Have Learned
What Researchers
from the National Longitudinal
Surveys About Youth Unemployment
U.S. Depaflment of Labor
Lynn Matin, Secretary
Bureau of Labor Statiatica
August 1992.
Repofi
a2a
w
Preface
Data from the National Longitudinal Sumeys of Labor
Market R~~rch (NLS) are “sd intensively by private
and government r~mrchers and the academic and busin=s communities: ~Is repofi iummarizm some of that
r=earch, with special reference to the mployment problems of minority and disadvantaged youth. The NLS are a
collection of five smeys. They ae Young Mm who were
14-24 yews old in 1966, Older Men who were 45-59 in
1966, Mature Women who were 30.44 in 1967, Young
Women who wme 14-24 in 196S, and Youth who were
14-21 in 1979 which include both sexes.
This report was written by David K. Howe, a summer
titem, and Harley J. Fmzis, an economist in the OffIce of
R=each and Evaluation, Bureau of Labor Skfisfim. Rita
Jain of the same office astited ti its prqaration.
Information in this rqoti will be made available to sensou impaired individuals upon requ=f. Voim phone
(202) 606-STAT; TDD”phone (202) 606-5897; ~b Mmsage Refemal phone 1-8W326-2577.
Mat~ial in this publication is in the pubfic domain and,
with appropriate credit, may be reproduced without
pemission.
Contents
Page
Tables:
1. Employmmt status by sa,. race, and Hispanic oritin,Spring 1979 .
2. Employment status byage, Spring 1979 . . . . . . . . . . .
.....
3. Selected characteristics of unemployed youth by race, Spring 1979 .
4. Percent distribution of employment during the sumey yea by race,
weeks worked, andnumber of jobs held, 1971 and 1980 .,,,,....
5. Summa~ of research in distribution/turnover s~tion
...
v
.......
..........
...............
.....................
2
3
3
.
................
. . . . . . . . . . . . . . . ...
4
5
Introduction
Unemployment rates of youth typi~lly gr=tly exceed
those of other workers. Therefore, it is interesting to examine the characteristic of unemployed youth, the length
and frequency of their spells of unemployment, and the
importance of the unemployment ciperimce on them,
both in the short mn and the long run.
This r~ofi summarizm some of the r=mrch that uses
the National Longitudinal Sumeys of hbor Market Experience @LS), with special reference to the employment
problems of minority ihd disadvantaged youth.’ The National Longitudinal Sumeys of bbor Market Experimce
are a collection of Lve surveys. They are.:Young Men who
were 14-24 in 1966, Older Men who were 45-59 in 1966,
Mature Women. who were 30-44 in 1967, Young Women
who were 14-24 in 1968, and Youth who were 14-21 h
1979, which includw both sex=.
Because of the large sampl= of youth and be=use NLS
respondents have been surveyed once evm y=r or two
over an extmded period, these data are well-suited to examining the Iong-mn consequences of youth labor market
experience=. In. panicular, the 1979 NLS Youth Cohoti
(NLSw contains tiee~y work histories detailing mch respondent’s labor force status, hours worked, and emPlOymmt at more than one job, pemitting” analyses that are
not possible with other data series. (Note The words
“sumey” md “qohort” are used interchangeably in thts
rqofi.)
All of the younger r~pondents were askd about their
work =periences, their family income and assets, their
parenk’ education, as well as various attitudinal and demographic charactefistim. The NLSY includ= week-byweek work hlstori= as well m gr=t detail on other topi=.
The surveys ovemmpled several groups of particular
interest. The original NLS sumeys during the 1960s oversampIed blacks. The NLSY ovmsampled blacks, Hispanics, and economi~lly disadvantaged wh]t=.z Few studies
using the NLS focus on”poor youth, although many examine blacks and whit= separately.
The first s-tion of this report gives an ovemiew of the
genmal characteristics of unemployed youths. The next
s~tion discusses iwue relating to the duration and incidence ofjobl=sness =ong youth. The third section surveys the literature on the consequmm of youth jobl=sn=s. Section four discusses the.longer tem consequence
of youth unemployment. In the fifth setion, job se=ch
strategi= of the young are discussed. Sction six provides a
brief conclusion.
—
Youth Unemployment: General
Characteristics -
In 1980, unemployment among 16-24 ye= olds was 13.9
perant, while those over 24 smtatied rates ofrrnly 5.1 percent. ln 1990, the comparable figures wme 11.1 percent
and 4.4 percent, r-p=tively.> The NLS allows resmrchem
to mamine the demo~aphic traits of youth experiencing
difficulties k the labor maket.
Many factom Emit youths’ abifity to work. Schooling, of
course, mkes full-time employment cliff]cult, at Ie=t outtidc of the summer month. Those who have left school
mn find the trmsition b&ttieen school and work rather
daunting. Some of the difficulties me outlind by Re=
(1986): Many youths have not developed an extensive network of job conticts, and some me unfamiliw with the
limits of acceptable behatior on the job. This faifing can
result in a youttis tirly dlsmissaLq Volun~
quits me
also higher.
“The higher rate of volunta~ quifi h= two quite
sepamte sonrt=. Fimt, youth are still trying to dscovw what they want to be when they grow up, and
in the procas tv many different job<, some of which
do not prove congenial... Second, youth living with
employed parents often work to emoney for a
specific purpose—to buy a mr or to take a tripand
oftm l=ve the hbor force for a time when this shotitem goal has been reached.”s
This weak attachment to the labor market letis economists to pay considerable attention to the qu~tion ofdlstinguishing different labor market situations. Thwe ticlude the following
The unemployed are persons who are not working but
are actively looking for work md are available for work.
Those out of the labor force are persons who are neither
working nor actively seaching for work. The nonejnployed
orjobless are those without a job. Tti mtego~ is more inclusive and includa those out of the labor force. The latter
group includes discouraged worker$ thQsewho would like
employment but are no longer =tively seeking it6
The labor force participation rate equals the sum of the
~plOyed and the unemployed dividd by the relevant
population.
Stice attmpts by youth to find work we often s~mdic,
the distiriction between being “unemploy& and out of the
labor force is not as clwr as it is for mattire..adults. Thus,
much of the recent literature US= the more ticIusive
masure of nonemployment rather than unemployment.
Flinn and H&kman (1982), however, emphasize that
there are behaviorally mmtingful difference betwem
those out of the labor force and the unemployd, evm for
youth. Nevetih>less, th~ note that, “.. job sarch activity
,occum in both stata...tbe dflermce between the two
states is only a matter of deflw ....”7
The 1979 NMY (16-21 yar-old subs-pie)
revds
much about the group’s labor market behatior. Table 1
provides data on the mploymmt status of various groups.
Blacks md H1spani= have lower labor force patiicipation
md higher unemployment than whites. Young women
dso =perience relatively low padicipation and high unemployment. Table 2 shows the trmtion into the winkingworld as youths age Not su~risingly, as young people
become older more of them find jobs and unemployment
mtes decreases
Table 1. Employment status by sex, ra=, and Hispanic origin
Spflng 1979
Per=nt
unemployed
Employ menl[
population
ratio
19,1
57.3
67,6
?4, 1
65.3..
