The Employment of People with Disabilities: Federal Data Sources and Trends
The Employment of People with Disabilities:
Federal Data Sources and Trends
Updated January 22, 2007
Dennis M. Roth
Specialist in Labor Economics
Domestic Social Policy Division
The Employment of People with Disabilities:
Federal Data Sources and Trends
Congress has enacted major legislation related to the employment of individuals
with disabilities. In 1938, the Congress passed the Fair Labor Standards Act (FLSA),
which, among other provisions, established a reduced wage for the employment of
individuals whose earning capacity was impaired by age or physical or mental
deficiency or injury (Section 14). The Americans with Disabilities Act (ADA) of
and the Civil Rights Act (CRA) of 1991 allowed people bringing employment
discrimination suits to seek compensatory and punitive damages under the ADA and
the Rehabilitation Act of 1973. In 1999, Congress turned its attention to the supply
side of the labor market in the Ticket to Work and Work Incentives Improvement Act
by encouraging people with severely limiting health conditions or impairments to
increase their work effort without loss of government-provided health care coverage.
Learning whether public policy is on the right track is especially important today
because of the direct relationship between age and disability. As the incidence of
disabling health conditions rises with age and as the large baby-boom generation has
been entering middle-age, disability is likely to become increasingly common and
(potentially) costly to society. Changing economic conditions and the varying
definitions of disability in programs and surveys make it difficult to gauge the
success of the legislation in enhancing the employment prospects of people with
disabilities. So, too, does the lack of accurate, consistent statistics over time on the
labor force status of individuals with disabilities. Despite the charge of the
Presidential Task Force on Employment of Adults with Disabilities (1998-2002) to
develop a valid, reliable measure of the employment rate of working-age adults with
disabilities, progress toward that end has been very slow.
To mitigate the data shortcomings with regard to conducting trend analysis,
researchers have utilized multiple databases to determine if they are telling the same
story. In general, they are. Despite the strong state of the economy through 2000,
full implementation of the ADA’s employment provisions in the mid-1990s, and the
employment rate of all working-age persons with disabilities failed to improve in
recent decades. Analysts have examined a variety of explanations for the lagging
employment opportunities of people with disabilities, including the seemingly
increased share of adults with disabilities so severe that they report being unable to
work; expansion of the Social Security Disability and Supplemental Security Income
programs; and enactment and implementation of the ADA.
The results of this research have prompted some to suggest that enhancing the
employment prospects of all members of the population with disabilities might
involve subsidizing the cost of employer accommodations and helping those already
employed retain their jobs, among other things. This report will be updated as
In troduction ......................................................1
Actions by the Congress and Administration............................2
The Data Dilemma.................................................5
Federal Databases on Disability and Labor Force Status................6
Census of Population and
the American Community Survey (ACS)...................6
Current Population Survey (CPS).............................6
Survey of Income and Program Participation (SIPP)...............7
National Health Interview Survey (NHIS).......................9
Other Differences Between the Federal Data Sources.................10
How Much Progress Has Been Made in the Development
of Timely, Accurate, and Reliable Employment Statistics
on Persons with Disabilities?................................11
The Employment Status of Persons with Disabilities in Recent Decades......12
The Impact of Changing Economic Conditions, 1981-1992............12
What Happened During the 1990s Economic Expansion?.............13
List of Tables
Table 1. Percent Change in the Labor Force Participation Rate
of 18- to 64-Year-Olds by Disability Status, 1981-1992...............12
Table 2. Employment Rate of 18- to 64-Year-Olds by Disability Status,
The Employment of People with Disabilities:
Federal Data Sources and Trends
The Congress enacted major legislation concerning the employment of people
with disabilities as early as the 1930s through the establishment of wage rates for
such individuals below the federal minimum wage. This and subsequent laws share
a focus on individuals with disabilities as current or potential workers. Learning
whether public policy is on the right track is particularly important at the current time
because of the positive relationship between age and disability. As
!the incidence of disabling health conditions rises with age (e.g.,
hearing impairments, arthritis, and limitations on activities due to
heart disease) and
!as the large baby-boom generation has been entering the mid-to-late
disability is likely to become increasingly common and, potentially, increasingly
costly to society (e.g., increased utilization of medical and rehabilitative services as
well as decreased receipt of income and payroll taxes from forgone individual2
earnings). Moreover, new advances in medicine have allowed people to live with
previously life-threatening impairments. And, among students — who typically are
expected to go from school into the workplace — those served by the Individuals
with Disabilities Education Act (IDEA), Part B have increased over time.3
This report begins with a summary of federal actions taken during the 20th
century related to the employment of people with disabilities. It then describes in
detail the existing data sources of government agencies that cover disability and
employment status as well as reviews the activities of the Presidential Task Force on
Employment of Adults with Disabilities aimed at remedying statistical shortcomings.
1 This report was originally prepared by Linda Levine, a specialist in labor economics in the
Domestic Social Policy Division of CRS, and has been updated by the current author on the
2 The Committee on Assessing Rehabilitation Science and Engineering of the Institute of
Medicine estimated that lost wages from disability in 1994 ranged from $93.1 billion to
$111.1 billion depending on the definition of disability. In addition, medical costs
associated with disability may have totaled $163.1 billion (in 1994 dollars). Edward N.
