Occupational Age Structure and Access for Older Workers
Author(s): Barry T. Hirsch, David A. Macpherson, Melissa A. Hardy
Reviewed work(s):
Source: Industrial and Labor Relations Review, Vol. 53, No. 3 (Apr., 2000), pp. 401-418
Published by: Cornell University, School of Industrial & Labor Relations
Stable URL: http://www.jstor.org/stable/2695966 .
Accessed: 19/04/2012 18:31
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact [email protected].
Cornell University, School of Industrial & Labor Relations is collaborating with JSTOR to digitize, preserve
and extend access to Industrial and Labor Relations Review.
http://www.jstor.org
OCCUPATIONAL AGE STRUCTURE
AND ACCESS FOR OLDER WORKERS
BARRY T. HIRSCH,
DAVID A. MACPHERSON,
and MELISSA
A. HARDY*
This paper examines covariates of the occupational age structureand
the openness ofjobs to older workers. Using a large number of data sets,
which together span the years 1983-98, the authors focus on the structure of compensation, job skill requirements, and working hours and
conditions as the principal determinants of occupational access. Older
male and female workers, they find, face substantial entry barriers in
occupations with steep wage profiles, pension benefits, and computer
usage. In addition, union coverage is associated with limited access for
older men, while older female hires are concentrated in occupations
where flex-time,part-timework, and daytime shiftsare common. Segregation across occupations among older new hires exceeds that for
younger workers,but there is no evidence that it has worsened over time.
In
on agingand
theeconomicsliterature
the labor market, scant attention has
been given to the employmentopportunities available to older workersor how these
opportunities are related to job content.
This lack of attention is not surprising,
given that older workersfare well by such
standard measures as earnings,unemployment, and displacement risk. Anecdotal
evidence and a limitedamount of scholarly
research,however,suggestthatolder workers have constrained employmentoptions.
BarryHirsch is E.M. StevensDistinguished Professor, Department of Economics, TrinityUniversity,
San Antonio, Texas; David Macpherson is Professor,
Department ofEconomics, and Director ofResearch,
Pepper Instituteon Aging and Public Policy,Florida
State University,Tallahassee; and Melissa Hardy is
Professor, Department of Sociology, and Director,
Pepper Instituteon Aging and Public Policy,Florida
State University,Tallahassee.
This paper examines thejob opportunitiesfaced byolder workers.We firstreview
previous literature and outline a frameworkbywhich the age structureof occupations is determined. In interpretingthe
evidence, we viewjob matches and the occupational age structureas the result of
choices by workers and employers. For
employers, we stress the roles played by
fixedtrainingcostsand fringebenefitsand
the divergencebetween currentproductivityand compensation among older workers
owing to implicit contracts and deferred
compensation. For workers,we emphasize
the returnto skill acquisition and the mismatch of occupational characteristicsthat
mayarise withage, in addition to standard
The occupational data set created forthe paper is
available fromBarryHirsch at [email protected].
Industrialand LaborRelationsReview,Vol. 53, No. 3 (April 2000). i by Cornell University.
0019-7939/00/5303 $01.00
401
402
INDUSTRIAL AND LABOR RELATIONS REVIEW
31). Among those displaced, older workers
have the lowest reemploymentprobabilities, the longest time to reemployment,
highprobabilitiesofpart-timeemployment,
and the largest wage losses (Farber 1997;
Chan and Stevens 1999). And consistent
withanecdotal evidence, thereis some suggestion of an increase during the 1990s in
relativeratesofdisplacementamong older,
educated, white-collar workers (Farber
1997, Figures 3a, 3b, 4a, 4b, 5). Hutchens
(1988, 1993) concluded thatjob opportunitiesare more constrainedforolder workers thanforyoungerworkers.AndJohnson
and Neumark (1997) provided evidence
fromthe National Longitudinal Surveyof
Men (NLS) thatperceived age discrimination is associated withsubsequent declines
in employmentand earnings.1
In this paper we emphasize the role
played byjob and occupational characterisBackground and Previous Literature
tics on job mobilityamong older workers.
Ruhm (1990), among others,measured the
By most measures, older workers fare
frequency of "bridge" jobs among older
well. Experienced workersrealize a subworkers.2An implication of our analysisis
stantial wage advantage relative to young
that bridge jobs would occur more freworkers,an advantage thatgrewduringthe
1980s (for example, Levy and Murnane
quentlywerejob opportunitiesfacingolder
1992). Unemploymentrates are lower for workersnot so limited. Our studyis most
older thanyoungerlabor forceparticipants. closelyrelated to workbyHutchens (1986,
In 1998,when theannual rateofunemploy- 1988, 1993) and Scott, Berger, and Garen
(1995), brieflysummarized below.3
ment among all men was 4.4% and that
Hutchens attemptedto measurewhether
among men ages 25-54 was 3.3%, rates for
men 55-64 and 65+ were 2.8% and 3.1%,
job opportunities were restrictedamong
older workers. Using Census data
respectively. Corresponding rates among
(Hutchens 1986) or CPS supplementswith
women were 4.6%, 3.8%, 2.4%, and 3.3%
(U.S. Bureau ofLabor Statistics1999, Table
3). And although 19% of men and 39% of
women ages 55 and overare employedpart'Age discriminationcannot be directlymeasured
time,fewoftheseworkers-5 % among both
analysis,since age is
men and women-report their part-time eitherbystandardwage equation
correlated with productivityand implicit contracts
statusas being for"economic reasons" (U.S.
break down the equivalency of spot marginal prodBureau of Labor Statistics1999, Table 8).
ucts and wages, or by the number of lawsuits,since
reportingis endogenous.
Older workers'riskofjob displacement is,
2Using the RetirementHistorySurvey,Ruhm deifanything,lower than thatformiddle-age
fined bridge or non-careerjobs as jobs held by older
workers,and substantiallylower than that workersthat are not theirlongest-heldjobs.
for young workers (Farber 1997, Table 1,
3Following completion of this paper, a related
Appendix Tables 3a, 3b).
article by Heywood, Ho, and Wei (1999) appeared.
The authors examined the hiringof older workersin
The positive picture presented above is
Hong Kong, where there exist no age discrimination
misleading. Although older workershave
laws. Heywood, Ho, and Wei found, as we have, that
relativelyfew unemployment spells, the
older workersare less likelyto be hired by employers
duration of given spells increases withage
in jobs with high training costs and in jobs with
deferred compensation.
(U.S. Bureau ofLabor Statistics1999,Table
labor supply determinantssuch as health
status, pension wealth, health insurance,
and the opportunitycost of time. Finally,
we consider how labor rentsresultingfrom
union coverage or employer size affect
mobilityamong older workers.
The paper next describes the construction of our occupational database, created
fromcompilationsacrossseveralmicrodata
sets. Included are measures of the occupationalcompensationstructure,skillrequirements, and workingconditions. Descriptiveevidence and statisticalanalysisare used
to examine the covariates of the occupational age structureand the "openness" of
jobs to older hires. Age segregation measures are then calculated in order to evaluate changes over timein opportunitiesfacing older (and younger) workers.
JOB ACCESS FOR OLDER WORKERS
job tenure information (Hutchens 1988,
1993), he measured theratio ofthe proportion ofnew hireswho were older workersto
the proportionofall workerswho were old.
Both thisratio and its two components are
of interest. The denominator provides a
measure of the age structureof an occupation (or industry-by-occupation),
while the
numerator provides information on the
typesofjobs into which older workersare
hired. Since the lattermeasure reflectsthe
preferences of both employers and employees, it can be a misleading indicator of
employment opportunities. A low ratio,
however,indicates that there are few new
hires of older workersrelativeto the number ofolder workersemployedin thatoccupation. This suggests that relative to the
number of older workersable or willingto
workin an occupation, feware hired.
Employmentbarriersfacingolder workers appear to result fromhigh fringebenefitcosts (Scott, Berger, and Garen 1995;
Garen, Berger, and Scott 1996) and from
implicit wage contracts wherebyyounger
workersare underpaid and older workers
overpaidrelativeto productivity(Hutchens
1986). In a related analysis, Hutchens
(1988) developed a segregation index for
older workeremployment;he reportedthat
employmentwas more segregated among
older new hires than among older workers
in general or among younger new hires.
