Unemployment Insurance: Role as Safety Net for Low-Wage Workers Is
Limited (Letter Report, 12/29/2000, GAO/GAO-01-181).

The welfare and unemployment insurance (UI) programs have been part of
the nation's social safety net since 1935. The welfare program provides
cash assistance to needy families without means of support, while UI
provides cash assistance to people temporarily unemployed. In 1996,
welfare reform put time limits on how long most people can receive cash
assistance and generally required recipients to engage in work
activities to qualify for income support. Since then, the welfare rolls
have dropped dramatically as large numbers of welfare recipients have
started working, many in low-income jobs. With this shift, the UI
program has become a more significant part of the social security net.
GAO examined the use of the UI program by low-wage and unemployed
workers. GAO found that low-wage workers are less likely to receive UI
benefits than other unemployed workers even though they are twice as
likely to be unemployed. Low-wage workers are less likely to receive UI
benefits because of (1) their tendency to quit work voluntarily, (2)
restrictive state eligibility requirements, and (3) their lack of union
memberships. Several UI reform proposals to expand the availability of
UI benefits to these workers are being discussed by the Advisory Council
on Unemployment Compensation.

--------------------------- Indexing Terms -----------------------------

 REPORTNUM:  GAO-01-181
     TITLE:  Unemployment Insurance: Role as Safety Net for Low-Wage
	     Workers Is Limited
      DATE:  12/29/2000
   SUBJECT:  Unemployment insurance
	     Social security benefits
	     State-administered programs
	     Eligibility criteria
	     Disadvantaged persons
IDENTIFIER:  Unemployment Insurance Program
	     Temporary Assistance for Needy Families Program

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Testimony.                                               **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
******************************************************************
GAO-01-181
A

Report to Congressional Requesters

December 2000 UNEMPLOYMENT INSURANCE

Role as Safety Net for Low- Wage Workers Is Limited

GAO- 01- 181

Letter 3 Appendixes Appendix I: Scope and Methodology 36

Appendix II: State Responses to GAO Questionnaire on UI Eligibility 50

Appendix III: Comments From the Department of Labor 53 Appendix IV: GAO
Contacts and Staff Acknowledgments 55

Related GAO Products 56 Tabl es Table 1: Unemployment Rates for Low- Wage
and Higher- Wage

Workers, March 1992- March 1995 14 Table 2: UI Rate of Receipt for Low- Wage
and Higher- Wage

Unemployed Workers Aged 18- 64, March 1992- March 1995 15 Table 3: UI Rate
of Receipt for Low- Wage and Higher- Wage Workers,

by Number of Weeks Worked, Combining SIPP Data for Years 1992- 95 15 Table
4: UI Rate of Receipt for Low- Wage and Higher- Wage

Full- Time and Part- Time Workers With at Least 35 Weeks of Employment Prior
to Unemployment, Combining SIPP Data for Years 1992- 95 16 Table 5: UI Rate
of Receipt by State Recipiency Rate, Combining

SIPP Data for Years 1992- 95 17 Table 6: UI Rate of Receipt for Low- Wage
and Higher- Wage

Workers Unemployed in March 1995, by Industry 19 Table 7: Comparison of
States' Employment History Requirements

for UI Benefits for Two Unemployed Workers in 2000 26 Table 8: Eligibility
for UI After Quitting Job for Compelling

Personal Circumstances 27 Table 9: Eligibility for UI Benefits for
Unemployed Workers

Seeking Part- Time Work or Limited Shifts 28 Table 10: UI Benefits in 2000
for Unemployed Worker Who Had

Worked Part- Time for 26 Weeks at $5.15 per Hour 31 Table 11: Classification
of Jobs Held by Unemployed Workers in

the 27- Month Period 40 Table 12: Data From CPS, BAM, and SIPP for Low- Wage
Workers in

the Employed Labor Force and Workers Paid UI, 1992- 99 45

Figures Figure 1: Average Recipiency Rate for the Past 5 Decades 10 Figure
2: Industry Sector of Workers Unemployed in March 1995 18

Figure 3: Average UI Rate of Receipt by Industry, March 1995 20 Figure 4:
Industry Sector Employment of Low- Wage and

Higher- Wage Workers, 1997 23 Figure 5: Union Membership of Low- Wage and
Higher- Wage

Workers, 1997 24 Figure 6: Comparison of Data Elements Available Among

Sources With Nationwide Information on UI Participation 37 Figure 7:
Formulation Demonstrating the Use of CPS and BAM