Brandt Jr. and Andrew M. Pope, ed., Enabling America: Assessing the Role of
Rehabilitation Science and Engineering (Washington, DC: National Academy Press, 1997).
3 Presidential Task Force on Employment of Adults with Disabilities, Re-charting the
Course, November 1998.
The report closes with an examination of the limited trend statistics on the
employment situation of working-age adults with disabilities to determine how they
fared in recent decades and mentions policies that have been suggested.
Actions by the Congress and Administration
The National Industrial Recovery Act (NIRA) of 1933-1935 addressed the
employment disadvantages of disabled individuals4 by establishing a productivity-5
based sub-minimum wage arranged through a system of certificates. With time,
these individuals came to be identified in one of three classifications:
!employees of sheltered workshops where the wage floor was set by
a charitable institution,
!employees in the private for-profit sector where the wage floor could
not be “less than 75 percent of the minimum wage” in the industry,
!employees who by virtue of disability were employed in industrial
The NIRA was declared unconstitutional in May 1935, and the above requirements
were no longer operative.
In 1938, the Congress passed the Fair Labor Standards Act (FLSA). Section 14
of the act established a reduced wage for the employment of individuals, under
special certificates issued by the U.S. Department of Labor’s Wage and Hour
Administrator, whose earning capacity was impaired by age or physical or mental
deficiency or injury. No statutory floor was established, but the wage floor for
disabled workers at private for-profit firms was administratively set at not less than
75% of the standard federal minimum wage. Workers in sheltered workshops run by
a charitable organization were to be paid on the basis of their earning capacity.
Under the 1966 FLSA amendments, the wage rate for persons with disabilities was
statutorily set at not less than 50% of the basic FLSA minimum for both competitive
industry and sheltered workshops, with an exception for work activity centers (P.L
89-601). In 1986, section 14(c) was again amended, removing the distinction
between workshops and work activity centers and eliminating the statutory minimum
wage for the disabled (P.L. 99-486). Instead, the wage rate was to be “commensurate
with those paid to nonhandicapped workers, employed in the vicinity ... for
essentially the same type, quality, and quantity of work,” and that was “related to the
The Americans with Disabilities Act of 1990 (ADA, P.L. 101-336) bans
discrimination in the workplace and elsewhere (e.g., in public accommodations and
4 Persons whose earning capacity was limited because of age, physical or mental handicap,
or other infirmity.
5 For more information see CRS Report RL30674, Treatment of Workers with Disabilities
Under Section 14(c) of the Fair Labor Standards Act, by William G. Whittaker.
services operated by private entities, transportation, and in telecommunications) on
the basis of disability.6 Title I primarily is intended to equalize access to employment
opportunities for otherwise qualified workers with disabilities by requiring employers
of 15 or more employees to make reasonable accommodations that would not cause
undue hardship to their operations. (The provisions banning employment
discrimination by employers with at least 25 employees went into effect in July 1992.
The ban’s extension to firms with 15 to 25 employees became effective in July 1994.)
In addition, the Civil Rights Act of 1991 (CRA, P.L. 102-166) allows people
bringing employment discrimination suits to seek compensatory and punitive
damages under the ADA and the Rehabilitation Act of 1973 (P.L. 93-112).
The FLSA makes disabled individuals more attractive to hire by creating a lower
wage rate for them. The ADA and CRA regard private and public employers who
fail to hire, retain, and promote members of the population with disabilities, at rates
comparable to other equally qualified workers, as contributors to the group’s poor
employment and earnings outcomes. Some believe that the success of the laws’
employment provisions is best determined by whether the employment rate of
working-age adults with disabilities increases, because lack of job opportunities
appears to be an even bigger problem than wage discrimination for this protected
class.7 Yet, it is no easy matter to separate the impact of legislation from a host of
other factors that are operating simultaneously (e.g., changes in economic conditions,
in private sector provision of health benefits, and in job requirements away from
manual labor and toward higher levels of educational attainment as well as toward
Toward the close of the decade, the Congress turned its attention to the supply
side of the labor market in the Ticket to Work and Work Incentives Improvement Act
of 1999. Among other things, P.L. 106-170 is intended to encourage persons with
severely limiting health conditions or impairments to increase their work effort by
alleviating a dilemma of Social Security Disability Insurance (SSDI) and
Supplemental Security Income (SSI) beneficiaries, namely, choosing between a job
which may not have employer-provided health benefits and continuation of
government-provided health care coverage through Medicare or Medicaid.8 These
provisions became effective in October 2000. The legislation also creates a voucher
program, to be phased in over a three-year period beginning in 2001, that allows
SSDI and SSI beneficiaries to obtain employment, rehabilitation and other services
6 The ADA applies to private employers as well as to state and local governments. The
federal government, federal contractors, and other entities that receive federal funds
previously were banned from discriminating against qualified persons with disabilities under
the Rehabilitation Act of 1973. For more information, see CRS Report 98-921, The
Americans with Disabilities Act (ADA): Statutory Language and Recent Issues, by Nancy
7 See, for example, Marjorie Baldwin and William G. Johnson, “Labor Market
Discrimination Against Men with Disabilities in the Year of the ADA,” Southern Economic
Journal, vol. 66, no. 3 (Jan. 2000).