60.1
71.0
20.7
17.6
36,5
41.2
36.0
53,6
61.1
40,2
35,4
45,4
61.8
53,4
70,7
72,4
70.0
74,9
23.2
24,7
22.0
15.~
17.5
14.4
47,5
40.2
55.1
60,9
57.7
64,1
*
Female. .
Male .. . . . .
Black ..,....
Female . . . .
Male, ,. ...
Hispanic
o rigid . . . . . .
Female.
Male, ..,..
White . . . . . . .
Female.
Male . . . . . .
1
Universe: ~viNans, age 16-21 on intemiew date. (N=24,580.000).
Source
SantOs (i981 a).
Table 3, from Sates (1982), d~crib= the share of
16-to-21 -y=r-olds repofihg
unemployment in 1979.’
tiost
Wf of the unmployed are high school students;
80 percent live with their parents. Minoriti@ suffm &spro~tionately
from unemployment-a black yOuth is mO=
aS a w.~t+but,
than twice as Wely to be unmployti
stice the majority of the population is whit% l= tb hdf
of the ~employed are membem of a racial minority. Of
those whine family income is bown, 22 ~rmt
of the unemployed are pwr, compared to 15 prcmt of the youth
population.
2
Table 2. Employment SW-by
fi” oermnt)
Age
Total
16-17 . . . . . . . . . .
18-19 . . . . . . . . . .
20-22 . . . . . . . . . .
Source
I
I
Employed
57
44
61
68
to have higher unemploymat rat= than white youths,
mostIy due to higher turnover, although longer duration
also plays a role.
Among whites, low= tenure at a giva job,smms to explain why younger whites have higher probabilities of unemployment than olda white accordkg to Leighton and
Mincer. When tenure is taken into xcount, age plays a less
kpofimt
role. Young white mal~ who have been
age, Spring 1979
I‘nemp’eyed
I ‘“’::F
l’412g
I
17
13
10
I
39
26
23
~plOyed by. a firm for up to 5 yas have similar chances
of unemployment as older white males with the sme
tenure. Young blacks, however, retaio a high incidence of
unaployment
(compared with older blacks) even after
tmure is t~en into account, espaially among those who
have spent less thm a y=r with their employer.
Calculated from Santos(1987aJ;tables2,1 and 2.2.
Mnomkts
such 88 Freeman Md Wise (1982),
Feldstein and Ellwoti ( 1982), and Rees ( 198@ have found
that unemployment is ody a serious problem for the small
propotiion of youth who stay out of work for extended
periods of time. Hdf of %de teemge unemplo~ent,
for
emmple, ocms among those who are out of work for over
6 months ~Is group consists of Iws thm 10 print
of the
youth labor force. NonempIoymenrspells tend to be longm
and va~ by mce Holzm (1986a) repofis the avemge
duration to be over 10 months for whites ad over 13
month for blacks.
In an effoti to shed light on the truly disadvantage,
Bores (1984) sepamtd out a subsmple of “hard-core unemployed,,. whom he defined as thow who (1) were out of’
school, (2) fived on their own or fivd with parents whose
income was below the povefiy level and (3) had bem unemployed for at lwt 10 weeks. By that definition (w.h]ch
Bores ackowledgm
as necasarily arbitra~) approximately 1 in 11 of the unemployed youth were ;hard-core,,
in the spring of 1981. They tendd to be “...older, more
likely to kve patiicipati
in trainin~ to be marnd, to
have children, to five in the central city of an SMSA, and to
live in an =ea of high unemploymmt than was tme of all
unemployed youth.,’ 10
Unemployment
spells
Too
often
or too
Table 3&e[etied Characterlstiw of unemployed youth by HIs
panic origin and race, Spring 1979
on percent)
Characteristic
Total
ilsDanic
Vtite
31ack
Sex
Female
Male
53
4?
47
53
54
46
51
49
16-17
18.19
20.21
44
34
23
46
29
25
45
34
22
40
35
26
Enrollment status
High school
dropout
24
36
45
44
45
46
10
10
10
9
21
11
22
22.
80
81
78
86
Away at college,
in dormito~
3
1
4
2
Has own
dwelkng
17
18
18
12
17
61
22
34
50
17
10
67
23
33
46
21
Age
Mgh school
student
College
student
Differences h“ jobl~sness or nnemploymmt ratw ca”
be analyzd by distinguishing betwem two factors Duration—how long spells ofjobl~sn~s Iast—md incidenceti
how oftm thq occur. Researihen disagree on which
aspect is more impofiant. Their etiination
of different
periods md use of diffeient masure (unemployment ~
Household status
At home, with
parents
OppOsed tO nOnemployment) makes comparison of thtir
results problematic.
Leighton and Mincer (1982) mphasim
incidence.
Using the data from the 1966-71 NLS s-eys,
they repofi
that hlgb job tumovm explains why unemployment mt~
among male youths me higher than those among mm 42 to
62 Y=KS old. Duration actually rduces the age dltTerential: Older ma tend to have longer unemployment spells.
This explanation applies to the &fference in uncmp]oy.
ment rate betwem adults and both studenti and nonstudents, whether white or black. Black youths were found
Povetiy Stal”s
Poor
Nonpoor
Not available
Universe G“ilians aged 16-21 who were unemployed
dale (N = 3,300,000),
Source
3
22
!
Nonenrolled
tigh school
graduate
long?
23. .
Santos (1962).
o“ i“ter”iew
Table 4. Percent distrlb”tio” of employment during the suwey
year by mce, week%worked, and number of jobs held, 1971 ati
.-..
,...
1971
1980
Total number
of weeksljobs
Wtite
White
Black
Total employed
(thousands)
287
1,503.
296
1,955
Total (percent)
00.0
100,0
00.0
100.0
Weeks
Consequences
3i.2
68.8
f8.1
81.9
40.8
59.2
22,6
77.4
Jobs
1
2+
47.4
52.6
61.4
38,6
60.5
>9.5
52.,8
47,2
Weeks/jobs
1-38
1
2+
3.8
27.4
5.8
12.2
20.9
20.0
8,9
13,7
43,6
~
55.6
~
39,6
~
43.9.
33.5
1-38
39+
39+
Table 5 summtiz= the literature suweyd in this s&tion. It is mpting
to conclude that Vhde incidence
seemed to produce most of the resdts for the origtial Whofis of young men and young women, duration becae
the key fact,or by the time the 1979 NLSY was conductd.
Such a conclusion is not completely wammtd, however,
stice =rtier studi- usd unemployment dza from the
NLS as a measure while more recent studia tmd to use
nonemploymmt dat6 from the NEY %a masure.