Our analysis expands the work of
Hutchens and others in two principal directions. First,we provide evidence that
relates differencesin employmentopportunitiesto a wide arrayofvariables measuring occupational compensation structure,
skillrequirements,and workingconditions.
Second, we provide evidence on changes
over time in job segregation among older
workers.4
4Filerand Petri (1988) examined the relationship
between occupational characteristics,as measured by
theDictionaryofOccupationalTitles(DOT), retirement
behavior, and the presence of pension and health
benefits. They concluded that a number of occupational characteristicsare related to retirementbehavior. In a paper using the 1992 Health and Retirement
403
Labor Market Transitions
and the Age Structure of Jobs
In thissection,we discuss the theoretical
frameworkused to interpretour empirical
evidence on age structureand job transitions. The sortingof individuals intojobs
and the resultingage structureof occupations are determined through the interaction of heterogeneous workers and employers. Workersmaximize the expected
present value of utility,subject to constraints; firms maximize the expected
presentvalue of profits,which are increasing in revenues and decreasing in costs.
Choices made by workers and firms are
affectedand constrained bynumerous factors: individual preferences,the value of
timein alternativeactivities,the wage rate,
privateand governmentpension structures
and entitlements,health,productdemand,
technology,productivity,
the cost and valuation ofworkplaceamenities,taxrules,and
governmentregulation of the workplace.
Employmentdecisionsare madejointlywith
decisions about the structureof pensions,
the wage profile,the organization of work
and technology,and public policy. Hence,
observed outcomes map out neither labor
supply nor labor demand functions but,
rather, an "envelope" curve representing
satisfactionof marginalequilibrium conditions among heterogeneous workers and
firms.
Because of the complexitybywhich the
age structureis determined,we regard the
subsequent statisticalevidence as largely
descriptive. But such evidence can be interpreted using a maximizing economic
framework,as discussed in thissection. We
focusbelow on howjob compensation, skill
requirements,and workingconditions affect the age structureof occupations.
Study (HRS), however, Hurd and McGarry (1993)
found little relationship between prospectivelabor
market transitionsand the physical or skill content
characteristicsmeasured in the HRS. They did find
that hours and financial considerations are important determinantsof expected behavior.
404
INDUSTRIAL AND LABOR RELATIONS REVIEW
Compensation Structure
Two elementsofthe compensationstructure are examined-"wage tilt"and fringe
benefits (specifically,pensions and health
insurance). Bywage tiltwe mean therateof
wage growthor steepness of the earningsexperience profile, following control for
othermeasurable wage determinants.The
slope of the wage profile in part reflects
past human capital investment,only some
ofwhichis transferableacrossjobs. A steep
wage profilemayalso reflectdeferredcompensation, with wages rising faster than
productivity(for firm-levelevidence, see
Medoffand Abraham 1980). Because wage
tiltis likelyto provide a proxyforthe excess
ofcurrentrelativeto alternativecompensation among older workers,it has implications for occupational access.
Wage tiltreflectingdeferredwageswould
lead workers,ex post,to retire or leave a
"career"job at an age beyond that which
maximizesthejoint worker-firm
surplus,ex
ante. The use of defined benefit pension
plans that provide financial incentives to
workerswho retirewithina particular age
range can offsetthistendencyand encourage earlier retirement (for evidence, see
Ippolito 1991). Wage growththatexceeds
productivitygrowthhas an indeterminate
effect on the age structure of the work
force,but should be unambiguouslyassociated with a low probabilityof hiringolder
workers.The substitutioneffectassociated
withwage tiltdelaysjob change by raising
the returnto the currentjob relativeto an
alternativejob or retirement. Wage tilt,
therefore,maybe accompanied bypension
plans that discourage continued work.
Moreover, wage tilt and defined benefit
pension plans willoccur in thosejobs where
an earlyexit age is planned (Filer and Petri
1988). Wage tilt will unambiguously reduce the hiring of older new workers if
firmspay older workerswithlow as well as
high senioritywages in excess of marginal
products.
The compensation mix also should affectthe age distributionofworkers.On the
supply side, the structureof defined benefitplans encourages participationamong
"young" senior workers and discourages
participationamong older workersbeyond
a "normal" age or years' service at which
the present value of benefitsis maximized
(Ippolito 1987; Gustman, Mitchell, and
Steinmeier 1994; Ruhm 1996). On the
demand side, providing defined benefit
pension eligibilityto an older new hire will
have a high cost to the firmand fewof the
benefitsthat attach to their use in a longrun employmentrelationship.
Health insurance has ambiguous effects.
On the one hand, older workers are less
likelyto exit fromjobs with health coverage, at least prior to Medicare eligibility.
On the other hand, higher health costs
associated witholder workerswill discourage firmsfromemployingand hiringolder
workers (Scott, Berger, and Garen 1995),
unless health costscan be shiftedbackward
to older workersvia lower wages. If costs
can be shifted,thenhealth coverage is likely
to be high in jobs employing and hiring
older workers owing to the tax and riskpooling advantagesfrompurchasinginsurance throughan employer.
Occupational Skill Requirements,
Hours, and Working Conditions
Our next area offocus is the effectofjob
skillrequirements.Returnsto trainingtend
to decline withage owing to high opportunitycosts and a shorterperiod over which
to realize benefits. Older workers who
change jobs are likelyto either remain in
the same occupation, thus permittingthe
transfer of occupation-specific skills, or
switchto an occupation in which required
skillscan be acquired at low cost. Employers are not likelyto hire older workersin
jobs requiring substantialfirminvestment
in workertraining.5
Jobworkingconditionsand hoursshould
have an impact on the age structure of
5Acemoglu and Pischke (1998) showed how low
levels of workerturnover,independent of skill specificity,lead employers to invest in general as well as
firm-specifictraining.
JOB ACCESS FOR OLDER WORKERS
occupations. Manyworkingconditionscannot be varied absent a change in job or
occupation. Over time a mismatch may
develop between workersand jobs withrespect to hours worked, physical demands,
and otherjob attributes(for an excellent
discussion, see Hurd and McGarry1993).
If a mismatchdevelops withthe currentjob
and viable alternativejobs, then we should
observe a more rapid rate of exit fromthe
labor force. Ifjob attributesdiffersignificantly across jobs, and older workerscan
readilytransfertheirskills,then we should
observe job transitions of older workers
correlatedwithchanges in job characteristics. In such cases, we expect older job
switchersto move into occupations with
lowerworkhours, less demanding working
conditions,and relativelylow trainingcosts.
For manyolder workers,however,newjobs
providing a preferredbundle of working
conditions cannot be obtained withoutsuffering a substantial wage loss. Here, job
switchingrates should be low among older
workers.
405
Bycontrollingforunion density,we ensure
thatwe do not incorrectlyattributeto wage
tilt or fringe benefits (or other variables
correlated with union density) effectson
the age distributionthatare a direct effect
of collective bargaining. A similar argument can be made forthe inclusion of firm
size,whichis correlatedwiththewage level,
fringes,and union density.
The frameworkoutlined above is next
used to interpretthe relationshipbetween
the age distributionof incumbentworkers
and new hiresand occupationaljob characteristics.
Measures of the Age Structure
and Occupational Access
Much of our analysisuses Census-delineated detailed occupations as the unit of
analysis. "Occupation" is appropriate in
that it approximates the concept of "job
type" in terms of skill requirements and
workingconditions. We develop complementarymeasures of the age structureof
occupations. The static age distribution
can be readily measured. For descriptive
Job Transitions and Rents
purposes, we use worker age at the 90th
(P90), 50th (P50), and 10th (PIO) percenThere mayexistworkerrentsassociated
tiles to identify"old" and "young"occupawith industrywage differentials(Krueger
tions. Although median age is a standard
and Summers 1987), unionization, employersize (Brownand Medoff1989), regu- measure, it provides limited information
about the age distributionof workers. A
lated industries,or other factors. Receipt
of a wage premium lowers quits and inmeasure highlycorrelated withP90, which
we use in subsequent regressionanalysis,is
creases applicant queues. But it is not
Age5O+, representing the proportion of
obvious how rents should affect the age
workers in an occupation who are 50 or
distributionof incumbentworkersor new
over. An additional staticmeasure of occuhires. On the supply side, worker rents
increase the payofffromworkingrelative
pational age structureis the coefficientof
to retirement,but may also be associated
variation,CV(Age), measuring the age diswith greater career savings and pension
persion of an occupation.