8 If SSDI or SSI beneficiaries earn above the “substantial gainful activity” level, they lose
health care coverage through Medicare or Medicaid.
from providers of their own choice to assist them to get jobs or increase their
Recent Administrations also have undertaken various initiatives that focus on
the poor labor market outcomes of individuals with disabilities. Executive Order
13078, issued by President Clinton on March 13, 1998, created the Presidential Task
Force on Employment of Adults with Disabilities. It was charged with seeking ways
to increase the share of the adult population with disabilities that has a job, among
other things. That same year, the U.S. Department of Education’s National Institute
on Disability and Rehabilitation Research (NIDRR) announced funding, under the
Rehabilitation Act of 1973, for six new Rehabilitation Research and Training Centers
to develop information that addresses the employment-related issues of persons with
disabilities.10 In 2000, the U.S. Department of Labor (DOL) requested applications
for work incentive grants to assist states creating their One-Stop delivery systems
under the Workforce Investment Act of 1998 (P.L. 105-220) to provide
comprehensive employment and training services to people with disabilities. And,
the following year, the Office on Disability Policy was established within the DOL.
The relatively recent implementation of these policies and programs makes it
difficult to gauge how successful they have been at enhancing the employment
prospects of persons with disabilities. The varying definitions of disability contained
in programs and surveys compound the problem.11 Perhaps most importantly, so,
too, does the lack of accurate statistics on individuals with disabilities derived from
surveys that have asked consistent questions over time. The National Council on
Disability noted in 1996 that despite
considerable discussion in the last decade about the limitations of ... databases
and the need for a database specifically designed to answer policy-relevant
questions about people with disabilities ... little improvement ... has been12
9 For more information, see CRS Report RL31157, Ticket to Work and Work Incentives
Improvement Act of 1999: Implementation Status, updated by Karen Tritz.
10 The NIDRR was established in 1978, as the National Institute of Handicapped Research,
through amendment of the Rehabilitation Act of 1973. It is part of the Education
Department’s Office of Special Education and Rehabilitative Services. NIDRR leads the
Interagency Committee on Disability Research.
11 Under the ADA, for example, an individual meets the definition of disability if he/she has
a physical or mental impairment that substantially limits one or more major life activities,
has a record of such impairment, or is regarded as having such an impairment. Precisely
who fits this definition is unclear, however, as indicated by recent court decisions. (For
more information, see CRS Report RL33304, The Americans with Disabilities Act: The
Definition of Disability, by Nancy Lee Jones.) Disability is much more narrowly defined
in the SSDI and SSI programs, where the emphasis is on one’s capacity to work. To qualify
as having a disability under the programs, a person must be unable to engage in any
substantial gainful activity because of a medically determinable physical or mental
impairment that can be expected to result in death or has lasted/is expected to last for a
continuous period of at least 12 months.
12 National Council on Disability, Achieving Independence: The Challenge for the 21st
In 2004, the Council effectively concluded that the situation had not much changed.13
The Data Dilemma
One of the charges of E.O. 13078 recognized the data shortcomings related to
the labor force status of persons with disabilities. It required the Presidential Task
Force on Employment of Adults with Disabilities to design and implement
a statistically reliable and accurate method to measure the employment rate of
adults with disabilities as soon as possible, but no later than the date of
termination of the Task Force [30 days after submitting its final report on Julyth
26, 2002, the 10 anniversary of initial implementation of the ADA’s
employment provisions]. Data derived from this methodology shall be published
on as frequent a basis as possible.
The Bureau of Labor Statistics and Census Bureau, in cooperation with the
Departments of Education and Health and Human Services, the National Council on
Disability, and the President’s Committee on Employment of People with
Disabilities, have been among the numerous organizations working toward this end.
The accomplishment of this charge has been complicated by the fact that:
(1) a universally agreed-upon definition of disability does not exist, and
(2) a data source is lacking that incorporates both labor force and disability
questions which have undergone testing to determine whether people understand
what is being asked (i.e., validity) and to which people respond in a consistent
manner over time (i.e., reliability).
At the present time, the U.S. Bureau of the Census collects self-reported
information from individuals on the disability and labor force status of household
members through the decennial census, American Community Survey, Current
Population Survey, Survey of Income and Program Participation, and National Health
Interview Survey.14 Disability typically is not the primary focus of these surveys.
Century (Washington, D.C.:U.S. Government Printing Office, 1996), pp. 26-27. Note: The
National Council on Disability was created in 1978 by P.L. 95-602 as an advisory board in
the Department of Education. It became an independent agency with passage of the
Rehabilitation Act Amendments of 1984 (P.L. 98-221). The Council is charged with
promoting policies, programs and practices that afford equal opportunity to those with
disabilities and that enable them to attain self-sufficiency as well as to become integrated
in all aspects of society.