1
2+
~arching
unemployment
ssarring
For youth in general, there are a variety of ways in
which emly jobl=sn~.s ~n Med their future prosp-ts in
Three
~conomic
th~~fi~,
~ch
bor.
the labor ~arke~,13
rowed from the broader unemployment fitemture, identify
these eff=ts. Thehuman capital mode[sugg=ts that the
knowledg% skills, and discipline msociatd with working
should enhance prospects for both hlgber wage and lower
unemployment. Ushg the Young Men’s Cohort, k(1976) =timated that approximately one-third of a young
worker’s total compensation was delivered in the fom of
human capital. This reason alone might make =rly jo~
l&n6s up=ially unfortunate, as =rly work experima is
thought to entail intensive on-the-job training. If oppotiunities to acquire mly human mpitil are fleeting, theeff=ti
of youth unemployment would be =p&ially
persistent.
Dual labor market theorists beheve that unemployment
“scars” its victims, leading fimt to dlscouragemegt and ultimately to decfining work habits and a sucmsion of job
with little payorupwwd mobility. Thus, mlyentv
into
the primav swtor (the one with d=tiable jobs) is critical if
one is to stay out of the cycle of d=d-end joh =sociated
with theseconda~ swtor. Adherents ofthis theory note
that potential employem make extensive use of job historiw when hiring applicants. Those who have experience
substantial unemployment emly in their careem risk being
tagged as poor workem.
h contiast, search theov emphasiz= the beneficial mpects of youth unemploymmt. In this tiew, youth unmplowent allows those new to the labor market to %ther
valuable itiormation shut thtijob prospects. Seen iri thk
way, job search -n be as much as m “inv-tmenl” as ontbe-job training, by allowing wo!ke~ to switch into more
lucrative positions at other firms.
Stevenson (1978a, b, c), Raefin (19S0, 1981), Becker and
Hills (19S0, 1983), Ellwood(19S2), Grcoran (1982), HLlis
(1989, Lynch (1986) and Ballen md Frman
(1980 all
address the smrnng hypoth~is. With the exception of
Hills (19 S5), none of the empirical work in three stud<=
attempts to explicitly dlstiriguish between the human mpital and dual labor market/s*g
hypoth==. Inst=d,
they simply a~ whethw youth unemployment d~r=s=
wage or future employment prmpats and, if not, whether
youth unemploymat in fact advmca labor market op-
Universe;. Males. ages 18-21 as of the beginning Of the surveY Year,
who were not enrolled in school, not in the mititaw, and who were
employed al least i week during the suwey year.
Source
of youth
and
Pollard (1984).
the older NLS cohorts with the Young Men
Comparing
and Young Women,s coho~, Frmk md Freeman (1978)
reprt that the higher unemployment of the younger
cohort3 redts from gr=ter job turnover. Within the
younger cohorts, however, the authors emphasize the importance of duration in explaining unemployment.
tirfier NLS sumeys were compared with the 1979
NLSY by Pollard (1984). Both white and black youths experienced rising unemployment and joblesm=s between.
the two surveys. For whit=, Pollard concludes that increass in joblwsness r-ulted from=rising turnover. Black
joblessnas, on the other hand, resulted from longer nonemployment spell% if anything, turnover among young
blacks decfined over the d~ade. (See table 4).
Recent studia using NLSY data have tended to focm
on duration. A group of NBER studies in 1986 compared
the NLSY data with a specially commissioned suryey of
black inner-city youth—what we will mll the NBER sample.J1In one of the studla, Ballen and Freeman (1986) selected a subset that was not k school; they found intensive
evidence of lengthy nonemployment spells. Twenty percent of their subsmple had been without work for a yew
or longe~ those never employed accanted for over half of
the sample’s weks of total nonemployment, Duration is
thus the dominant factor, both in comparisons between
NBER blacks and NLSY whh= and between NLSY
blacks and whites. h imphmtion of these r=tits i8 that a
va~ing propensity to lose jobs do= not drive. the blackwhite aployment
differatial so much...as the difflc~ty
young blacks have in findktg an actiptible job in the fi=t
place. 12
4
I
pofluniti=. ~efin (1980) uses a subsarnple focusing mreclusivelyon dimdvantagd youtk Bdlen md Freemn
study black inner-city youth; the other authors distinguish
merely between black and white subgroups. All of thae
studies focus on indlviduds who are out of school.
These studl= imply that those individuals with =rly labor market dtT1culti= tend to have gr=ter unemployment
later in their working fiv~. This finding does not nec=szily imply a musal relationship, however. It is @ssible
(even fikely) that some third factor or set of factors (e.g.,
education or aptitude) muses both the early ahd later
problems. In panicular, workem who are less productive
may have trouble in the labor market all.throughout thti[
lives, but that would not necwady
imply a scarring effeck Both arly and later aployem may simply recognize
less productive workers whm hiring and firing. If this is
the -se, mrly unemployment would be correhted with
later unemployment, but it would not follow that the former ~used the latter. 14
Both cross=sectional and longitudinal data sets =n address ttis issue by controlling for such characteristics as
yam of education and pychologiml
indicators that are
thought to be comelated with worker quahty. But longitudinal data sets in addition have the pote”tiaI to control for
since tiey track tie sme tidiunobse~ed cbacteristics,
viduds over tie.
exist for males who were originally defind as “out of the
labor force, “ i.e., out of school, without a job, and not
actively sexctig
for aployment.
Usfig a different method, Wefin (1981) foun{no relationship between youth untiployment
aid either future
wage rat= or occupational smtus, although unemployment did lead to dwlines in job sa~fwtion, 16Using a subsample of disadvantage youth, Raelin (1980) reported
StatUS,
that a variable summarizing data on..occupati0nal
and hours of the respondent’s fimt job si~ificantly
wagaffected future wag= and occupational status.’7 Unfortunately, the latter study made. no attempt to control fol
other factors.
In contrast to the above r=ults, Becker and Hills (1980,
1983) found support for the semch hypothesis. Thoseexperiencing nabnemploymeqt but who changd jobs more
frequently =m@ higher wag& y“ars liter. Longer spells
of unemployment, howevm, did appez to depress wag=.
Thee resdts are broadly consistent with Raefin’s finding
that part-time status (which allows both sarch and the acquisition oFjob sk14s) is assqcisted with higher future pay
rates and occupational stitus.
Racial br=kdowns of data presentd by Becker and
Hills (1983) are striking. Blacks who experienced no unemployment as twnagas and who changd jobs 1-s than
two time hti even lower wages than blacks who experienced unemployment spells of 26 weeks or longer. Blacks
who remained employed but changed jobs two tire= or
more fared worse than” those who had unemployment
spells of 5 weeks or fewer, indl~ting that job sarch associated with unemployment may be more beneficial than
s=rch that occurs while employed. Whhes also appeared
to benefit from job search, although not as dramatically as
blacks. It is also notable that whites who were unemployed
for over 25 weeks were ptid more than blacks with any
Studies using rhe NLS Young Men k and Young Women k
cohorts. Stevenson ( 1978b) repotis that out-of-school
whites who were employed during the sumey week in 1966
@md approximately $2W more per yer a d-de
later
than out-of-school whites ,who were unaployed
at the
time. For out-of-school black maks, the improvement
amountd to $7W a “YW. For out-of-school females, the
differences
about $9W per y~r.15 Greter differentials
●
Table 5. SUrnmaV of research In duration ftirnover secfio”
Author
,
.