Data to Infer How the Rate of UI Receipt and the Unemployment Rate Relate
Across Wage Groups for 1996 Through 1999 47

Abbreviations

AFDC Aid to Families With Dependent Children BAM Benefit Accuracy
Measurement CPS Current Population Survey FUTA Federal Unemployment Tax Act
SIPP Survey of Income and Program Participation TANF Temporary Assistance
for Needy Families UI unemployment insurance

Lett er

December 29, 2000 The Honorable Daniel Patrick Moynihan The Honorable John
D. Rockefeller IV The Honorable Bob Graham United States Senate

The Honorable Benjamin L. Cardin The Honorable Sander M. Levin House of
Representatives

Since 1935, the welfare and unemployment insurance (UI) programs have
operated side- by- side as major parts of the nation's social safety net.
The welfare program provides cash assistance to needy families without means
of support, while UI provides cash assistance to people temporarily
unemployed. In 1996, federal legislation fundamentally changed the welfare
program, putting time limits on how long most people can receive cash
assistance and generally requiring recipients to engage in work activities
to qualify for income support. Since that time, the welfare rolls have
dropped dramatically, and large numbers of welfare recipients have started
working, many in low- wage jobs. With this radical shift, the UI program is
left as a more significant element of the social safety net, particularly
for lowincome families formerly assisted by the welfare program.

In contrast to the welfare program, which focuses on assistance to needy
families with children, UI is a social insurance program intended to
partially replace lost earnings for people with prior work experience who
become involuntarily unemployed and who are able, available, and actively
seeking work. Premiums are paid in advance by employers as a payroll tax on
wages earned by their employees. Although state law varies, this payroll tax
is applied to, at a minimum, the first $7, 000 of most employees' annual
earnings. State law specifies the factors (for example, minimum earnings or
employment period) that qualify a person to collect UI benefits.

You expressed concerns about the ability of low- wage workers to qualify for
UI benefits and asked us to examine the adequacy of the UI system as a
safety net for low- wage workers and, in particular, former welfare
recipients who lose employment through no fault of their own. In response to
your concerns, we addressed four questions: (1) what is the overall trend in
usage of the UI program among the unemployed and what are the reasons for
this trend; (2) how likely are low- wage workers to receive UI benefits,
compared with other workers; (3) what factors might reduce the

likelihood that low- wage workers receive UI benefits; and (4) are states
restructuring their policies and practices to better ensure that low- wage
workers, and former welfare recipients, are included in the safety net of
the UI program?

To conduct our study, we used a combination of methods. To determine the
long- term trends in the usage of the UI program, we analyzed data from the
Department of Labor. To compare the likelihood that low- wage workers
receive UI benefits with that of other workers, we examined data from the
Survey of Income and Program Participation (SIPP), a national database
maintained by the Bureau of the Census. For our purposes, SIPP data were
available only for the 4- year period 1992 through 1995; to extend our
analysis through the rest of the decade, we supplemented SIPP data with UI
administrative data from the Department of Labor and with data from a
national database jointly maintained by the Bureau of Labor Statistics and
the Bureau of the Census- the Current Population Survey (CPS). To determine
factors that may affect the likelihood a person will receive UI benefits, we
reviewed the available literature. We also surveyed UI program directors for
the 50 states on eligibility criteria that may affect low- wage workers in
general and former welfare recipients in particular. To obtain detailed
information on state policies and practices, we talked with officials in the
four most populous states- California, Florida, New York, and Texas- as well
as collected data on other states nationwide. A fuller description of our
methodology is included in appendix I.

Results in Brief Since 1950, the percentage of unemployed people who apply
for benefits under the UI program has gradually declined. According to the
Department

of Labor's data, about one- half of the unemployed filed claims for UI in
the 1950s, while only about one- third filed for UI in the 1990s. Various
factors have been cited as contributing to this decline, with several
factors persisting over the entire 50- year period. Because labor practices
encouraged by manufacturing industries and unions (such as providing
information about UI benefit requirements) tend to increase the likelihood
that unemployed workers will apply for UI benefits, the reduction in both
manufacturing and union membership over this time is seen as a major factor
in the decline in the use of UI. Another commonly cited reason for the
decline is tighter state eligibility criteria that prevent workers with
certain characteristics- for example, those who have worked for a relatively
short period of time- from collecting UI benefits.