13 National Council on Disability, Improving Federal Disability Data, Jan. 8, 2004.
14 Unless otherwise indicated, information in this section was drawn from conversations
with staff at the U.S. Bureau of Labor Statistics, the U.S. Bureau of the Census, and the
National Center for Health Statistics, as well as from the following websites
[http://www.census.gov/hhes/www/disability.html]; and [http://www.cdc.gov/nchs].
Federal Databases on Disability and Labor Force Status
Census of Population and the American Community Survey (ACS).
Disability questions are included on the long-form of the decennial census of
population. The number and nature of disability questions have differed over time.
The 1970 form asked about work disability; the 1980 form, about work disability and
the ability to use public transportation; and the 1990 form, about work disability, the
ability to go outside the home alone to shop or visit a doctor’s office, and about self-
care (e.g., bathing or dressing without assistance). The work disability questions
were fairly consistent from 1970 through 1990 (i.e., having a physical, mental or
other health condition that prevents an individual from working, or that limits the
amount/kind of work an individual can do).
The disability questions were substantially revised on the 2000 long-form based
on the judgment of a panel of experts and follow-up testing. Households that
received the revised questionnaire were asked whether a physical, mental or an
emotional condition of at least six months duration creates difficulty in working at
a job or business; going outside the home alone to shop or visit a doctor’s office;
dressing, bathing, or getting around inside the home; or learning, remembering, or
concentrating. In addition, a question was included on whether respondents have
long-lasting sensory impairments (e.g., blindness) or substantial limitations on
physical activities (e.g., reaching or climbing stairs).
The Census Bureau intends that the long-form of the decennial census be
replaced by the more frequently administered, comparatively new ACS. The
disability questions in the 1999-2002 ACS were changed from their original (1996-
1998) version to agree with those on the 2000 long-form. The ACS questionnaire
was changed in 2003 to address the seeming over-reporting of disability in the 2000
census and in earlier surveys by moving the questions’ location and altering
instructions to reduce misinterpretation. It appears that these changes have reduced
the presumed over-reporting.15 In addition, the Census Bureau will be conducting a
national content test in 2006 which, among many other things, will determine
whether new disability questions will be introduced in the 2008 ACS.
Thus, at present, neither the decennial census nor the ACS can provide data
suitable for analyzing changes over time in the employment status of persons with
Current Population Survey (CPS). The purpose of the CPS is to enable
the U.S. Bureau of Labor Statistics (BLS) to provide timely information on the labor
force status (i.e., employed, unemployed, or not in the labor force) and characteristics
(e.g., demographic and occupational) of the population. Rather than replying yes or
no to the question about whether they had worked in the preceding week, respondents
to the monthly (basic) CPS sometimes say they are disabled. If, based on additional
questions, it is determined that an individual who said they are disabled meets the
15 Sharon Stern and Matthew Brault, Disability Data from the American Community Survey:
A Brief Examination of the Effects of a Question Redesign in 2003, U.S. Census Bureau,
Housing and Household Economics Statistics Branch, Jan. 28, 2005.
criteria for being employed or unemployed, the response “disabled” is deleted from
the record. “Because of this, researchers have no information on the disability status
of employed and unemployed persons from the basic CPS.”16
In March, the CPS includes supplementary questions about income.
Respondents have been asked annually since March 1980 whether a health problem
or disability prevents them from working or limits the kind or amount of work they
can do. It is much like the work disability question that appeared on the 1970-1990
census long-form. Those who say they have a work disability are then asked whether
they receive income based on the disability (e.g., workers’ compensation, the SSDI
and SSI programs, and the Black Lung program).
Some researchers cross tabulate the work limitation response from the March
supplement with the labor force status question in the basic CPS to develop
employment rates for persons with disabilities. Because the basic and March
supplement questions were not designed for this purpose, testing to determine
whether they provide a valid and reliable estimate of the number of people with work
disabilities has never been undertaken. The BLS therefore considers use of the
current version of the CPS for this purpose to be inappropriate. While
acknowledging that “work limitation-based questions are not the ideal way to identify
the size of the working-age population with disabilities,” some analysts nonetheless
“argue that nationally representative employment-focused data sets like the CPS can
be used to monitor trends in the employment outcomes of the working-age
population with disabilities.”17 (Emphasis added.)
Survey of Income and Program Participation (SIPP). The survey
gathers demographic and economic data by asking a core set of questions that is
periodically supplemented with questions on different subjects (e.g., the health and
disability topical module). The two principal ways that have been used to define
disability in surveys are in terms of (1) limitations on sensory, physical or mental
functioning and (2) limitations on activities (e.g., paid employment) due to chronic
health conditions or impairments.18 By and large, the SIPP relies on the former
approach and the National Health Interview Survey (described below), on the latter.
Some observers believe that a definition based on functional limitations is superior
to one based on activity limitations. Both approaches generally rely on the medical
model of disability that focuses on individuals rather than on environmental factors
which not only can affect people’s actual performance but also how those with
equally disabling conditions reply to survey questions.