Measure :
(nonemploymentl
unemployment
Leighton
and Mincer
(1982)
Unemployment
Frank and
Freeman
(1978)
Unemployment
Pollard
(1984)
Pdncipally
nonemployment
Phenomenon
Database
NLS
NLS
to be explained
Explanation
(incidence
or duration).
Wtite age differential
Black age differential
Racial differential
incidence
incidence
incidence
Age differential
Variation among youth
incidence
dura~on
NLS
incidence
duration
Bane” and
Freeman
(1986)
Nonemployment
D<fferentia!: NBER inner cily
Blacks vs. NLSY whites
NLSY racial differential:
NLSY
5
duration
In addition, due to the economic downturn during the exly 198Vs, the NLSY dIows us to evaluate the eff=ts of unemployment in a w=ker labor market.
h gened, recent malys= uskg the NLSY have tended
to assmiate long spells of arly nonaployment
tith subsequent l~or market problems. HNs ( 1985) prducd
results that app= at odds ~th his prior work with Becker.
While @rlier afiicl= had linked Io@er peritis of ititifl
unemploy mat with lower eventual wage among white
maim, the later papm revded the opposite (at lat for unemployment spells of 23 week or 1-s). Su~risingly, in
contrast to the implications of search theoq, job chang=
did not I&d to signifitintly higher .mings among the un~ployed. The effect of longer perioti of initial unaPloYment for black male and all females wme negative but insignificant.
The difference may be due ti the specification of the
model in H]lls’ 1985 paper (though waker labor markets
during the erly 198Vs may have erodd the advantag= of
sarch). In addition to a m=me of reemployment dufing
197.9, Hills also incIuded a variable indicating the we~
worked since the respondent turned 18. Thus the unem.
ployment m=sure used by HilI pr~umabiy refl%ts “sarring” rather than lmt human capital. This latt= m=sure
w= h]ghly significmt. One plausible ht~retation
of
Hills,s re”lts is that although early jobl=m=s do= have
long-lasting effects, a given spell of unemployment, if it is
r=sonably shofi, do= not prove epecia~y crippling.
Thus, the 1985 paper is consistent with human mpital
models, which strss the tiportance of =lY work m~rience, although the impact of Iost experience is 16s severe
than some might imagine. The criti~ factor app-s to k
lost working expefimce ooblmm~s or nonemployment]
s opposed to the fmstration of s=rchlng for new work
(unmploymmt).23 On the other hand, Hills, evidence k
also consistent with the interpretation that long peritis of
nonemployment can “tag~ one as a low quality worker. .
In their study of nonemployment duration” and labor
market turnover, Ballen and Fr=man (1986) compared
the NBER wple
of inner-city black with data from the
NLSY74 Nter controlling for unobsemed chmacttistiw,
they found that longer spefls of nonaployment
lower&
marginally the inner city bI=ks’ probability of mploymmt, but did not app=r to affect either blacks or whit= in
the NMY. Inter~tingly, long periods of mployment signifimtly reduced NLSY blac~ md whit@, probability
of nonemployment, but did not ‘improve the NBER
group’s chances. Wd
evidence suppo~ing incidence dependenc+lage
numbem of spe~s of nonmployment
l=ding to subsequent nonemployment—w~
preented
for the NBER sample. Data from the NLSY, how,ever, did
not support the existince of incidence dependenm.
Llmititions Of the NBER data set led BalIen md Freeman to pumue more eridence. They intemiewed employers in a primarily black dktrict of one city to win fufihm
levd ofjob mobtity. Btiat=
were based on recessions
which mntroUd for du=tion,
experience and various
dmogmphic vtiable. 18
Two ~icl= by BIau md Ktin (198 la, 198lb) point towads a poxible rmondation
of the semching and scar*g MewPints. One afiicle a“ddr~s~ the consequences of
quits while the othm studl= layoffs. This resemch sugg=ts
that voluntw quits Iead to improvements h wmings for
all race ad sex groups. hyoffs lad to lower wage among
reds, but not for femties.
EUWOM(1982) noted that an unobsemed facmr or set of
factom, such m three relating to the worker’s mw ~ility,
could act tidependently on both the =rlier md bter job
~tienc~
fiere maybe qualities of the worker that are
ob%mable to the potential mployer but not reflected in a
household swey. Mtemately, =rly unemployment could
tidlwte w~ attachment to the labor force, which in turn
might mmehte with lower worker productivity. 19
Mw@
exploited the fact that the NLS, like all longitudinal dab sets, tmcks the same individuals over time.
Cros-=tional
studies, in contrast, select new responden~ eve~ sumey. In Longitudinal dah sets, by examining
how btir force outcomes change for each individual after
a SWU of unaploymmt,
one can emmine the effect of
events such s spells of employment
without the confonndmg itiuence of mob=mtile
traits such as w=k attachment to the labor forw. Consequently, Ellwood could
use the NLS to control for respondents’ othemiie unob*mable trtits to the extent that the traits stayed constant
over time.’0
hatbpting
to control for such unobsemable chzacteristics, Ellwood concludd that early unaployment
huti future job prospects only mildly, ad only for a shoti
time. “Even a 6“month spell out of work tinds to generate
ody an addhional 3 to 4 week out of work 1 YW Iater.”z1
In contrmt, ‘~ly
work ex~fience hm a tiable impact
on wages. ~ntrblling for individud effects, experience in
the wnd,
third, or foutih y-r out of school tends to be
=stiatti
with wage increm~ of kt ween 10 and 20 pr=nt a yar.,,z2 (Itti= added.)
Usrng a similar method, Corcorm ( 1982) analyzed the
con~uence
of unemployment mong young females. She
found a stronger relationship between present and future
~PIOyment than EIIwood did for men, even after control.
hng for unobsewable individual chamcteristics, while the
wage &Sts of ~rly nonemployment for women are IW
tha they are for men but still appretiablc Unfotiunately,
her analysis doa not take into acmunt marriage or childbifih, =h of which muld plausibly r=ult in a volunta~
etit from the bbor force (and which obviously changes
ova time). Futih~or%
the shofin~s of the panel prohibitd ~rco~
from ~timating how long these “scars,, last.
Studies Wing the NUY The N2SY obbi”d
more dettil
on its rapndents, work experimc~ than earher sumeys.
6
.
.
insight into the caus~ of labor mket problas of tiercity youth. While 7.8percat of those em-ployem who inked
about theti hires, ” work record considerd it “a strike
against the youth if he h= a Msual work histoW~ only 28
per-t
considmed it bad “for youths to have been “out.of
work for a long pefiod of time before applying for a job.,,zs
The authors inte~reted this as evidence of “incidence dependence,’ among innm-city blacks.
In order to detemine why employment duration did not
appea to help inner-city blacks as much asit did others, the
authors exmined jobs held by this group. They found that
fewer than 20 percent of these jobs requird a high school
degree or higher, while just over 15 percmt were whh&collar. The authors sumised that they were “d=d-end jobs,”
i.e, of the type that protided fittle humm capital.