In order to measure marketopportunibenefits. On the demand side, a high
ties and the dynamics of the market for
wage expands the applicant pool and is
older workers,we calculate measures similikely to decrease hiring of younger, less
experienced workers. But absent knowl- lar to those developed byHutchens (1986,
1993). The measure Hire5O+ represents
edge about how the wage distribution
the proportion of workersby occupation
with respect to age is affected, we can
who are "new"hires in a company,defined
predict little.
here as workerswith 5 or fewer years of
We do know thatunionization is associated not only with rents, but also with a
company tenure who are age 50 and over.
flatterwage profile and greaterfrequency HireSO+providesa measure ofthe age structure among recent hires. A low value of
of pension and health insurance coverage.
406
INDUSTRIAL AND LABOR RELATIONS REVIEW
Hire50+ may indicate either that firmsare
restrictingthe hiringofolder workersin an
occupation or that few older individuals
are able or willingto work in that occupation. Hutchens has proposed a measure
combining informationon new hires and
the existing age structurein order to approximatewhetherhiringopportunitiesor
access, relativeto preferencesand abilities,
are restrictedfor older workers. Letting
itrepOpen50+be equal to Hire50+/Age50+,
resents an index measuring the openness
or accessibilityof an occupation to older
new hires, relativeto the number of existing older workers.
Open50+ and its component parts provide useful information. For example, occupations requiring physicalstrengthmay
have low levels of existing older workers
(Age5O+) and older new hires (Hire5O+),
but need not have low accessibilityrelative
to worker preferences and abilities
(Open5O+). Occupations that defer compensationviawage tilt,pensions,and health
benefitsmay employ many senior workers
yethire fewolder workers(a high Age5O+,
but low Hire50+ and Open5O+). The negativerelationship between wage tiltand accessibilityforolder workersis preciselythe
relationship considered by Hutchens
(1986).
Following descriptive evidence on the
occupational age structuremeasures described above, we examine their determinants, estimating gender-specificmodels
of the followinggeneral form.
(1)
+ SkillsI3a
Age5O+= Compmixc(x
+ Conditionst
+
XV
+ ea
a
a
(2)
Hire50+= Compmixax
+ Skillsfib
+ Conditionst
b + Xrb+ eb
(3)
Access= Compmixcx
+ Skills,3c
+ Conditionst
+
+ ec
XTn
c
c
(4)
CV(Age) = Compmixc(x
+ Skillsid
+ Conditionstd + Xrd +
ed
Estimation is by weighted least squares
(WLS), with gender-specificoccupational
employmentas weights.6 All variables are
measured at the occupation level (we omit
a subscriptindexing occupation). Compmix
representsa vectorincludingthewage level,
wage tilt, health insurance, and pension
coverage; Skillsa vector including education,company-providedtraining,computer
usage, and required numerical aptitude;
Conditionsincludes measures of the frequency of shiftwork, long hours (more
than 42 hours workedper week), part-time
employment,flex-time,outdoor work,occupational strength requirements, exposure to extremeenvironmentalconditions,
job hazards, and physicaldemands; and X
includes variables measuring the proportion of workersunionized, the proportion
employed in large firms,and the rate of
employmentgrowth.
As indicated in the previous section, the
coefficientsa, P, t, and r should varyin a
predictable way across equations. For example, we expect skillvariables to be associatedwiththepresence ofincumbentolder
workers but few young workers and few
older new hires. Wage tilt and pension
benefits may be associated with low age
dispersionand limitedaccess to older workers. Occupations withlow trainingand skill
requirements will attract young workers,
but theymayalso attractolder workerswho
have lefttheircareerjobs, assuming physical demands and hours requirements are
not onerous.
Data and Descriptive Evidence on
Occupational Age Structure and Access
Our studyrelies on data assembled from
a large number of data sets. Much of the
occupational analysis is based on genderspecificvariables compiled byus fromvarious micro-levelCurrentPopulation Survey
(CPS) files. The data setsused in the paper
6WLS estimation using employmentweights provides coefficientestimatesrepresentativeof the occupation of an average workerrather than the average
occupation, while also weighing less heavilyobservations whose variables are calculated with greater error.
JOB ACCESS FOR OLDER WORKERS
include the CPS Outgoing Rotation Group
MonthlyEarningsFiles (CPS-ORG) forthe
years 1983-95; the March CPS Annual Demographic Files for 1983-95; CPS training
supplements for January 1983 and 1991;
CPS computer use supplements for October 1984 and 1989; CPS dualjob/shiftwork
supplements for May 1985 and 1991; and
CPS tenure supplementsforJanuary1983,
May 1983,January1987, May 1988,January
1991, April 1993, February1996, and February1998. Occupational skill and working condition variables are obtained from
ofOccupational
the fourtheditionDictionary
Titles (DOT), as mapped to 1980 Census
occupation codes byEngland and Kilbourne
(1988) and made time-consistentby us to
account forthe minorchanges betweenthe
1980 and 1990 Census codes. DOT variables are not gender-specific. Definitions
of and sources forall variables used in our
analysisare provided in the appendix.
We turnfirstto descriptiveevidence on
the age structureof occupations. Table 1,
with panels for men and women, respectively,includes selected occupations based
on size, number of old or young workers
(P90 or Pl0), proportionofolder newhires
(Hire50+), accessibility to older workers
(Open5O+), and age dispersion (CV(Age)).
Among men, occupations with the oldest
workers (a high P90) tend to be those requiring few physical demands, flexible
hours and schedules, and, for the most
part, low skill and training requirements
(forexample, crossingguards,messengers,
private guards, and taxi and bus drivers).
Chief executivesandjudges are exceptions
to the generalization regarding skills,presumablybecause skillsin these occupations
depreciate slowly.Most of the occupations
withhigh P9Os also hire a high proportion
of older workers, as compared to the
economy-widemean for Hire5O+,which is
.10. Interestingly,
many"old" occupations
(for example, messengers,parking lot attendants, and private guards) have high
age dispersion and many young workers,
PIO being below the economy-widemean
of 23.
Occupations with low trainingrequirementsand high physicaldemands or unde-
407
sirable hours tend to employyoungbut not
older workers(forexample, stockhandlers
and baggers,cooks). Well-paidjobs requiring lengthytrainingtend to have fewyoung
and fewold workers,the latterresultowing
to retirementcombined withlittlenew hiring of older workers. For example, skilled
administrativeand managerialoccupations
have low age dispersion. The accessibility
measure Open5O+(the ratio of the proportionsof older new hiresto older workers)is
largelyunrelated to age (the correlationof
Open50+and P90 is .03), and obtains some
ofitshighestvalues in occupations withfew
older workers (for example, recreation
workersand food counterworkers). Occupations requiringsubstantialtrainingtend
to have low accessibilityfor older workers
(the correlationof Open5O+and firmtraining is -.26).
We obtainsimilarpatternsamongwomen
(panel B ofTable 1). Occupations withthe
oldest workers,on average, among women
include household workers,welfareservice
aides, religious workers, and crossing
guards. In many occupations, large numbers ofyoungand old women are hired; for
example, sales occupations, cashiers, and
privatehousehold child care. Occupations
in which older women appear to have limited access (low values of Hire5O+ and
Open5O+) tend to have high training requirements, demanding working conditions,or, in a fewcases, high-valuedphysical attributes(forexample, announcers and
dancers). Occupations with skilled workers, a demanding work environment,and
high ratesofpension coverage tend to have
relativelylow age dispersion (for example,
registerednurses and teachers).
Occupational Age Structure
and Access: Regression Results
Equations (1)-(4) provide a framework
for examining the covariates of occupational age structureand access-Age5O+,
Hire5O+, Open5O+,and CV(Age). Table 2
presentstheWLS regressionresultsforthese
equations for men and women. We organize our discussion around the fourgroups
of explanatory variables-compensation
408
INDUSTRIAL AND LABOR RELATIONS REVIEW
Table 1. Descriptive Statisticson Male Age Structure
and Female Age Structurefor Selected Occupations, 1983-95.
Occupation
N
All Men
1,192,721
SelectedLargeMale Occupations:
Managers and administrators,
n.e.c.