16 Thomas W. Hale, “The Lack of a Disability Measure in Today’s Current Population
Survey,” Monthly Labor Review, June 2001, p. 38.
17 Richard V. Burkhauser, Mary C. Daly, Andrew J. Houtenville, and Nigar Nargis, “Self-
Reported Work-Limitation Data: What They Can and Cannot Tell Us,” Demography, vol.
18 A third approach less often taken to define disability is receipt of SSDI and SSI benefits.
The number of beneficiaries will be less than the number of persons who report they have
functional or activity limitations because the government programs require that disability
be proved medically and because some eligible individuals may not apply for the programs.
Respondents to the SIPP are considered to have a disability if they indicate that
they experience difficulty performing certain sensory or physical functions (e.g.,
seeing words in newspapers or lifting an object weighing 10 pounds); activities of
daily living, ADLs (e.g., eating or getting around inside the home); or instrumental
activities of daily living, IADLs (e.g., going outside the home or keeping track of
bills). People who indicate that they have a physical, mental or other health
condition that prevents them from or restricts the kind/amount of work they can
perform around the house or at a job also are deemed to have a disability. In
addition, respondents who state that they use assistive devices (e.g., wheelchairs or
canes), have specific conditions (e.g., developmental disorders or Alzheimer’s
disease), or are under age 65 and covered by certain benefit programs (e.g., SSI) are
considered to have a disability.
Over the years, there has been some cognitive pre-testing of the SIPP disability
questions to ascertain their validity. It has long been argued, however, that asking
people whether they are limited in the kind/amount of work they can do lacks “face
validity” because their responses would depend upon environmental conditions (e.g.,
the presence of workplace accommodations).19 For this reason, a question was added
to allow respondents to indicate whether they have a condition which makes finding
a job or remaining employed difficult without stating that they are “limited”
regardless of environmental considerations. The SIPP’s disability module and core
questionnaire have been fairly consistent since 1990, but as just discussed, questions
were added to try to better capture the meaning of disability.20
The Presidential Task Force on Employment of Adults with Disabilities found
that the SIPP produced sizeable fluctuations in the disability status of individuals
within just a year’s time,21 which suggests that at least some of its disability estimates
lack reliability.22 As individuals in the SIPP are interviewed at four-month intervals
over two or more years, the longitudinal survey is subject to attrition and to time-in-
sample bias,23 both of which can adversely affect the quality of estimates derived
from one period to the next. Moreover, a change in administration of the work
disability question to the 1996 panel as opposed to earlier panels of respondents has
yielded noncomparable estimates over time. These observations led a Census Bureau
analyst to conclude in 2000 “that there is currently no satisfactory vehicle for
producing” a measure “over time of the employment status of individuals with
19 John M. McNeil, Employment, Earnings, and Disability, paper prepared for the 75th
Annual Conference of the Western Economic Association International, June 29-July 3,
20 Other added questions concern having difficulty using hands/fingers and having mental
problems that seriously interfere with everyday activities.
21 Presidential Task Force on Employment of Adults with Disabilities, The Second Report,
22 McNeil, Employment, Earnings, and Disability.
23 Attrition has to do with an increase in the non-response rate among sample units over the
life of a survey. Time-in-sample bias has to do with the propensity of individuals to offer
different replies to the same questions depending on the number of times they have been
asked those questions.
disabilities.”24 Other analysts dispute this contention. They have shown “that
certain questions from select [comparable] interviews can be used in trend
analyses.”25 That is to say, the problem may reside less with the questions
themselves than with the way analysts have utilized the SIPP.
National Health Interview Survey (NHIS). The purpose of the National
Center for Health Statistics’ annual survey is to provide information on chronic
health conditions, health status, health care utilization, and disability. In addition,
several federal agencies and a private foundation funded a two-part disability
supplement that was conducted along with the 1994 and 1995 NHIS. As the most26
comprehensive survey on disability since 1978, it contains a plethora of information
that allows construction of a variety of definitions for disability. The supplement’s
questionnaire was widely reviewed by interested parties in the disability community,
voluntarily tested by persons with disabilities in the National Center’s cognitive
questionnaire lab, and pretested by some 250 households.
The annual NHIS is ongoing. It asked consistent disability and labor force
questions between 1982 and 1996. Although the survey underwent a major revision
in 1997, some of the earlier disability questions were carried forward. Results from
the annual NHIS, like those from the disability supplement, typically are made
available for tabulation and analysis by members of the research community rather
than regularly being released in Center publications.
In the annual NHIS prior to 1997, individuals were considered to have
disabilities if they reported a chronic condition or an impairment that prevented them
from or limited their ability to perform age-dependent major life activities (e.g.,
attending school in the case of children age 5-17 or taking care of oneself in the case
of someone at least 70 years old), or limited their ability in any way to perform any
other activities (e.g., social or family pursuits). Respondents aged 18-69 who
reported that an impairment or health problem completely prevented them from
working or restricted the kind/amount of work they could perform were deemed to
have work limitations. The NHIS also asked people if they needed assistance in
performing ADLs or IADLs. As in the SIPP, respondents with limitations who
needed assistance did not necessarily indicate that they had conditions or
impairments that restricted their ability to work. In those instances where individuals
identified an activity limitation, they were asked about the specific health condition
that led to the limitation and how long it lasted.