Finally, BaOen and Frmman used the NLSY data set to
study the effects of nonemployment on wa8es. Among
blacka, the number of spells mattered more than spell duration. Resdts for whites revealed a positive relationship
between the number of spells and wages, after controlling
for unobsaved characttistics.
Aside from the long-tqu~tion of “warringfl there
is the short-tern
quetion
of ‘Zdumtion dependence’,—whethw the Ionger a pmon is unemployed, the
more fikely he or she is to remtin unemployed. Lynch
(198@ used the NLSY to focus on nonemployment, which
she defined as being neither employd, in school, or in the
military. After controlling for various demographic facton, she found that longer spells of nonmployment substantially reduced the chancw of findlrig work during the
next period. Lynch writ= Where-employment probability for either a ,typical’ male or female who has not been
work]ng for 1 week is slightly gr=ter thm 30 percmt. If
they have not been working for 8 week this drops to 8 percent and if they have not bm employed for 52 weeks thek
re-employment probability is only 2 peycent.’,26Results for
a subsample of high school dropouts were much the sme
While tbe papem discussed here differ in many details,
the overall picture is r=sofibly
cleu. h the shofi tern,
nonmployment
may l-d to futher nonemploymmt,
wheras in the long run, emly unemployment has little dfect on future employment for male although posibly
more of an effect f~r.females. Long periods of =rly unemployment have a negative effect on w~ges,’”howevm. This
effwt appars to be stronger for bIacks ad feti=
than
for white mala. Counteract~rig the negative effmt or unemploymmt on wages is the positive effect ofjob s-rch—
the job mobility sometimes associated with periods of unemployment.
semch mme ticiently. me Natioml Longitudlnd Suiveys mntain numerous quetions on various faceh of job
SWCL The 1981 NLSY investigated the methods youths
med whm they Iookd for aplwment.
Wie1g6sz and Ca~ter
( 1987) r~ofied that individuals utig multiple job s=rch methods spen! less time lo.Oking for ajob. Infomal methds, spmifidly asking friends
or relatives md d]rect appfimtion b the employ-, were
jMged more effective thm fomal methods such = apply
ing through a State employment agmcy .27Among unaployed workers who wme not new entrmts to the labor
market, checking tith the local union w= pafiicularly effective. Those who were alrady employd, however, may
have found checking tith a union counter-prductive due
to seniority roles. No single method affected subsquent
job satisfaction stistmtiaUy?a
Holzer (1987a) disaggre~td
the monthly probability
of becoming employed into probabihties that ~e conditional on v~ious =ch
metho~. For both whit= and
blacks, infomal methods of job search wme found to be
the most frequently used and most productive in terns of
generating job offers md acceptanc=.’q
Holzer also disaggregate the difference between” the
probability of a black md white baming
mployed into
different= betwem. the use of different s=ch methods.
All methods of swch gave whites a geater chmce of receiving a job offer, but difimenca k dkect application to
the employer explained a patiicularly Imge share of the racial emplopent
differential. In fact, the two infomd
semch methods together =plaind 87-90 percent of the total black-white dlffermtial. “Fufihemore, vifiually au of
this reflwts Wereqm
h the abihty of th=e methods to
produce job offe=, as oppos~ to differmc- in methods
used or job acceptance rates.”
One elaent of a worker’s job smrch strategy that economists have ptid sp=ial attention to is the reservation
wage A reservation wage is the Iow=t wage that a worker
will accqt at a given point in time. The NLSY m=sures
this conc~t with two cate~ri~ of vatiabla. The fimt
notes whether a r=pondmt would acqt
a s~cifid job
(for e=mple, washing dishes) at a patiicular wage level
($2.50, $3.50, and $5.00). The second meas~= reemation wages by asking what minimum wage the youth
would accept in a new job. We refer to this secqnd m=sue
as asking wag~.qo
Bores, et al. (1981) and Bores (1982) analyzd the determinants of both me%ures of resmation wages. As one
might expeti, reservation wag- rise with age and are higher among employed workem. Females have lower asking
wages than mal~ those out of school have higher a&lng
-ges than students, probably beause many of the latter
mticipate working in a tempora~ summer job.3 L
Youths from For famifies were more fikely to accept
jobs at S2.50 in 1979 than youths from wwlthler fmiliw.
Each job categmy w= sufficiently attractive at $2.5&a
Search strategies
The qu=tion of the methods and strategy that youth use
in job search is interesting both theoretically—as a way of
explaining unemployment dumtion—.and from a policy
perspective tince it may be possible to ttich youth to
7
rate of pay that was 86 percent of the minimum wage it the
time of the intemiew—to elicit at least a 20 percent accep.
tance rate from the smple of all youths. >1
Wcial differences in asking waga were small. Howev.
er, black youth were more likely to exprms a willingnms to
work at the amay of hypothetical jobs than whitw, even
after cent rolling for various demographic variabl~. 33
But, while black resemtion wages ~e comparable or
lower than white resewation wages, the market wag= for
blacks are quite a bit worse. Thus, while unemployed
whites will ultimately receive wag= ~eater than thti r~evation wags, blacksapper to have to scale down their
wage exp=tations. Two empirical results follow: Blacks
have long nonemployment and unemployment spells and
the ~tra serch time nairows the black-white wage differential to a de~ee. Holzer summarizes th’e situation as
follows
Young blacks seek wagm comparable to those sought by
young whites but which are less available to them. B-use
ofthls, their durations of unemployment rise substantially.
As th=e lengthy durations progress, many young blach
either lower their reswation wags or accept other positions which they consider to be tempora~. Othem gain no
employment at all. Thus, the unemployment durations of
young blacks will reflect their high resemation wages
while their received wig= will do so to a leser extent.
Although tlot all elements of this scentio have been cl=rIy documented, they are consistent with the evidence that
is thus far available. jg
Intermtingly, results by Borus, et al. (19S 1) suggest that
black employment probabilities are 1=s sensitive to drops
in resemation wages than are white ernplo~ent probabili-
ties. A drop in a white male studegts’ remation
wage
from $0.60 above hIs fimputed) market wage to hls market
wage can adjust downwards hls probability of unemployment from 96 percmt to 10 percent. Comparable black
mde students experience a 17-pementage point drop in
their unemployment probabihty from 98 “percent to 81
percmt. Among non-students, the phenomeno~ is I=s
dramatic.
To summarize, NLS rmearch h= mown that blacks and
whit- appur to sarch for jobs in stillar ways, both with
regard to the search methods used and with re~rds to r~.
emation wag= for accepting a job offer. However, whit=
have more succes in generating offers.
Conclusion
The Nat ional Longitudinal Surveys have provm to be
an invaluable r=wch tool for the study of the labor market experiences of youth. The sumeys explore many
aspwts of labor market behavior, includlng -ings,
and
provide unusually detailed information on unemployment
and nonemployment, search strategies, and raemation
~ag-. The detil of the surveys md the large sample siz=
would by themselve insure the value of the NLS in understmding the labor market. In addition, the Longitudinal
nature of the survey-tracking
the sme sample membcn
over several yam—allows rs=rchus
to answer qumtions about the long-tern effects of labor market experienc-, and additionally allows resachers to control for
traits not directly obsemable that might bias the r=ults of
cross-section studies. This review has attmpted to show
the contribution that r=earch using the NLS has made in
~derstanding the problem of youth unemploymmt.