73,226
Truck drivers
50,326
Supervisors and proprietors,sales
34,574
Janitorsand cleaners
31,645
Supervisors,production
occupations
23,783
Sales representatives,
mining, manufacturing
23,072
Laborers, except construction
20,758
Cooks
19,874
Construction laborers
15,030
SelectedOld P90 Male Occupations:
231
Crossing guards
Chief executives and general
administrators
272
Messengers
2,162
Clergy
6,384
Taxicab driversand chauffeurs
2,883
Bus drivers
5,180
SelectedHigh Hire5O+Male Occupations:
231
Crossing guards
Tailors
450
638
Judges
Guards and police, excluding
public service
12,136
SelectedHigh Open5O+Male Occupations:
Recreation workers
589
Protectiveserviceoccupations, n.e.c.
851
Athletes
1,054
Food counter, fountain and
related occupations
1,869
SelectedHigh CV(Age)Male Occupations:
News vendors
1,225
Sales workers,apparel
1,585
876
Parking lot attendants
Attendants,amusement and
recreation facilities
2,320
Cashiers
9,381
Stock handlers and baggers
15,598
SelectedYoungPl0 Male Occupations:
News vendors
1,225
Stock handlers and baggers
15,598
Miscellaneous food
preparation occupations
7,258
Food counter, fountain and
related occupations
1,869
Cashiers
9,381
Cooks
19,874
SelectedLow Hire5O+and Open5O+Male Occupations:
690
Physicians' assistants
Drillers, earth
327
Data-entrykeyers
1,291
183
Carpenter apprentices
SelectedLow CV(Age)Male Occupations:
Supervisors,police and detectives
1,762
Administrators,education and
related field
6,288
Personnel and labor relations
managers
1,185
Airplane pilots and navigators
1,927
Open CV
50+ (Age)
P90
P50
P1O
Age
50+
Hire
50+
55
A. Men
35
23
0.191
0.101
0.529
33
57
56
54
62
40
36
36
38
26
23
24
19
0.242
0.204
0.162
0.313
0.128
0.134
0.092
0.219
0.526
0.655
0.565
0.700
28
33
30
40
56
40
27
0.248
0.088
0.356
26
57
54
43
53
38
31
23
30
25
19
17
19
0.219
0.150
0.062
0.139
0.124
0.092
0.029
0.080
0.569
0.615
0.467
0.574
30
38
42
38
77
68
50
0.900
0.672
0.747
21
69
66
65
65
64
49
32
45
40
45
34
19
30
24
27
0.493
0.285
0.404
0.323
0.398
0.341
0.239
0.224
0.277
0.282
0.692
0.838
0.555
0.857
0.707
26
48
28
36
30
77
63
68
68
46
53
50
25
37
0.900
0.418
0.592
0.672
0.445
0.439
0.747
1.066
0.741
21
32
23
64
40
22
0.340
0.307
0.904
39
50
43
47
28
20
27
19
16
18
0.105
0.062
0.069
0.314
0.140
0.112
2.985
2.252
1.613
40
48
38
28
18
16
0.014
0.019
1.383
36
57
61
65
22
24
29
16
17
19
0.153
0.162
0.235
0.068
0.131
0.242
0.441
0.811
1.028
56
53
49
58
50
41
28
22
20
17
17
16
0.159
0.101
0.065
0.106
0.055
0.031
0.662
0.542
0.476
49
49
49
57
41
22
20
16
16
0.153
0.065
0.068
0.031
0.441
0.476
56
49
46
21
16
0.082
0.059
0.719
47
28
50
43
18
22
23
16
17
17
0.014
0.101
0.062
0.019
0.055
0.029
1.383
0.542
0.467
36
49
42
44
50
49
31
32
31
29
21
24
22
20
18
0.029
0.116
0.100
0.016
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
24
32
36
29
52
42
32
0.163
0.019
0.120
19
58
45
33
0.349
0.144
0.412
22
55
54
42
41
30
29
0.262
0.207
0.141
0.056
0.540
0.272
23
23
Continued
JOB ACCESS FOR OLDER WORKERS
409
Table 1. Continued.
Occupation
N
P90
P50
B. Women
All Women
1,104,886
55
35
SelectedLargeFemaleOccupations:
Secretaries
86,450
56
37
Cashiers
44,555
52
25
Managers and administrators,
n.e.c.
42,464
55
37
Registered nurses
36,573
55
38
Bookkeepers, accounting,
and auditing clerk
35,623
58
38
Teachers, elementaryschool
30,398
55
40
Nursing aides, orderlies,
and attendants
29,763
58
38
Waiters and waitresses
27,522
48
26
Sales workers,other commodities
23,554
60
32
SelectedOld P90 FemaleOccupations:
Cooks, privatehousehold
71
294
56
Musicians and composers
69
827
42
Demonstrators,promoters
and models, sales
696
68
47
Privatehousehold cleaners
and servants
67
47
10,273
Crossing guards
735
67
46
Welfare service aides
65
1,973
44
Religious workers,n.e.c.
1,010
64
44
SelectedHigh Hire50+ FemaleOccupations:
Cooks, privatehousehold
294
71
56
Housekeepers and butlers
664
68
48
Postmastersand mail superintendents
63
614
49
Clergy
618
62
44
SelectedHigh Open5O+FemaleOccupations:
Actuaries
43
122
30
Sales workers,motor vehicles
and boats
435
48
34
Protectiveserviceoccupations, n.e.c.
933
40
19
SelectedHigh CV(Age)FemaleOccupations:
Ushers
147
57
19
Child care workers,private
household
56
8,278
22
Sales workers,apparel
61
8,483
26
Food counter, fountain and
related occupations
37
6,117
18
Sales workers,other commodities
60
23,554
32
Cashiers
44,555
25
52
SelectedYoungPl0 FemaleOccupations:
Child care workers,private
household
56
8,278
22
Food counter, fountain and
related occupations
37
18
6,117
Sales workers,shoes
54
1,549
22
57
29
Waiters'/waitresses'assistants
3,483
Stock handlers and baggers
53
28
5,167
Farm workers
53
31
2,739
SelectedLow Hire50+ and Open5O+FemaleOccupations:
Mechanical engineers
283
44
29
43
Photographers
503
28
Announcers
234
43
28
Dancers
270
33
25
SelectedLow CV(Age)FemaleOccupations:
Railroad brake, signal, and
switchoperator
43
20
37
Social work teachers
34
57
44
Law teachers
44
47
38
34
Metallurgical and materials engineers
40
32
PlO
Age
50+
Hire
50+
23
0.187
0.102
0.537
34
23
17
25
26
0.200
0.123
0.180
0.187
0.100
0.065
0.085
0.110
0.499
0.524
0.471
0.589
33
45
30
27
24
27
0.245
0.203
0.142
0.051
0.577
0.251
33
25
22
18
18
0.233
0.089
0.232
0.141
0.043
0.142
0.606
0.478
0.615
34
40
45
28
23
0.619
0.366
0.509
0.255
0.822
0.697.
30
38
20
0.460
0.311
0.676
40
26
32
25
27
0.455
0.424
0.381
0.352
0.300
0.338
0.365
0.101
0.659
0.797
0.957
0.288
33
28
32
31
28
25
35
29
0.619
0.461
0.497
0.353
0.509
0.447
0.444
0.388
0.822
0.971
0.894
1.100
30
32
22
28
23
0.025
0.093
3.801
25
21
16
0.080
0.061
0.178
0.113
2.215
1.853
31
47
16
0.136
0.268
1.970
57
16
17
0.144
0.227
0.119
0.150
0.824
0.662
54
50
16
18
17
0.044
0.232
0.123
0.027
0.142
0.065
0.610
0.615
0.524
47
45
45
16
0.144
0.119
0.824
54
16
17
17
17
17
0.044
0.138
0.183
0.135
0.137
0.027
0.126
0.113
0.102
0.107
0.610
0.910
0.620
0.758
0.786
47
49
48
44
40
23
19
18
20
0.060
0.052
0.043
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
n.a.
29
34
34
21
26
34
27
25
0.050
0.324
0.068
0.029
0.079
0.000
0.000
0.000
1.574
0.000
0.000
0.000
18
19
19
20
Open CV
50+ (Age)
N is the sample size fromthe 1983-95CPS-ORG earningsfiles.
Allvariablesin panel A are calculated formen and in panel B forwomen. For variabledefinitionsand sources,see the appendix.
