re!ative to dropouts
dropouts
by
tie most time in low-wage jobs during tie 10-
less experience.
of the dropout population
employment
maintained
Sectors?
Wgh-schooI
The pticipation
experience
having considerably
nonemployment.
critical
in the order listed.
sector during their first 5 years after school is.a suwtisingly
year period after school,
education,
provides
have a 75%-80 %‘ chance of holding a low-wage job at sometime
High-school
segment
mat leads @ upw~d
suppositions
We consider
Participates
tie frost 10 years after Ieaving school.
low-wage
of tie alternative
entrants as are tie
and graduates
amount
average
Approximately
this percentage
reflecfi
hat a signifi~nt
of time in
1-2 years of. low-wage
50% of high-school
dropouts
steadily drops for higher levels of
falling to wefl below 20% for college graduates.
Measurements
designed
to gauge the extent of low-wage
7-2
employment
in any
start
population depend critidy
the age of individuals.
on tie time fmrne used to register participation
For high-school
dropouts and tigh-school
pardcipation
rate in low-wage employment
participation
as holding any low-wage job. in a 2-y=r
sma~ as 11% when participation
quarter
when individtis
Regardless
graduates,
tie
is as large as 59 % when registering
period in tie late teens, and it is as
refers to holding a low-paying
job during an ~bitrsry
are in their kte twenties.
of the age and education
of young men, participation
is twice as Wely to occur over 2-year horizons
this horizon.
as weu as on
respective
tian in any pardcular
of the time frame used to cablog
in low-wage
quarter
participation,
m~ng
low-wage jobs declines with age @y as much as 50%),
with tie most
significant
decreases
As individuals
age, education
a stronger
factor de_@mining the lkelihood
Race-ethnic
differences
at the earlier ages.
iri”low-wage
college graduates,
wage sector,
participation
differentials
7.2
are inconsequential
ordering
longer horizons
minor at young
at lower leveIs of edu=tion.
between
rates in the low-
Whites and Hispanics.
Race-ethnic
for college graduates.
How Much Does Earnings
The fraction
are relatively
Blacks tend to have tie highest participation
with no systematic
becomes
of holding a low-wage job.
ages, but they tend to widen at older ages especially
Ignoring
up
the Wefihod
of holding
occurring
jobs
from Low-Wage Jobs Contribute
made up by earnings
for accumldating..income
to Income?
from low-wage jobs falls as one c~sider:
and broader
7-3
sour:=
of incoine.
Earnings
from
.
low-wage jobs is a relatively high fiction
of individuals’ labor income eared
during the
quarter during which these jobs wire held; it is a much smafler fraction of toti
income received over 2 y~rs.
For high-schooI
dropouts
and high-school
as much as 86% of toti..bbor
farnify
graduates,
earnings
from !ow-wage jobs average
income during tfrose
qu’-rs
during which such jobs were held; this average drops to as Iittie as 30% when
.
calculating
~riods
tie contribution
dutig
analogous
of these earnings
which low-paying
percerrmg~
to to@I family income
employment
occurs.
For col!ege graduates,
men who hold low-wage jobs,
~mings
on average for 65 %“-80% of to~l family income received
labor income alone, the fraction of earnings
of these earnings
considering
the
participants
to quarterly
the contribution
fr_oln these jobs account
in a“quarter,
family income received over 2-year periods when low-paying
poorIy-educated
over 2-year
are 71 % and 15%.
For less-educated
individuals’
received
and 40%-50%
of
jobs are held ... .Considering
coming from low-wage jobs for
ranges between 69 %-88 % ““when considering
labor income; and it ranges between
to total labor income received
the contribution
33% -6”1% when
during 2-year
periods when
low-wage jobs are held.
The contributions
of earnings
from low-wage jobs to “any measure
any period declines with age and with higher levels of education.
associated
with aging is much steeper for higher edu=tion
tie fraction of any measure
typically
of income made up by earnings
highest for BIacks, with Hispanics
levels.
of income over
The rate of decline
At the lower levels,
from low-wage
jobs is
second, and with Whites having the lowest
7-4
fractions.
7.3
Rcc*thnic
differentials
k Employment
A centi
labor-market
predicdng
&s
employment,
occurred, remntiy
or whether
eduwtion
past employment
it occurred
future employment.
future employment
diminishes
ethnic differentials
observed
employment
7.4
h Participation
paying employment
groups,
previous
levels.
labor-market
Sectors
of mobility
“trap”.
any
the stabifity of
Iabor-mmket
experience
A large component
Temporary
out of low-wage
of the race-
and the. fraction of this
of limited employment
earIy in
~S a
or Persistent?
jobs.