Endnotes
‘k conumt, the CPS had xpofied that only about 15Fcmtof141S
y.ar-.ld$ $veFemPl0Yed. S?.!0s(19S 1). MOsl,t.d,es .ftiela~rfor.c
ovmlmk 14.and 15 year..olds. Michael md T.ma (1984) arg.. that.
..,.students age 14a.d 15.acq”i* substantial aploymmt a“d that exvricnce is vastly tiffemnt for black a“d wh,teyo”ths.,,
9The NLSYpr”d”ced
a””nemplaymmt
rate of 19.1 Frcmt among
16-21 y=r-ol&. For comparison, ~S tits indicate a 16.1.F=enl unemploymentrate among 16-19 year old, in 1979. Handhk of tibrStnt&
de., BLSB”lletin
2217, June 1985.
‘0 BorU(1984) pp. 30,
II ~o,zer(19s6a)
notes thal <<TheNBERsufleY w~cOnd. cl&&-
1 [n fomaci.n o. NLSre$earch ispr.vided in Bzelby, elal. (1977 and
1979), bigh (1982), a“d Day”on[ a.d A”drka.i (1983). TWO CO1[K.
[i,>nsof NUabstrace
are The’\,atio”ol Lo,r8irudinal Sumeys of Labor
,Wurker E.Yper?c,rcr:A!z A””orortd Bibliogwphy and A7LSBib/iogmphy
Updotc, 1990. Bo<”s(1982a) surveys other 10ngittiinal da[a sets.
RCeS( i986) presents a“ excellent surve>, 0“ youth jobl=sn=s.
‘ Using da!, .“ family income, family size Ixation and a rnrd/”rba.
variable, ya”ths &ho K,,cd i“ fami~i~ that were below the poverty level i“
Ja””aq 1978we,e $eI%ted aspar[of the..econamhtiy
dkadva" taged
r,<,,,-Hispanic, “o”.blac~ove-pie,
R~nl[$ for.versamplcd
rmonde”ts.= \.eigh[ed when a rcpresentarive sample.f the Iargerpopulation
is “ceded. See Fra”kel, McWilfiams a“d Spencer (1983) s“d Fcankel
(1981).
3 ,V”,zlhly Labor Rcvie*,, Ja..aq, 1983and March, 1991. Dara from
Current Pc>pubat
ion S“rvq (cPS).NLS figures aftm van from CPS MtimaIw. Fceeman and MedoN(19S2b~ noIel hat different-bctweintie
two surveys may arise from the facf that wh!le [he CPS interview h-ti
of ha”scholds, #he NU inten,ie%s its youth rspo”de”ls directly.
“Rw(1986)
pp. 617-618.
5 Rws ( 1986) pp. 617. It k ako noted [hat the mi”im”m wage may &s.
<riminateagai”st youth by dryi”g”p thes”pply oflow wage, lowskiU
jobs. Rws, hox.tver, dow”plays tKtsar8”ment.
‘Some of[heamb(g.itiimplicit inthenotion ofinvol””taq
““em“Iovment and. bv exlension, di$coumsd workers are diwused i“ Sum.
me;s (1986).
7Flin. .nd Iieckma. (1982), pp. 19
Iwem November 1979 md May 1980 among 2,~
yo””gbl.ck
men.
.g& 16 through X, who wem 1ivin8in che i..er-cit i= of Boston, Chica.
go,. a“d PfiIadelphia. The intemims were limited to i“habtta”[s of city
block with..atl~t 70 percent black reide”ls and 30 percc”t familie hav.
i“g incomes below the pov~y K“.. The q“~tions i“ the S“RCY fm..d
on che r=po”dmts, daily ac[ iviliw...; thtir f-ily backgro””d% IhGrjob
sach behavior md experience, including r=emation wags..; their =trespective work tistoti
sforthcpreedi.g
12mo”t~...
The”ti”lnS
of the NBER sumey masuppleme”t [o the NLS, which pmvid- lhc bulk
of the black-white cmparimn% fiainits focus o”nofihemi..mtils
blacks...; a”d thedir%t co.mparability ofmmyof
theq.mtionsi”
the
NBER S“FVW to those k the NM (after wtich some of the fomm were
,!‘:*~a:/;ri
mdFr~man.(1986)
pp. 85.
ad
titisani
(I 983) ad Lei& (1982).
13 See Daymo”t
8
Small mmple size phgued many of th=e st”di-. For .-PIc
EUwod ( 1982) andB=kerand Hills, (1980, 1983) .$mple $US ragd fmm
91 to 217. See B_ow (1982), pp. 289. Mili@V ind”ctiomdutis the fietmm W= wone”ed attrition ram of the 1966 cobofi,
2’ Ellwti> PP. 383.
22 Ibid.
23 To {f~ti”g~
betwm
b“tie~
qcle a“d model s~ificatio.
tif-ti, m exmt r~Eation of Bwker a“d Hills, (1983) paper, usin~ NLSY
&b, mi8ht prove interesting.
24 Note ~hat ..ptcntjally ,~ti.”s” m-remmt
problms ~d $mall
=mple sizes phped this pan of Bane” and Fr-”>s
study. Ball.. and
FEmm,
pp. 89.
2s Ballm and F,-.”,
PP. 95-96.
% Lynch (1986), pp. 16-17.
a8m.i=,
17Ot~erfomal methods incl.de “Si”gprivateemployment
sch.ol p[acme”t off,ws,hbor unionhin”g hallsa“d neqapm advwtiseme”ts,Holzer(1986.)producedsimila %tits.
28,o~ ,atj%faccion
is primmly aff~ted by vocasion,e~i~g$. ad
g?qob ~cepta”ce, .,. ~~,”red .$ thepeme”tofjob offersaccepted.
= Freman(1978)discusses
the“seof s“~-tive vtiabl~ i“ economic
2..1”.;.
:“TfB:m$ (1981.).
3’ Bo,”$1482b).
to this ge”emlization w-e that
33 Bom$ ( 19s I~ I g82b) Exqtions
bhck$ te”dd to h. more reluctant to “Clea” “p ntighborhoti
and
“Work away from home al a national forat or par~ than whites.
3’ Holzer (1986b), pp. 173.
References
Ah,
Ad
V. ad Ma”g”m, Gtih L. 1978, The Li”zen.”~ Cfiis of
Yo.rh Unemplqmen(, Wamzm,
MI: W.E, Upjohn I.stit”te
for Employmmt %e~h,
hdtiti,
Paul J. 1978. “The Estabhshmt
of Stable and Sncc-sful
bploymmt
-m
me Role of Work Attit”dm a“d Labor
:y~
Knowltize.- ~ U.S. Depaflme”t of bbor ( 1982,), pp.