410
INDUSTRIAL AND LABOR RELATIONS REVIEW
level and structure,occupational skills,job
working conditions and hours, and a category including union density,firmsize,
and demand growth.7
Compensation Level and Structure
tion of mean schooling, with which it is
highlycorrelated.
Pension coverage is positivelyrelated to
the proportion ofworkersover 50 and participationofmiddle-ageworkers(see Ruhm
1996), but negativelyrelated to Open5O+
and to age dispersion. That is,the presence
of pension benefitsacts as a barrier to job
change. The evidence for this pattern is
fairlystrong for older men, but weak for
older women.8 No statisticallysignificant
relationshipis foundbetween health insurance coverage and the age structurevariables formen or women. The absence of a
negative relationship with Hire5O+ and
Open5O+is at odds with expectations and
the conclusion reached by Scott, Berger,
and Garen (1995), who testedthisrelationship more directly.
Perhaps our most strikingresult is the
statistically
significanteffectofwage tilton
theage structure.Occupationswithsteeper
profiles, aln W/alnEXP, are less likely to
have a high proportion of older workers
(Age5O+)and less likelyto hire older workers (Hire5O+). The point estimates imply
thata .10 increase in the slope of the earnings profile (the mean is .13) is associated
with a .035 decrease in the proportion of
older workersand a large .047 decline in
the proportion of older new hires (mean
Hire5O+is .101). Wage tiltalso is associated
with occupations that are less open or acOccupational Skill Requirements
cessible to older new hires, measured by
the ratio of older new hires to older incumSkill variables are systematically
related
bents (Open5O+). This finding provides
to the age structure,but the typeof skill
support for implicitcontracttheories prematters. As seen above, the occupational
dictingrelativeunderpaymentofyoungand
wage (skill) level is associated positively
overpaymentof older workersrelative to
with age and negativelywith age disperproductivity (see, relatedly, Hutchens
sion. Occupations withfirm-provided
train1986). It also helps explain both the barri- ing, those requiring high numerical aptiers older workersface in obtainingemploy- tude, and those with high computer use
ment and the substantial earnings losses
have fewerolder male workers. Although
among older displaced workers (for exhigher skillrequirementsare generallyasample, Farber 1997, Tables 11, 12).
sociated with lower age dispersion, comAs predicted, higher wage occupations
puter use and numerical aptitude (among
are found to be associated witholder work
men) are exceptions,being associated with
forces, but with substantiallyless age disgreater dispersion due to a large concenpersion. To some extent,thewage variable
trationof youngworkers. Occupations remay reflecttrainingor other skill-related quiring computer use not onlyemploy few
job attributesnot fullymeasured by other
older workers,but also are less accessible to
variables. Exclusion of the wage variable
older workers,at statistically
significantlevhas a relativelymodest effect on coeffi- els, than are other occupations (that is,
cients of other variables, with the exceplower Open5O+, as well as Age5O+ and
Hire5O+). Occupations requiring high numerical aptitude exhibit the same pattern,
but only among men, not women.
7In order to gauge the relative importance of the
manyvariables,we also estimated all equations using
standardized beta coefficients. The relative size of
the standardizedcoefficientsfollowedcloselythe relativesize of coefficientt-ratios,withthe one exception
that the relative impact of the pension variable was
larger than suggested by its significancelevel.
'Garen, Berger, and Scott (1996) provided evidence thatdefined benefitpension plans discourage
employersfromhiring older workersfor entry-level
positions.
JOB ACCESS FOR OLDER WORKERS
411
Table 2. WLS Regression Results:
Determinants of Male and Female Occupational Age Structures.
Age5O
Variable
Coeff.
Hire5O+
t
Coeff.
t
Open5O+
CV(Age)
t
Coeff.
ln(Wage)
lnW/alnExp
Pension
Health
0.078
-0.347
0.306
0.014
2.87
-4.55
4.65
0.22
Schooling
Firm Training
Computer
Numerical
ShiftWork
Overtime
Part-time
Flex-time
0.008
-0.077
-0.156
-0.033
1.81
-2.82
-6.29
-3.44
A. Men
0.085
3.38
-0.466 -6.65
0.055
0.91
0.106
1.75
0.009
2.07
-0.044 -1.76
-0.132 -5.78
-0.033 -3.75
-0.023
0.066
0.106
0.068
-1.09
2.20
1.88
1.91
-0.006
0.031
0.076
0.020
Strength
Environmental
Hazards
Physical Demands
Work Outdoors
Union
Large Firm
EmploymentGrowth
Constant
-0.010
-0.026
0.054
0.008
0.017
DependentVar. Mean
R-square
N
-0.002 -0.22
-0.026 -3.20
0.022
1.15
0.009
1.26
0.004
0.29
-0.156 -4.84
-0.010 -0.33
-0.031 -4.11
-0.054 -0.79
0.191
0.381
497
ln(Wage)
alnW/alnExp
Pension
Health
0.056
-0.244
0.372
-0.131
2.02
-2.23
5.97
-1.81
B. Women
0.031
1.40
-0.338 -3.81
0.155
3.09
-0.016 -0.27
-0.017
-1.189
-0.332
0.264
-0.19
-3.23
-1.59
1.08
Schooling
Firm Training
Computer
Numerical
ShiftWork
Overtime
Part-time
Flex-time
-0.032
-0.039
-0.068
0.000
-6.57
-1.32
-3.30
0.06
-0.022
-0.012
-0.063
0.001
-5.54
-0.52
-3.79
0.11
-0.012
0.073
-0.101
0.006
-4.43
2.42
3.21
4.84
0.61
0.35
0.03
-0.13
-3.03
1.04
-5.18
-1.65
9.11
0.187
0.379
494
-0.077
0.090
0.113
0.140
0.003
-0.002
0.014
0.000
-0.021
-0.019
-0.085
-0.004
0.305
-4.22
2.10
2.97
5.08
0.47
-0.15
0.57
0.01
-1.55
-0.67
-3.90
-1.08
7.65
-0.082
0.134
0.058
0.151
0.001
-0.013
0.034
-0.001
0.011
-0.72
0.75
-1.48
0.22
-1.09
0.75
0.37
1.32
0.05
-0.27
0.33
-0.04
0.20
-1.94
0.01
-0.18
4.94
Strength
Environmental
Hazards
Physical Demands
Work Outdoors
Union
Large Firm
EmploymentGrowth
Constant
DependentVar. Mean
R-square
N
-0.100
0.129
0.152
0.166
0.005
0.005
0.001
-0.001
-0.052
0.037
-0.140
-0.007
0.449
-0.174
-0.004
-0.011
-0.105
-0.29
1.13
1.46
0.61
-1.18
-3.47
2.99
1.24
1.41
-5.85
-0.13
-1.62
-1.68
0.101
0.257
497
0.102
0.386
494
Coeff
0.223
-1.109
-0.640
0.362
2.11
-3.77
-2.52
1.42
-9.073
-1.126
-10.973
3.915
0.033
0.032
-0.345
-0.070
1.85
0.30
-3.61
-1.89
0.104
-4.065
2.744
1.413
-0.48
-0.25
0.33
-0.93
-0.88
-2.37
2.92
0.39
1.29
-3.24
0.71
0.33
0.28
0.529
0.233
493
-1.488
-5.361
10.838
-0.977
-0.039
-0.028
0.073
-0.127
-0.030
-0.075
0.219
0.010
0.066
-0.403
0.087
0.010
0.073
-0.231
0.001
-0.002
0.812
0.537
0.206
463
t
0.799
0.023
-0.573
-0.528
1.507
-8.79
-0.39
-4.43
1.58
0.59
-3.94
2.93
3.89
-1.86
-4.76
5.09
-0.73
2.41
0.07
-0.78
-2.06
3.00
-3.032
2.660
-0.528
54.988
-2.49
2.22
-1.86
21.44
32.813
0.882
497
-16.295 -14.48
-3.546 -0.80
-13.997 -5.53
4.498
1.53
1.843
-4.143
3.602
0.232
1.835
-3.156
2.256
-5.361
-0.316
-0.267
1.514
0.604
0.049
-5.440
11.654
-0.049
44.970
9.46
-3.49
4.32
0.72
2.00
-1.46
1.18
-3.86
-1.05
-0.46
1.20
2.52
0.07
-3.76
10.65
-0.30
22.45
33.501
0.870
494
All variables are defined at the Census detailed occupational level. Apart from the DOT measures, all
variables in Panel B are calculated based on female-onlysamples.