The evidence
Young men spend relatively
jobs during his- peri~.
Individuals
does not
litfle time in low-
typically
experience
Even for fie lowest education
jobs 2 times or !nore during the 10-yearperiod;
7-5
and
activities.
in the first 10 years after school.
only 50% s~rt low-paying
~thin
not time spent in low-wage jobs--that
in Low-Wage
very few entries into low-wage
jobs.
factor governing
in the amount of employment
tie notion of a low-wage
is the main factor
in low- or bgh-wage
at higher education
in future
of whether this past experience
spent in Klgh-wage jobs reflects the influence
There is a great dd
support
irrespective
The link beweeq
effect on subsequent
Labor Markets?
tie ro~e of work experience
is the primary
careers--i. e., time spent out-of-work,
compounding
in hw-Wage
The total amount of past employment
the extent of futun
group,
for college graduates.
SWbIe for Workers
finding in this study concerns
activities.
individuals’
are hsignifiat
only
Further,
10% start such jobs more than 4 times.
employment
are surprisingly
high-school
dropouts,
short.
“
school;
the majority
tisn 4 yms
well within 2 YWS.
1-2 years of low-wage
spend less tisn
h low-paying
dropouts
fmd a high-wage
job within
15 months;
While the typid
members
of race-ethnic
exceptionauy
employment,
90%
dropou@ and graduates
in toti
during the 10 years after
and ordy 10% of Wls popuktion
a component
or graduates
10% are not in such employment
entetig
spend more
low-paying
90% find such employment
of Black high-school
even after 9 years.
spent in low-wage
dropouts
employment
employment
within 3-5 years.
groups wait similar amounts
long waits; 25% are not in high-paying
not reflect long periods
and beyond;
jobs.
About 50% of high-school
high-paying
graduates
High-school
employment
1 yeu
of spe~s in low-wage
spells last for less than 5 months for the
and around 6 months for high-school
of all speUs are completed
average a modest
Typid
the durations
of time to enter
experience
before 3.5 years; and
Most of this wait, however,
employment,
instead it reflects
does
much time
spent out of work.
Mobility
out of low-wage
jobs does not vary systematically
levels, except for college graduates
high-wage jobs.
,
across race-ethnic
dissipate
Substantial
who experience
differences
noticeably
exist in mobility
groups at lower levels of education,
for tie higher edueation
across education
more rapid movement
out of low-wage
but these differentials
groups and become inconsequential
graduates.
7-6
employment
steadily
for college
into
H &ere is any notion of a “trap” in the Iabor market,
spent out-of-work
along with low-paytig
of nonemployment
relevant
*an
individuals
time spent in Iow-payirig
in the subsequent
in tie combined
alone for judging
stages of theti working
careers.
smtus of low-wage jobs and out-of-work
levels of education
status more
the prospects
of
The extent of participation
d-shes
sharply for higher
and for older workers.
~ss+ducated
Blacks spend an exccptiondy
and tie experiences
better.
a notion of a kbor-mwket
employment
drne
Time spent in the combined s=*
employment.
and low-wage jobs provides
it must incorporate
for some poorly-educated
More “*an 50% of Black high-school
large amount of time in ~s
White and Hispanic
dropouts
dropouts
s~tus,
are not much
spend more tian 6.5 years in low-
paying jobs or out of work. d.gring the first 10 years after school; and 10% spend more
than 9 years.
About 50% of Hispanics
wage employment
White dropouts
spend about half of tie 10-yesr period in low-
or out of work; and 10% participate
fare bettir
than Blacks and Hispanics,
almost 9 years in these statuses.
and yet 10% of Whites spend 7.5
years or more @ low-wage jobs or out of work.
The large amount of time spent in this combined
and nonemployment
experience
acquire
reflects tie influence
early in the crsrmr which predicts
labor-market
employment
primarily
experience,
are significantly
state of low-wage
of insufficient
labor-market
low future employment.
even in low-wage jobs, @eir prospects
enhao.ced.
7-7
employment
When individuals
for fiture
.
7.5
k bw-Wage
Incr=sing
tie future.
,
Employment
The eviden=
supports
in low-wage
the work experience
and additional
The acquisition
the famihsr conmpt
sectors as offering
of Iife-cycle wage growth> which
individuals
of work experience
in low-paying
ties
jobs.
The findings generally
paying employment
fraction of the high-school
wage employment
in the not-too-distant
M% of high-school
hand, over ~%
dropouts
Whereas
dropouts and graduates
low-wage
entering
become involved
are in low-paying jobs within about a year..
in lowjob,
On the other
avoid such jobs for over 8 years.