Beginning with the redesigned 1997 NHIS questionnaire, some changes were
made to the above-described items. For example, the questions which determine
24 McNeil, Employment, Earnings, and Disability, pp. 17-18.
25 Elaine M. Maag and David C. Wittenburg, Real Trends or Measurement Problems?
Disability and Employment Trends from the Survey of Income and Program Participation,
Washington, DC: The Urban Institute, May 2003, p. 26. (Hereafter cited as Maag and
Wittenburg, Real Trends or Measurement Problems?.)
26 Intermittently between the early 1960s and 1978, the Social Security Administration
conducted the Survey of Disability and Work to study the prevalence of disability in the
working-age population and the experience of SSDI beneficiaries and their families.
work limitation were reworded to ascertain if a physical, mental, or emotional
problem prevents or restricts the kind/amount of work a person can perform; to
eliminate an upper age limit; and to allow acute in addition to chronic conditions as
causes of work disability. Questions similar to those in the SIPP were added to
ascertain the presence of functional limitations in a context that is not activity
specific (e.g., the degree of difficulty walking or lifting). The questions are asked
only for one randomly chosen adult per family included in the survey.
Other Differences Between the Federal Data Sources
The timeliness of results differ. The population census is conducted once every
10 years and statistics derived from it are released with a considerable lag, a
shortcoming that the ACS was developed in part to overcome. Information from the
annual NHIS and the SIPP are available with a fairly lengthy lag as well. In contrast,
results from the March CPS supplement are released about six months after it is
conducted, and selected labor force data from the monthly CPS are published the
month after it is conducted.
The composition of the samples also differs. The labor force questions in the
redesigned NHIS are administered only to those respondents age 18 or older; in the
decennial census, ACS, and SIPP, to those age 15 or older; and in the CPS, to
persons age 16 or older. Moreover, unlike the NHIS and CPS, both of which cover
the civilian noninstitutional population, the SIPP includes members of the civilian
and military resident population (excluding Armed Forces personnel living in
military barracks). These three surveys exclude the institutionalized population,
which is covered by the decennial census and is expected to be covered by the ACS.
The questions in the surveys also have reference periods of varying length which
can affect an individual’s labor force status. SIPP respondents are asked about their
activities over a four-month reference period during which lengthy interval a person
could very well have been employed and unemployed. Before the 1997 revision,
NHIS respondents were asked about their labor force activity during the two-week
period preceding the interview. The reference period subsequently was changed to
the week before the interview is held. In the CPS, an individual is counted as
employed if he/she worked for pay or profit for one hour or, in the case of a family
business, for 15 hours on an unpaid basis, during the one-week reference period that
precedes the survey week; or as unemployed, if he/she was available for work and
“actively” sought a job during the four weeks preceding the survey week. As neither
the SIPP nor the NHIS has the same precision as the CPS in determining labor force
activity, an individual who is deemed “out of the labor force” in the CPS could be
counted as an unemployed worker in the two other surveys.27
27 For example, while in all three surveys a person without a job must look for work to be
classified as unemployed, the CPS specifies the actions the individual must undertake to
fulfill its definition of “actively” seeking employment. If the person does not take one of
these actions, he/she will be counted in the CPS as a non-participant in the labor force; in
the NHIS and SIPP, as an unemployed member of the labor force.
Although these various factors can impact the quality of survey estimates,
perhaps the most important factor that affects efforts to collect quantitative
information on persons with disabilities is one of definition. Some observers have
work disability is a multidimensional, complex concept that investigators do not
know how to define well. ... Disability or limitations in participation are a result
of a complex interaction among an individual and his or her attributes, the
physical and social environment, and the accommodations and barriers to
participation. The measurement of disability is further confounded by the fact
that all characteristics, whether intrinsic individual characteristics or28
characteristics of the environment, change with time.
How Much Progress Has Been Made in the Development
of Timely, Accurate, and Reliable Employment Statistics
on Persons with Disabilities?
The Presidential Task Force established the Employment Rate Measurement
Methodology Work Group (ERMM), which is composed of some 15 federal
agencies. The ERMM has been working on implementing E.O. 13078’s charge
concerning the development of timely, accurate, and reliable labor force data for
persons with disabilities.
The ERMM’s long-term goal has been incorporation of questions in the CPS
that would enable the release of labor force estimates for people with disabilities as
is done for other population groups (e.g., women and racial/ethnic minorities).
Toward that end, after they had undergone extensive testing, numerous questions
on disability and labor force status were asked as part of the National Comorbidity
Survey (NCS) in 2001 and 2002.29 The hope was that a small subset would perform
as well as the full battery of questions in order to minimize the burden that would be
placed on respondents were disability questions integrated in the basic CPS.