For variable definitionsand sources, see the appendix.
412
INDUSTRIAL AND LABOR RELATIONS REVIEW
A broad generalization that emerges is
thatthe older workforce is relativelyhighskilled,but not strongin quantitativeskills.
Older workersare unlikelyto select or be
selected forjobs providingsubstantialonthe-job training or requiring computerbased skills.
We also examine occupational working
conditions fromthe DOT. Withfewexceptions,workingconditions are not strongly
related to the age structureorjob access for
men or women. Occupations with exposure to extreme environmentalconditions
(noise, chemicals, and so on) have fewer
older male workers and are less likely to
hire older men. No such relationship is
Work Hours and
found for women, but few women are in
Occupational Working Conditions
suchjobs. Required strengthis not related,
We consider four "workhour" features at conventional levels of statisticalsignifiof occupations: the frequencies of shift cance, to the age structurefor women or
work, overtime, part-time,and flex-time. men. Surprisingly,hazardous occupations
are positivelyassociated with Hire5O+and
The overall picture that emerges shows
Open50+formen,while number ofphysical
these features to be very important for
demands (climbing, reaching, stooping,
women, but farless so formen.
and so on) is not related to the age strucJobswithsubstantialamountsof evening
ture for men or women. Occupations reand night shiftwork are less likelyto emquiring workoutdoors have a high disperployor hire older women. Shiftworkis not
sion in male workerage and displaya somestronglyrelated to the male occupational
what higher older male hire rate than do
age structure. Occupations with a high
other occupations. There is a statistically
proportion of employees working long
significant inverse relationship between
hours (more than 42 hours a week) have
outdoor work and employment of older
male and female work forces that are
women.
older but less age-dispersed. OccupaTaken as a whole, the weak relationship
tions with high proportions of part-time
of these DOT variables with the occupamale and female workers tend to have a
tional age structureand access supports
more dispersed age distribution, and
the findingsof Hurd and McGarry(1993),
higher proportions of older workers and
older hires. The former effect is stron- who, using microdata fromthe Health and
ger among men, since most part-time Retirement Survey,found little effect of
workingconditions on prospective retiremen are young, and the latter effect is
ment. Our resultsremain a bit surprising,
stronger among women, reflectinglarge
however,giventhe predictionsfromtheory
numbers of older part-time women.
and our assessmentof the descriptivedata
Open50+ is unrelated to part-timework,
(Table 1). This discrepancy may arise in
suggesting that the relationship between
part from our use of a broad rather than
part-timework and the age structure is
narrowgroup ofolder workers(thatis,ages
primarilya labor supply rather than de50+ versus,say,65+) in the regressionanalymand phenomenon.
sis and collinearity of the DOT variables
Whereas the presence of "flex-time"
programsin theworkplace is not an important withthe skilland other included variables.
Or the limited mobilityof older workers
covariateformen, among women it is assobecause ofjob skillrequirements,wage tilt,
ciated witha substantiallyolder workforce
and pensions maymake itdifficultforthem
and a higherproportionof older hires,but
to switchjobs in response to a mismatchin
lower age dispersion, reflecting the fact
working conditions. At any rate, absent
that there are fewyounger women injobs
more compelling evidence, it is difficultto
with flex-time. Either companies hiring
argue that there exists a substantial misand retaining older workers choose to
match between occupational workingconadopt flex-time policies or companies
ditions and the preferencesof older workadopting such policies attract and retain
ers remaining in the labor force.
older workers.
JOB ACCESS FOR OLDER WORKERS
Rents, Unions, and Wage Growth
Older workers' age structure and job
opportunities may also be affectedby the
presence of workerrents or labor market
institutions,captured empiricallybyunion
coverage. The principal effectof unionizationshould be to lowerage dispersionwithin
a workforce. Rents associated withunion
compensation lowerturnoverand increase
the number ofprime-ageworkers.Pension
benefitsand more demandingworkingconditions (Duncan and Stafford1980) lead to
few very old workers. And selection by
union employersfroma lengthyapplicant
queue should lead to fewveryyoungworkers.
As expected, union densityis associated
withlower age dispersion among men and
women. Among men (but notwomen), the
proportionof older workersand older new
hiresis also lowerin highlyunionized occupations. For both men and women, job
access for older workersis lower in highly
unionized jobs, perhaps indicating employer reluctance to hire older workersin
jobs withgenerous health and pension benefits.
The proportionofworkersin large firms
(1,000+ employees) is generallyunrelated
to the age distributionfor men, following
othercontrols (see, also, Scott,Berger,and
Garen 1995). Among women, however,
thereare fewerolder employees and fewer
older hires. But age dispersion is higherin
large firms,the latteroutcome reflectinga
verylarge population of young female employees in large firms.
Finally,theemploymentgrowthrateover
the 1983-95 period is associated withfewer
male workersages 50 and over,reflectinga
concentrationofyoungworkersamong new
hires. Apart from this relationship, employmentgrowthshows little relationship
with the hiring of older workers or job
access.
Overall, our regression results support
and help clarifythe general framework
outlined in theprevioussection. There are
predictable relationships between the occupational age structureand older worker
hiringwithrespect to the skilland training
413
requirements of jobs, the compensation
level and mix,unionization,and workhour
arrangements.In particular,older workers
have highlyrestrictedaccess to jobs with
steep wage profilesand those offeringpensions, occupations requiring computing
skills,and jobs thatare unionized. By contrast,job working conditions, apart from
hours, exertlittlenet influenceon occupational age structureor occupations' openness to older new hires.
Age Segregation and
Occupational Access:
Have There Been Changes over Time?
Hutchens (1988, 1991) has proposed the
use of "segregation curves" to evaluate
whether job access for older workers is
He provided evidence
constrained.
(Hutchens 1988), based on the January
1983 CPS tenure supplement, that jobs
among newlyhiredolder workersare more
highlyconcentrated across occupation-byindustrycells than the distributionofjobs
among youngernew hires or among older
workersin general. He concluded thatjob
opportunitiesare more segregated forold
than foryoung new hires,and thatthere is
greater segregation among newly hired
older workersthan among older workersin
general. Althoughevidence of greatersegregation does not prove that occupational
access is limitedamong older workers,it is
surelysupportiveof thatview.
We extend Hutchens's analysisby examining age segregationforfivetimeperiods:
1983, 1987, and 1991 using JanuaryCPS
tenure supplements, and 1996 and 1998
using FebruaryCPS supplements. Two issues are investigated.First,we ask whether
Hutchens's conclusions hold up when we
use alternativedata sets. Second, we examine whether age segregation has changed
over time.
Hutchens's segregation curve is best illustratedwitha diagram, shown by Figure
1. We firstgroup all employed wage and
salaryworkersinto occupational cells-346
formen and 269 forwomen. We differentiate among alternativegroups of workers;
for example, newly hired older workers
414
INDUSTRIAL AND LABOR RELATIONS REVIEW
Figutre1. SegregationCurves and the
MeasurementofOccupationalSegregation.
A
0
X
B
0
Cum.% ofType2 Workers
Explanation:Lettingx1be the numberof type1
workers(forexample,older new hires) and x2 the
numberoftype2 workers
(all others),theverticalaxis
measuresthecumulative
percentageoftype1 workers,withoccupationsorderedbyx1/x2.The horizontalaxismeasuresthecumulative
percentageoftype2
workers,again withoccupationsorderedby x1/x2.
The Giniindex,whosevaluesareprovidedinTable 5,
is measuredbyG = (5,000- B) /5,000,whereB is the
area underthesegregation
curve.
(defined as workersage 50 and over with5
or feweryearsof tenure) are referencedas
"type1" workerswhile all otherworkersare
referenced as "type2" workers. Letting xi
and x2 be the numbers of type 1 and 2
workers,the ratio xl/x2,measuringthe mix
of older new hires to all other workers,is
calculated foreach occupation. Each occupation is then sorted from low to high
values of xl/x2.In Figure 1, theverticalaxis
measures the cumulative
percentage of type
1 (older new hire) workers,with occupationsordered by xl/x2.The horizontal axis
measuresthecumulativepercentageoftype
2 (all other) workers,again with occupations ordered by xl/x2. A plot of points is
formed from the set of P(c2,c), where j
representsoccupations ordered ly xI/X2, c2
is the cumulativepercentof type2 workers,
and cl is the cumulative percent of type 1
workers. If older new hires were distributed across occupations in exactlythe same
wayas other workers,P(c2,cl) would plot a
45 degree line of equality; if there were
complete segregation,all points would lie
along the horizontaland rightverticalaxes.