BOVS Michael E 1980& ‘,K”owltige
et d (1980), pp 340-343. .
of the World of Work?
Bores, Michael E, 1981.. “Introd”ctia”.,,
a“d Wise (198ti),
Bor”s, Michael E 1982a. ‘<A. I“ve.toq of tingitudinal Data Sets of Interest to Economists?
Review of fibric
Dora
uxe.
10(1-2):113 .126.
pp.
&ker,
Brian E. a“d Hills, Stephen M. 1980. “Tee”age Unemployment:
%me Evidence of the Lo”g-m. Effects on WaZm.X,Jour”ol ofH.mon Rmo”mes 15(3} 35+372
kker,,
Brian E. and Hills, Stephen M. 1983..-The Lo”g-m” Effects of.
Job Chmg~ a“d Unmployme”tamensMaleTemgers.,,Jour.
“cl of H“mo” Re,ourccs 18(2) 197-212.
In Bores (1981 b), pp 1-17,
BOW. Michaei E,, ed. 198 lb. Pathwaysto the Future:A repor! on the Notionar Lon8ir.dinar SUm,y of Youth Lob., Mo,kel Experience in
1979,Columbus, OW @nter for Human Resource R~mrch, The
Ohio State University.
Bdle”, Joh md Fmma”, &chard B. 1986. “Tra”titio”s betw~” EmPlOymml a“d No”emplomat?
1“ F,&ma” and Holzer (1986.),
pp. 75- I 12.
Bmow~8&3fiWS 1982. “Comrnentfl 1“ ~-”
In Bows,
Bows, Michael E. 1982b. ‘.Wilti”fl~
to Work Among Youth.,,Joura./
of Human Resotirces, 17(4}58 1-593.
Bor”s, Michael E., ed. 1983. Tomormw’s Workers. Laingt on, MA Lexin~ton Book.
&r”$, Mrchael E. 1984.. ‘<A De$ctiptio” of Employed a“d Unemployed
Youth in 1981> 1“ Bor”$ (1984 b), pp 13.55.
Bor”s, Michael E,, ed. 1984b, Yo.zh and {he Lobor~arket:Anaryss fmm
the Natio”ul Lon~itudi”ar S“mey, Kalamazoo, Ml: W.E. Upiohn
lnsti”le
for Employment Research.
Bor.s, Michael E,; Crowley, Joan E; R“mberEer, Russell W.; S..10s, Richard; and Shapiro, Datid, 1980. “RSmrch on Youth Employment and Employabi fity D.evelopmmt: Fi”drn@ of the Nat io”al
Longit”di”al S“WV of Young Americans, 1979? Youth K“nwl.
&ge De.elOPmmt R~oti 2.7. Washington U.S. G..om,nt
Printing Off,ce. PB80-203217.
Bielby, Wilfiam T.; Hawley, C. B.; and Btlls, David, 1979 ,’Re%uch USof the Na!io”al bngitudiml
SUmWs.,, R & D Monograph 62
uS. Depanmcnr of Lb,.
B!an, Fra”cine D. a“d K&n, bwrme
M. 198 la. ‘.CSUS ad Cons..
que.c= of ~Yoffs” Economic Inquiw. 19(2):270-296.
Bor”s, Michael E.; Kim, Choong$~o, and Sates, Richard. 19S1. “Job
Se=ch Activides of Youth.,> 1. Bor”s (198 Ib), pp 223-276.
Bla”, Fra”ci”c D. ad Kahn, bwre”ce M. 1981b. ..Race a“d Sex Differ.
a=
in Qtits by Young Worker,.,, Indusz,iar ond Labor Rela.
[i.”, Review. 3~4)563 .577.
Corcomn, Maw. 1982. ‘<The Employment a“d Wage Co”seq”mces of
Tec”aEe Womenss Nonernployme”t.,, 1“ Freernin and Wise
(1982a), pp 391-419.
9
Damont, Thom~ N. a“d &drismi, Paul J. 1983. “The R=mrch ff=
of the National Longitudiml S.WW: .“ U@ate.,, Re.iew of &b.
Ii. Data fJ$e.11(3):203-310.
Ho1m, Haw J. 1987b. “Job s~mh by Emp]oyti md UnemPloyti
Yo”th~, t“d~triulazd tibor Belatio”s Rmiew. ~4}@1
-611.
tie,,,
Ellwoad, Darid T. 1982. ‘.Ttemse Unaplo~mk
P-mat
Scam or
Tmporav
Blemitie?”
k Frwma” and Wise (1982.), pp
349-305, . .
Edwzd. 1976. ‘.A8e> Experience md Wage Gmwth,”Ame,fcon
Em”omic Review.66(4> 548.558.
high. D“...
E. 1982. ‘~e
SUWV of Recmc
1W3}185 .201.
FeI&t&”, Main and E1lwo&, Darid T. 1982. ‘,Te”#ge U“employ ~,~:, What is tie Probla?
In Freeman md Wise (1982a), pp
Natio”d Lo”8buti”al Sumw
Evidence? Review of @bfic
A Stl-tive
~ra
Use
Leighton, Linda and Mincer, Jacob. 1982. “bbor T.movm a“d Youth
Unmployme”t.,, In Fr~ma” a“d Wi= (1982s), pp 235-270.
Fli”n, Christopher J. ud Heck.”,
lam= J, 1982 ‘~tie Unemplaymmt
a“d O“t of the Labor For= &&vi.&y
Di$ti”ct kbar Force
StatmV Jou,”al of Labor Economics 1(1}28 .42,
Lynch. Lisa M. 19s6. “Tb. Youth h~r
M=ket in the 8& Detcmim“ts afRe.mployment ~ob.b~fhim for Yo”.g Me” a“d Worn.
en.,, National B“r~u of Economic R_ch
Worti.g Pap S..
rie, No, 2021. Gmbridgc, MA NBER,
Fmnk, Robefl H. and Fr=man Kchud T. 1978. ‘~e D,strib”tion of the
Unemploymmt Burden Do tie ht Hired Luve Fint?” Review
of Ecmomiwand S,ottiti& W3)380.39 1.
Frankel, Martin. 1981. “Sampli”s, Wdghti, a“d Sapling
BOmS (1981 b), PP. 19-58.
Emors:
Michael, Rohn T. a“d T-.
Does Life Bean at
2(4>464.476.
b
NLS motatd
Bibliography 1968-1989. 19S9. Glumbus Cmter for
Humm Resource R--ch,
The Otio State Univenity.
Fra”kel, Martin, McW%ams, Harold A. and Spencer, Bruce D. 1983.
“Natio”al Longit”&”al S“mey of Labor Force Behavior, Youth
S.wey (NM), T%hnical Sampling Rtpofl.,, Chicago, IL: Nation.
al Opinion R=czch Cmtm, Uni.ctity of ~cago.
Freeman, Richard B. 1978, ‘.Job Satisfwtio” z a“ Economic Vsriable7
A,nerican Economic Review. 68(2} 135.140
Fr_a”,
Nanq B. 1984. ..You& tiplommc
16Y Jou,”al of tibor Economim
NLS Etbliography Update 1990. Colwbis
Cater for HRes=rch, The Ohio State U“ivetiry.