Analysis of the questions was conducted by outside experts, and the best seven
questions were selected.30 This set of questions underwent further testing before BLS
presented it and the related research to the ERMM. The ERMM raised some
concerns that were addressed through further testing. All testing outside the context
of the CPS was completed in 2004. The final test is inclusion of the question set in
the February 2006 supplement to the CPS. After the supplement receives clearance
from the Office of Management and Budget and if it subsequently is determined that
the questions perform well without too adversely affecting the response rate to the
28 Nancy Mathiowetz and Gooloo S. Wunderlich, ed., Survey Measurement of Work
Disability: Summary of a Workshop (Washington, D.C.: National Academy Press, 2000),
29 The National Comorbidity Survey is conducted by Ron Kessler at Harvard Medical
School’s Department of Health Care Policy. Primary funding for the survey comes from the
National Institute of Mental Health.
30 Terence M. McMenamin, Tom Hale, Haejin Kim, and Douglas Kruse, Designing
Questions to Identify People with Disabilities in Labor Force Surveys: A History of the
Work of BLS to Measure the Employment Level of Adults with Disabilities, draft.
survey, they could be added to the monthly CPS. This might occur in 2007, at the
The Employment Status
of Persons with Disabilities in Recent Decades
As just discussed, all of the available federal surveys that collect data both on
persons with disabilities and on labor force status suffer from various shortcomings.
In an effort to mitigate the data limitations, researchers sometimes have taken the
approach of utilizing multiple surveys to determine whether they are telling the same
story. Generally speaking, they are.
The Impact of Changing Economic Conditions, 1981-1992
As shown in Table 1, the proportion of the population with disabilities that was
in the labor force (i.e., employed or unemployed) dropped by 4.5% during the
recession of the early 1980s. At the same time, the labor force participation rate of
the population without disabilities declined by a lesser degree (1.0%). Similarly,
during the recession of the early 1990s, the labor force participation rate of the
working-age population with disabilities fell to a greater degree relative to the rate
among people without disabilities (4.1% and 0.8%, respectively). Conversely, over
the 1983-1990 period when the economy was expanding, the labor force activity of
individuals with disabilities rose much more than activity among individuals without
disabilities (14.0% and 8.8%, respectively).
Table 1. Percent Change in the Labor Force Participation Rate
of 18- to 64-Year-Olds by Disability Status, 1981-1992
Percent change in the share of the civilian
Periodnoninstitutional population that is in the labor force
With disabilitiesWithout disabilities
1981-1983 -4.5 -1.0
1983-1990 14.0 8.8
1990-1992 -4.1 -0.8
Source: Analysis of NHIS data as shown in Edward H. Yelin and Patricia P. Katz, “Labor Force
Trends of Persons With and Without Disabilities,” Monthly Labor Review, Oct. 1994.
One might conclude, based upon these data, that the labor force status of
persons with disabilities is more sensitive to the business cycle than that of persons
without disabilities. In other words, persons with disabilities appear to be harmed
more by downswings in the economy and helped more by economic upswings. As
discussed in the following section, however, this relationship seemingly did not hold
up during the strong and lengthy economic expansion that followed the 1990-1991
What Happened During the 1990s Economic Expansion?
The 1990s could have been among the best of times for the working-age
population with disabilities because of (in no particular order) enactment of the
previously mentioned legislation, advances in assistive technologies, and the
existence of an extremely tight labor market.31 Not only would robust economic
conditions be expected to encourage people who might have given up looking for
jobs to reenter the labor force, but also a dwindling supply of workers would be
expected to make firms more willing to hire and, if necessary, train persons whom
they previously might not have considered or might have avoided. The timing of
employers’ clamoring for workers could have been particularly fortuitous for the
large minority of adult beneficiaries with disabilities in the Temporary Assistance to
Needy Families (TANF) program who face time-limited cash assistance under the
Personal Responsibility and Work Opportunity Reconciliation Act (P.L. 104-193).32
As shown in Table 2, this does not appear to have been the case. The employed
share of 18-64 year olds with disabilities, broadly defined, did not change
significantly through 1996, according to data from the NHIS. (See column 3.) CPS
data on work disability reveal the same lack of improvement from the mid-1990s33
through the remainder of the decade as the labor market tightened further.
Additionally, based upon a variety of definitions of disability available in the SIPP,
the employment status of 25-61 year olds with disabilities was found to have
deteriorated over the 1990-1996 period.34 In contrast, the proportion of the working-
age population without disabilities who had jobs rose substantially according to data
from all three surveys. Defying expectations, then, the employment situation of
persons with disabilities vis-a-vis persons without disabilities seemingly worsened
during the economic expansion of the 1990s.
31 Michelle Conlin, “The New Workforce: A Tight Labor Market Gives the Disabled the
Chance to Make Permanent Inroads”; and John Williams, “Enabling Technologies,” both
in Business Week, Mar. 20, 2000.
32 Almost 46% of individuals age 21-64 in the 1996 panel of the SIPP who reported
receiving TANF benefits also reported that they had disabilities. As unmarried women with
children under the age of 18 appear more likely to receive TANF benefits if the mothers
have a disability, disability may have “an important role in determining whether individuals
become TANF recipients.” McNeil, Employment, Earnings, and Disability.