The segregationcurve,much like Lorenz
curvesused to examine income inequality,
lies below the line of equality, as seen by
curveOA in Figure 1. As shownbyHutchens
(1988,1991), givenreasonable assumptions
one can unambiguouslyargue thatone distributionis less equal than another ifit lies
everywherebelow the other. Hutchens
(1988) showed thatjob segregationamong
older new hires is unambiguously greater
than among eitheryoungernew hiresor all
older workers.AlthoughHutchens (1988)
did not do so, one can calculate a Gini
coefficient(G) measuring the ratio of the
area between the segregationcurve and 45
degree line to the area under the 45 degree
line (that is, G = [5,000-B]/5,000, where
the area under the 45 degree line is 5,000
and B is the area under the segregation
curve). A G = 0 would implyperfectequality(the 45 degree line) while a G= 1 would
implycomplete segregation. Among segregation curves that do not intersect,G values provide the same rank ordering as do
segregation curves (Hutchens 1991). The
G coefficientis not withoutproblems,however, since a higher G for one group than
another does not rule out the possibility
that their curves intersect. Despite this
reservation,G provides a useful metricfor
measuring the degree of segregation, although we are reluctant to place much
emphasis on small differencesin G.
Table 3 provides values of G for three
groups of male and femaleworkers-older
new hires (workersage 50 or over with 5
yearsor fewercompanytenure), youngnew
hires (new hires younger than 50), and all
older workers.Values are calculated at five
points over the 1983-98 period.
Our analysismakes twocontributionsto
this line of research. First, we confirm
Hutchens's previous findings, which he
based on the January1983 CPS. Occupational segregation is substantiallygreater
for older new hires than for either young
new hires or all older workers. Young new
hires appear to have a slightlymore equal
distributionthan all older workers.For the
JOB ACCESS FOR OLDER WORKERS
415
Table 3. GiniMeasuresofOccupationalSegregation.
1983
1987
1991
1996
1998
Older New Hires
Young New Hires
Older Workers
.397
.240
.276
.359
.230
.276
.416
.228
.290
.390
.236
.289
.370
.250
.307
Women:
Older New Hires
Young New Hires
Older Workers
.350
.226
.275
.355
.200
.252
.357
.195
.273
.349
.226
.250
.315
.211
.250
Group
Men:
Data sources are CPS Supplements for January 1983, January 1987, January 1991, February 1996, and
February1998. There are 346 occupational cells formen and 269 forwomen. The Gini index (G) ranges from
0 (no segregation) to 1 (complete segregation). The Gini index of occupational segregation is described in the
text and in Figure 1.
occupations and frequency of older new
hiresvarywithrespectto the compensation
level and mix, skill requirements,working
conditions and hours, and union statusof
thejob.
The results are broadly consistentwith
matchingframethe economic worker-firm
workwe outlined. Among the more important findings are that steep wage-experience profilesare associated withfewerolder
male and femaleworkers,fewerolder hires,
and a lower ratio of older hires to incumbent older workers. Pension benefitsare
more prevalentin jobs witholder workers,
but such jobs limit access to older hires.
Occupations with extensive computer use
have few older workers,few older hires,
Conclusions
and limitedopenness. Workhours have an
impact, particularlyfor women. Jobs reOur largelydescriptiveanalysisenhances
knowledgeabout theoccupational age comquiring night and evening shiftshave few
positionand employmentopportunitiesfor
older women and older hires, while jobs
older workers.Occupations in whicholder
withflex-timeschedules and in which partworkersare employedneed not be thesame
time work is prevalent have many older
female workersand older hires. For male
as those in which older workersare hired.
workers,union statusis associated withfew
We have found that the age structureof
older workersand hires,and more limited
access. Least consistent with theory and
expectations is evidence thatmost occupational working conditions (for example,
measuresdisplaysensitivity
to
9Theconcentration
hazards, strength,physicaldemands) have
samplesize. Each ofthesupplements
includestenure
littleeffecton age composition and access.
responsesforall eightCPS rotationgroupsin a single
The
notable exception is extremeenvironso samplesizesare similaracrossyears.Data
survey,
mentaljob risks,which limithiringof and
for1983 and 1991 correspondto recessionary
periods,whereasdata fortheotheryearsdo not.
access to older men.
1998 sample, values of G for men are .37,
25, and .31 forolder new hires,young new
hires, and all older workers,respectively.
Corresponding numbers among female
workersare .32, .21, and .25. Second, the
results in Table 3 indicate there was no
increase between 1983 and 1998 in occupational segregationfacingolder workers. G
formale older new hires changes from.40
to .37, and the change among older women
is from .35 to .32. Although the recent
decrease in G is suggestiveof improving
job prospects facing older workers,we are
unwillingto attachweightto small changes
in the concentrationmeasure.9
416
INDUSTRIAL AND LABOR RELATIONS REVIEW
Our analysismakes clear thatthe generally positive outcomes observed among incumbent older workersprovide an incomplete picture of the labor market. Positive
outcomes forincumbentsare not representativeof thejob opportunitiesfacingolder
workerswho switchjobs, or opportunities
thatwould face non-switchingolder workerswere theyto change jobs. Althoughage
segregationamong older new hiresexceeds
that among older workersin general and
thatamong youngernew hires,we find no
evidence that age segregation has worsened over time.
The long-runtrendtowardearlierretirement may reverse itselfin the futureas a
resultof slowergrowththan anticipated in
pension wealth and reductionsin the work
disincentivesassociated with Social Securityand definedbenefitpension plans. The
currentstructureof the labor marketdoes
not appear well suited forlater retirement,
however,given the restrictedemployment
opportunities facing older workers, par-
ticularlyinjobs withhigh levels of deferred
compensation or requiringuse of computers. But labor markets and jobs change.
For example, we may see flatterwage profilesevolvein response to such thingsas the
increase in the number of older workers,
the legal restrictionson mandatoryretirement,the increase in Social Securityretirement age, and the declining importance of
defined benefit pension plans. If firms
structure compensation to correspond
older
morecloselywithcurrentproductivity,
workers' mobilityshould increase. Later
retirementis also likelyto lead employers
and employees to search for alternative
workarrangementsand increased flexibilitythatincrease thejoint value of the work
relationship. Such labor marketdevelopments have the potential to alleviate some
of the difficultiesfaced by older workers.
But there is likelyto remain a sizable number of older workers facing constrained
opportunities both within and following
theirlong-termcareer jobs.
APPENDIX
Variable Definitions and Sources
Variable
Definition
Age Distributionand Access Measures
Age at the 90th percentile of the age distribution.
P90
Age at the 50th percentile of the age distribution.
P50
Age at the 10th percentile of the age distribution.
P1O
Age5O+
CV(Age)
HireSO+
Proportion of workerswho are age 50 or over.
Coefficientof variation of the age distribution
(100 times the standard deviation divided by the
mean).
Proportion of workerswith5 years or less of company
tenure who are ages 50 and over.
Source
CPS-ORG, 1983-95.
CPS-ORG, 1983-95.
CPS-ORG, 1983-95.
CPS-ORG, 1983-95.
CPS-ORG, 1983-95.
Six CPS supplements containing
informationon company tenure
(time withcurrentemployer):
January1983, May 1983, January
1987, May 1988, January1991, and
April 1993.
Continued
JOB ACCESS FOR OLDER WORKERS
417
APPENDIX
Continued.
Variable
Definition
ExplanatoryVariables
Open5O+
Ratio Hire50+/Age50+.
ln(Wage)
Mean of log wage in 1995 dollars.
Source
CPS-ORG Male W&S, 1983-95.
alnW/alnExp Wage equation regression coefficienton log of
potential experience (Age-Schooling-6), estimated
by occupation. Schooling and other premarketcontrol
variables included in the regression. Represents wageexperience elasticity.