75% of tie college graduates
low-wage
employment,
jobs within 3.5 years.
low-paying
of educatiomd
high-
From time of entry into a high-wage
future.
attainment
the length of time before participation
before entering
of peopie entering
low-wage jobs for quite some time, though a
Once in a high-wage job, educational
determining
jobs in the
of this future employment
support the view that tie majority
will not experience
to acquire
jobs at eartier ages not only
in the future, but a greater amount
place k high-paytig
the opportunity
SWS needed to move on to higher-paying
irnpfies more employment
nontrivial
Jobs?
employment at any wage implies more time spent in high-wage jobs ti
depicts pardcipation
fu~re.
a Port of Entry into Wgh-Paying
employment
is a significant
occurs
who stirt high:wage
in low-wage
groups have broadly
after working at a high-wage
attainment.
7-8
employment.
jobs wait at least 5 years
about 50% of high-school
Race-ethnic
factor in
dropouts
are back ti
similar experiences
job, regardless
in
of the level
7.6
Concludhg
Remark
The two definitions
very low thresholds
similtity
of low-wage
findings based on these two definitions.
both of our wage thresholds
accept as a low wage.
movements
used in our empirid
While our definitions
in minimum
for tie close
Expressed
may be relevant for considering
wages, many would define tie relevant concept
one were to ,use a higher tireshold
imply
in 1990
fall below $5.50 per hour, which most peopIe would
Iabor market using a much higher level for hourly earnings
answers
anaIysis
for tie hourly wage rate, and this factor may account
in our empirid
doks,
employment
in our empirid
.
of a low-wage
than adopti
analysis,
propo=d
in this study.
it is quite possible
If
that the
offered above would have been quite different.
No matter what dimensions
plays a critid
role in neutilzing
The relationships
between
fumre employment
~tegories
graduates.
prospec~
previous
Education
spent in high-paying
race-ethnic
tie major factor influencing
employment
After education,
eduation
or array of experiences.
and the characteristics
differentials
becoming
is the single most important
jobs.
in our study,
the higher tie level of edtication.
with differences
for the extent of fu~re
attributes
labor market experiences
experiences,
atninments,
considered
any disadvantageous
become less pronounced
of labor-market
higher educational
of”work activities
steadily
negligible
factor in. improving
of
For au
diminish
for
for college
individuals’
and for We fraction of his employment
the to~l amount of time spent
future employment
7-9
prospects.
employed
is
.
References
Amemiya,
T. (1985):
Advance d Econometrics.
Cambridge:
~s ha tic M
Bartholomew,
D.J. (1982): Sto
NY: John Wiley and Sons.
Harvard
ocial Pr
Press.
, 3rd d.
New York,
“Estimating Dynamic Random Effects Models
Bhargava, A. and J. D. Sargm (1983):
from Panel Data Covering Short Time Periods,” in~s,
A. B.
Atisorrand
F. A. CoweU, eds. London:
London School of Economics.
Cameron, A. C., R.M. GriWand T. MaCurdy (1989): “TheEffectso fUnemployment
Compensation onthe Unemployment
of Youths,” Report to U. S. Burmuof
Wor
statistics.
Chamberlti,
G. (1984) :-”’’Panel Data,” in Handbo ok of Econometrics
Gfiches
and M. Intriligator, eds. Amsterdam:
North-Holland.
stems,
Howard, R.A. (1971): ~
John Wiley and Sons.
Kohen,
A. L, H. S. Pamesand
J. R. Shea (1975):
and Middle-aged Men, “ in~,
Smith, ed. New York: Columbia Universi~
Lflard,
L. (1983):
“AModel
VOIS.. I and..~.
“kcomeh stabilitya
Vol. II, Z.
New York:
mongYoung
J. D.
Press.
of Wage Expectationsin
Incomes, A. B. Atkinson and F. A. CoweIl, eds.
bbor S~lpply, ” in Panel Data on
London:
London School of
Economics.
Liflard,
L. and R. Willis (1978): ‘(Dynamic Aspects of Gmings
Econometrica,46,
985-1012..
Mobility,”
MaCurdy, T. (1982): “TheUseofTime
Series Processes to-Model the Error Sticture
18, 65oftings
in a Longitudinal Data Analysisl” Jou mal of Econometrics,
102.
Sctiler,
B.R. (1977): “Relative Earnings
Economic Review, 67, 926-941..
Taubman,
P. (1975):
Sources of IneotIalitv
Mobility
in the United States,”
of Eimin~s.
Amsterdam:
American
Nortll-Holland.