33 H. Stephen Kaye, “Improved Employment Opportunities for People with Disabilities,”
Disability Statistics Report 17, (Washington, DC: U.S. Department of Education, National
Institute on Disability and Rehabilitation Research, May 2003). (Hereafter cited as Kaye,
Improved Employment Opportunities for People with Disabilities.)
34 Maag and Wittenburg, Real Trends or Measurement Problems?. See also Richard V.
Burkhauser, Andrew J. Houtenville, and David C. Wittenburg, “A User’s Guide to Current
Statistics on the Employment of People with Disabilities,” in David C. Stapleton and
Richard V. Burkhauser, ed., The Decline in Employment of People with Disabilities: A
Policy Puzzle (Kalamazoo, MI:W.E. Upjohn Institute for Employment Research, 2003).
(Hereafter cited as Stapleton and Burkhauser, The Decline in Employment of People with
Table 2. Employment Rate of 18- to 64-Year-Olds
by Disability Status, 1988-1996
Employed Share of the Civilian Noninstitutional Population
With disabilities, anyaWith disabilities, ableb
YearWithout disabilitiesactivity limitation to work
1988 79.0 48.1 70.5
1989 79.4 49.0 71.0
1990 79.1 49.0 71.8
1991 78.1 47.9 70.8
1992 78.5 47.0 70.2
1993 78.9 47.9 70.6
1994 79.6 47.7 72.2
1995 80.2 48.2 73.3
1996 80.5 46.6 72.3
Source: NHIS data as shown in H. Stephen Kaye, “Improved Employment Opportunities for People
with Disabilities,” Disability Statistics Report 17, Washington, DC: U.S. Department of Education,
National Institute on Disability and Rehabilitation Research, May 2003.
a. This definition of persons with disabilities includes survey respondents who said that they had a
health condition or impairment that limited their performance of major life activities (e.g., paid
work or housework) or other life activities (e.g., family, community or social pursuits) as well
as persons who said that they had a health condition or impairment that prevented them from
working (i.e., severe work disabilities).
b. This definition of persons with disabilities includes survey respondents who said that they had a
health condition or impairment that limited the kind/amount of work they could do but did not
prevent them from working.
A much different picture of the decade emerges based upon unemployment data.
Both NHIS and CPS data suggest that the decrease in unemployment among labor
force participants was significant and steeper for persons with than without
disabilities.35 The explanation for this seeming paradox is that the employment rate
is a population-based measure, and some individuals in the population are
uninterested in working (e.g., retirees) or consider themselves unable to work (e.g.,
due to poor health). In contrast, the unemployment rate is based upon a more
narrowly defined group of people — those who already have or are actively seeking
jobs, that is to say, those who want to and who consider themselves able to work. An
alternative employment rate of people with disabilities who report they are willing
and able to work shows statistically significant increases during the 1990s expansion.
(See column 4 in Table 2.) Using this alternative measure as the standard, the
35 Kaye, Improved Employment Opportunities for People with Disabilities.
employment gap between persons with and without disabilities shrank substantially
during the 1990s expansion.
One explanation that has been posited for the poor labor market performance of
the aggregate population with disabilities during the last decade is a change over time
in its composition — specifically, a shift toward people who report they are unable
to work. The increase in inability to work seems evident among both men and
women, and across the age spectrum (except for youth age 18-29). These self-reports
appear to be supported by less subjective measures that show an increased prevalence
of poor health status, more severe functional limitations, and greater need for
personal assistance with ADLs and IADLs among persons with disabilities. The
increased self-reporting of inability to work among the population with disabilities
may thus be “a legitimate and accurate reflection of a real worsening in the degree
of work limitation.”36
Assuming that the share of persons with severe work disabilities remains
elevated, it could prove difficult to narrow the wide gap between the employment
rates of people with and without disabilities. This finding also raises a fundamental
question about whether all adults with disabilities — including those with very work-
limiting impairments — are likely to be able to participate in the labor force and to
succeed at finding and keeping jobs. The 106th Congress responded affirmatively to
this question when it passed the Ticket to Work and Work Incentives Improvement
Act of 1999, as its intent is increasing work activity among a portion of the adult
population with severe disabilities, namely, SSDI and SSI beneficiaries who have
medically demonstrated physical or mental impairments of long duration that impede
their participation in the labor force.
Policy analysts have examined other explanations for the comparatively poor
labor market performance of the population with disabilities during the 1990s. They
include expansion of the SSDI and SSI programs, unintended consequences
associated with enactment and implementation of the ADA, increased health care
costs, and changes in the makeup of the population (e.g., educational attainment).
The results of this research have prompted some to suggest that improvement in the
employment outcomes of persons with disabilities might be attained by, on the
demand side, subsidizing the cost of employer accommodations and by, on the supply
side, assisting already employed persons with disabilities retain their current jobs,
among other things.37
36 Ibid., p. 37.
37 Richard V. Burkhauser and David C. Stapleton, “A Review of the Evidence and Its
Implications for Policy Change,” in Stapleton and Burkhauser, The Decline in Employment
of People with Disabilities.