Rsmrce
NE Handbook 1990, Col”mb~ Gnler for H”.ma” Rwnr~
The Ohio Staa U“ivetity.
R_ch,
Olsm, Randall J., 1989. The Future of tke NLS:A Reptifrom che ,VSF
Co”fe,ence on tke Fur”m of rhe NBand the NQ Tmknicoi R+
,,i,w Commirfee. ti”mb”s,
Ohim Cmler for H“mm R~”rce
R-h,
Ohio State U“i.ersity.
Richzd B. md Holztr, Ham J., eds. 1986a. The Black Youth
E,nplaymenI Crisis. Chicago Univwsity of Chicago Prms, 1986.
Fr=man, Ricbmd B. and Holzer, Harm J., ~, 1986b. ‘The Bbck Youth
Employment Critis, Summary of Fbdin~? In F,-.”
a“d HolZ.F ( 1986.), PP 3-20.
Par”=, Herbm S.a“d K.hen, Andrew I. 1975. ‘.Occ”p.tio”d
Ifiomation a“d Labor Market Stat”% The *C of Young Mm..- Journal
of Hum=” Re,oumes. 10(1~W55,
Fremm, Richard B. ad Medofl, Jme L. 1982. “Why Do,, the Rate of
Youth bboc Form Activity Differ Acrms S“me~W In.Fre=afi
and Wke (1982a), pp 75-114.
Pa%more, Darid L.; Ay, UDd; Rmkel, Shevl; Wade. Barb-;
Jam=. 1983. ‘,H-hh a“d Youth Emplowmt.-App/ied
iCS 15(6):715-729.
Freeman, Richard B. md Wis% David A., eds 1982. The Youth tibor
Market Problem: In Nature, Came. and Comeqze”em. ChicaEo
U“iversiry of Chicago Pres.
P.Uard, Tom K. 1984. “Cha”S= ovc the 197VSin TheEmployment Pattms of Black s“d ~ite
Ymns Men:- In Bo~s ( 19S4b), PP
57-79.
Frema”. Rich=d B. and Wke, David A., ~. 1982b. ‘.The Youth Labor
Markti Problem lti Nature, Ca”s~, md ConSq”e”ces.,,
1.
Freemn a“d Wise (19828), pp 1-16,
Raefi”, Joseph A. 1980.Btii/di”Sa Coreez Be EffcczofInil!alJob fiperiexces a“d Related Work A tt;~ud= m tire, EmploJme”L Kda.
mazoo, MI: W.E. Upjohn I“stit”te for Employment Rw&.
Hills, S.tephe” M. 1985. .$The~ects of High Uriemployrne”t i“ the Late
197LYsa“ [he WaEm of Young Me” and Women; Wortig PaPer.
Columbus, OH: Ce”rer for H.ma” Rsoume Resmrti, The Ohio
Stale U“ivetily.
-.E”,
JOumaf of Ec*
R“mberger, Rus.11 W. 1983. ,The Influence of Family BackSound on
Education, &i”@,
md W~kh,,> SocfaI Forcex 61(3)755-773.
Holzw, Harry J.. 1986b. “RS_atim
Wa8m and their Labor Market Effects for Black and Wbi[e Male Youth.” Journal of Human ReSOU,C,S,21(2)157-177,
Sa”tos, Nchrd. 198la. ..Dc Employrnmt S1aCUof Youth.,- In Bows
(1981 b), pP 59-116.
Holzer, Harw J., 1986.. ..Sarch Method Use by. Unemployed Youth?
National B“r~n of Eco”omic R-ch
Working Papr Setiw,
No. 1859, Cambridge, MA: NBEK
Holzer, Harry 1. 19S7a. ..hformd Job Sarch a“d Black Yo”ti
ployment? Americm Economic Review. 77(3)44642.
Joseph A, 1981.. “A. Comparative Study of htm Work Exti. . . . among Full-Tim% Pafl-Xme, and U“empIoy4
Mde
Youth? Journal of Vmrio”ol Behavior. 19(3)315.327.
Rms, tib~.
1986. “h Essay on Youth Jobl=sn=s:
nomic Lile,ature. 2q2)613 -628.
Holzer, Harv J. 1986& ‘,Thc Blink Y.”& No”employmenk Duration
a“d Job Sarch.>, 1“ Fr-a”
a“d Hal=. (1986a), pp 23.70.
md WE,
fianom-
Sante% Rchati. 1981b. ,~eH=hh
pp 465480.
Stat”$of Yo”w”
1“ Ba(1981
b),
Smtas, fich=d. 1982, “klhatinE
Youth Emplopmt
a“d ff”cmploy menfi me National SuW~s of Youth Utir Markti &fiMc..”
Roiew &PubOc Dar. Use, 10(1-2) 127-135.
fJnm-
10
S.WM1, Iabl V. 1988. ‘Povetiy i“ the U.S.: Why k it so Paiste”tY
Journal .f Economic fit,m,urc. 26.(3Z 1073.1119.
U.S. Dqaflmmt
of Lbor, 1978a, Conference ReWH o“ Youth U“employme”?:Ir.Measurement ondMeo”i”x, Wti”gton,
DC: U.S.
Government Printing OE*.
Stwmwn, WaWe. 1978.. ‘,The Relatia”ship Between Youth Employment a“d F“t”re Employ abifi$y.‘. 1“ Adam a“d Ma”g”m ( 1978),
pp 93-124.
Stevmso”, Wayne. 1978b. “The Relatio”stip Betwe”
mat ad F“l”re bployability
a“d timings.”
ment of tibor ( 1978b). pp 191-205.
U.S. Dqatime”t of tibor. 1978b. Suppleme”tam Papersfrom the Conference 0. youth Unemploymmr It. Me.swreme.t and Meaning.
Eti14 by Naomi Bergm Davidso”. Wmhi”gto”, D. C.: U.S. Govemmmt Pri”ti”g O~ce.
Youth Employ1“ U.S. Depati-
Wielg.sz,JohnB. a“d CVnter, Susan.1987..$The Effective”es of Altemat iv. Methtis of *arching for Jobs and FI”dt”s Them: An
Exploratoq Analysis of the Data B-ri”EUpO” Copin8 tith Job.
lessnes?
American Jour..!
of Economics ond SocioloD.
46(2)151-164.
Stevmm”, Wayne. 1978b. 1978., .<ThcTra”sition from School to Work?
In Adsms md Ma.g”m (1978) pp 65-91.
Summa, Lwx”ce
H. 1985, ‘Why is the U“employme”t Rate So VeW
HiSh Near F“ll Employment?- Hamard Institute of &o”omic
R-w~h
Disc”ssim Papr Seri6, No. 1284. &mbridge, MA:
Hamard U“ivemity.
Wink, Robcfi J. 1982. ‘Comment,,,
386.3.89,
11
1“ Freeman a“d Wtse (1982a), pp