CPS-ORG Male W&S, 1983-95.
Pension
Proportion of workerswithemployer-provided
pension coverage.
March CPS, Male W&S, 1983-95.
Health
Proportion of workerswithemployer-providedhealth
insurance.
March CPS, Male W&S, 1983-95.
Mean years of schooling completed.
Schooling
CPS-ORG Male W&S, 1983-95.
Firm Training Proportion of workersreceiving company-provided
CPS-Supplement, Male W&S,
training.
January1983 andJanuary 1991.
Computer
Proportion of workersusing computer on thejob.
CPS-Supplement, Male W&S,
October 1984 and October 1989.
Numerical
Numerical aptitude required forjob. Rescaled to
DOT (England and Kilbourne
range from0 (low aptitude) to 4 (high aptitude).
1988).
ShiftWork
Proportion of workerswhose work shiftis not during
the day (that is, sum of evening, night,irregular,and
CPS-Supplement, Male W&S, May
rotatingshift).
1985 and May 1991.
Overtime
Proportion workinglong hours (more than 42 hours
CPS-ORG Male W&S, 1983-95.
per week).
Part-time
Proportion workingpart-time(less than 35 usual hours
per week).
CPS-ORG Male W&S, 1983-95.
Flex-time
Proportion whose work schedule allows them to vary
CPS-Supplement, Male W&S, May
the times at which theyarriveand depart fromwork.
1985 and May 1991.
Index of required strengthin occupation, ranging
Strength
DOT (England and Kilbourne
from 1 (sedentary) to 5 (veryheavy).
1988).
Environm.
Number of severe non-weatherenvironmental
DOT (England and Kilbourne
conditions, from0 to 5 (cold, heat, wetness,noise,
atmosphere).
1988).
Hazards
DOT (England and Kilbourne
Proportion ofjobs withinCPS occupation involving
significanthazard.
1988).
Number of physical demands (significantclimbing,
Physical
DOT (England and Kilbourne
Demands
stooping, reaching, seeing), from0 to 4.
1988).
Work
Proportion ofjobs where significantamount (at
Outdoors
least 25%) of workis outdoors (combines DOT variables DOT (England and Kilbourne
for outdoors and both indoors/outdoors).
1988).
Union
Proportion of workerscovered by collective bargaining
CPS-ORG Male W&S, 1983-95.
agreement.
Large Firm
Proportion of workersin firmswith 1,000+ employees. March CPS, Male W&S, 1989-95.
CPS-ORG Male W&S, 1983/84,
Employment
Growth
Ratio of 1994/95 to 1983/84 employment.
1994/95.
N is the sample size fromthe 1983-95 CPS-ORG earnings files.
418
INDUSTRIAL AND LABOR RELATIONS REVIEW
REFERENCES
Acemoglu, Daron, and J6rn-SteffenPischke. 1998.
"WhyDo FirmsTrain? Theoryand Evidence. " QuarterlyJournal
Vol. 103, No. 1 (February),
ofEconomics,
pp. 79-119.
Brown, Charles, and James Medoff. 1989. "The
Employer Size-Wage Effect." Journal of Political
Economy,
Vol. 97, No. 5 (October), pp. 1027-59.
Chan, Sewin, and Ann HuffStevens. 1999. "Employment and RetirementFollowing a Late-Career Job
Loss." AmericanEconomicReviewPapersand Proceedings,Vol. 89, No. 2 (May), pp. 211-16.
Duncan, GregJ., and Frank P. Stafford. 1980. "Do
Union Members Receive Compensating Wage Differentials?"AmericanEconomic
Review,Vol. 70, No. 3
dune), pp. 355-71.
England, Paula, and Barbara Kilbourne. 1988. Occupational MeasuresfromtheDictionaryof Occupational
Titlesfor1980 CensusDetailed Occupations.Ann ArConsortium for Politibor, Mich.: Inter-university
cal and Social Research, No. 8942, Fall.
Farber, Henry S. 1997. "The Changing Face ofJob
Loss in the United States, 1981-1995." Brookings
Papers on EconomicActivity:Microeconomics,
pp. 55128.
Filer, Randall K., and Peter A. Petri. 1988. "A JobCharacteristicsTheory of Retirement." Reviewof
Economicsand Statistics,
Vol 70, No. 1 (February),pp.
123-29.
Garen, John, Mark Berger, and Frank Scott. 1996.
"Pensions, Non-Discrimination Policies, and the
Employmentof Older Workers." Quarterly
Reviewof
Economicsand Finance,Vol. 36, No. 4 (Winter), pp.
417-29.
Gustman,Alan L., Olivia S. Mitchell, and Thomas L.
Steinmeier. 1994. "The Role of Pensions in the
Labor Market: A Surveyof the Literature." Industrialand LaborRelationsReview,
Vol. 47, No.3 (April),
pp. 417-38.
Heywood, John S., Lok-Sang Ho, and Xiangdong
Wei. 1999. "Determinants of Hiring Older Workers: Evidence from Hong Kong." Industrial and
Labor RelationsReview,Vol. 52, No. 3 (April), pp.
444-59.
Hurd, Michael D., and Kathleen McGarry. 1993.
"The Relationship betweenJob Characteristicsand
Retirement." National Bureau of Economic Research WorkingPaper 4558, December.
Hutchens, Robert M. 1986. "Delayed PaymentContractsand a Firm's Propensityto Hire Older Workers." JournalofLaborEconomics,
Vol. 4, No. 4 (October), pp. 439-57.
. 1988. "Do Job Opportunities Decline with
Age?" Industrialand LaborRelationsReview,Vol. 42,
No. 1 (October), pp. 89-99.
. 1991. "SegregationCurves,Lorenz Curves,and
Inequality in the Distributionof People across Occupations." MathematicalSocial Sciences,Vol. 21, No.
1 (February), pp. 31-51.
. 1993. "RestrictedJob Opportunities and the
Older Worker." In Olivia Mitchell, ed., As the
Workforce
Ages: Costs,Benefitsand PolicyChallenges.
Ithaca, N.Y.: ILR Press, pp. 81-102.
Ippolito, Richard A. 1987. "The Implicit Pension
Contract: Developments and New Directions." Journal ofHumanResources,
Vol. 22, No.3 (Summer), pp.
441-67.
_
1991. "Encouraging Long-TermTenure: Wage
Tilt or Pension?" Industrialand Labor RelationsReview,Vol. 44, No. 3 (April), pp. 520-35.
Johnson, Richard W., and David Neumark. 1997.
"Age Discrimination,JobSeparations, and Employment Status of Older Workers: Evidence fromSelfReports." JournalofHuman Resources,
Vol. 32, No. 4
(Fall), pp. 779-811.
Krueger,Alan B., and Lawrence H. Summers. 1987.
"Reflectionson the Inter-Industry
Wage Structure."
In Kevin Lang and Jonathan Leonard, eds., Unemand theStructure
ployment
ofLaborMarkets.New York:
Basil Blackwell, pp. 17-47.
Lazear, Edward P. 1979. "WhyIs There Mandatory
Retirement?" JournalofPoliticalEconomy,Vol. 87,
No. 6 (December), pp. 1261-84.
Levy,Frank, and RichardJ. Murnane. 1992. "Earnings Levels and Earnings Inequality: A Review of
Recent Trends and Proposed Explanations." Journal ofEconomicLiterature,
Vol. 30, No. 3 (September), pp. 1333-81.
Medoff,James L., and Katharine G. Abraham. 1980.
"Experience, Performance, and Earnings." QuarVol. 95, No. 4 (December),
terlyJournal
ofEconomics,
pp. 703-36.
Ruhm, ChristopherJ. 1990. "BridgeJobsand Partial
Vol. 8, No.
Retirement."JournalofLaborEconomics,
4 (October), pp. 482-501.
. 1996. "Do Pensions Increase the Labor Supply
of Older Men?" JournalofPublicEconomics,Vol. 59,
No. 2 (February), pp. 157-75.
Scott, Frank A, Mark C. Berger, and John E. Garen.
1995. "Do Health Insurance and Pension Costs
Reduce theJob Opportunities of Older Workers?"
Industrialand Labor RelationsReview,Vol. 48, No. 4
(July),pp. 775-91.
U.S. Bureau ofLabor Statistics.1999. Employment
and
Earnings. Washington, D.C.: GPO (January).