In the last decade, unemployed low- wage workers 1 appeared far less likely
to receive UI benefits than other unemployed workers, even though lowwage
workers were twice as likely to be unemployed. In March 1995, for example,
only about 18 percent of unemployed low- wage workers were collecting UI
benefits, while about 40 percent of the higher- wage unemployed collected
benefits. The relative difference in UI receipt remained even among workers
who worked for similar periods during recent employment or who lived in
states where more unemployed apply for UI benefits. The prior jobs of the
low- wage unemployed were twice as likely as the jobs of other workers to be
in the retail trade and services industries- industries where workers are
much less likely to receive UI benefits than are workers from other
industries such as manufacturing, mining, or construction. Although this
data analysis was limited to the 4- year period 1992 through 1995, other
data suggest that these patterns persisted throughout the rest of the
decade.

Many factors may explain why low- wage workers receive UI benefits less
frequently than other workers. These factors include tendencies to
voluntarily quit work, for example, to look for a better paying job. Fewer
low- wage workers are in manufacturing jobs and they are less apt to be
union members- two factors that may explain some differences in UI receipt.
Moreover, compliance with some state eligibility requirements can be
particularly difficult under certain circumstances for low- wage workers-
especially former welfare recipients, who are often single mothers with
intermittent employment histories. Such circumstances include:

State earnings requirements are more difficult to meet for low- wage workers
than for higher- wage workers, even when the low- and higherwage workers
were employed for the same period of time. For example, a worker who was
laid off from a job in 2000 after 20 weeks of work for 20 hours each week at
the federal minimum wage of $5. 15 per hour would not be eligible for UI in
13 states, while another worker with the

1 Throughout this report, we refer to unemployed people as low- wage or
higher- wage workers. To classify an unemployed person as such, we first
determined whether the unemployed person had had a job within a 27- month
period before the time of unemployment. If so, we divided these people into
two groups- those who had earned $8. 00 per hour or less (low- wage) and
those who made over $8. 00 per hour (higher- wage) in the last month of
their most recent job. The $8. 00 hourly rate is approximately the amount
necessary to support a family of four at the poverty level in a full- time
job for one year; it is based on 1999dollarsand isadjustedforinflation.

same work history earning $10.00 per hour would be eligible in all but one
state. Voluntarily quitting a job for personal financial problems often

disqualifies claimants from UI benefits, but some limitations can be
especially hard for low- income single parents. For example, if a worker
currently available for work had quit his or her job because child care was
temporarily unavailable, the worker would not qualify for UI benefits in 32
states. If the same worker had quit his or her last job to care for a sick
child, 26 states would disqualify the worker from benefits. An otherwise
qualified claimant may be disqualified if the claimant is

looking only for part- time work, even if the person's job history (like
many former welfare recipients) includes only part- time employment. An
unemployed retail worker, previously in a part- time job, looking for a job
with the same 30- hour work week, would be ineligible for UI in 30 states. A
claimant's most recent earnings will not count toward UI eligibility in

most states because of the time lag allowed to process wage records. As a
result, a claimant who is otherwise eligible for benefits may need to wait 3
to 6 months, a significant delay for someone with little savings or other
financial support. Currently, only 11 states allow accelerated determination
of a person's most recent earnings if needed for the person to qualify for
benefits.

While in the 4 years since welfare reform many former welfare recipients
have joined the labor force, often in low- wage jobs, nationwide few states
have adjusted their UI programs to eliminate practices that may present
difficulties to low- wage workers, particularly these new workers. For
example, during this time only one state lowered its minimum earnings
requirements, while 29 states kept the same requirements and 19 states
raised theirs. Further, only two states added provisions to shorten the time
lag for processing wage records for UI claimants whose most recent earnings
are needed to qualify for UI benefits; according to our survey of UI
directors, only one state is likely to adopt such provisions in the near
future. Because the UI program appears to provide only limited protection
for low- wage workers, the role of UI as a safety net for all workers
warrants attention, particularly in light of the recent sweeping changes to
the national welfare policy. Although the Advisory Council on Unemployment
Compensation 2 in 1995 and, more recently, a workgroup that includes state
UI directors and the Department of Labor have made proposals that would
expand the availability of the UI program for lowwage workers, consensus has
been difficult to achieve, in part, because of concerns about increased
benefit costs and effects on state autonomy in managing the UI program.

Background In 1935, as part of the Social Security Act, Congress established
two programs- Aid to Dependent Children and the federal- state system of

unemployment insurance- to provide income support to two different groups of
unemployed people. Aid to Dependent Children added federal support to state
systems of pensions for widows with children. The UI program, on the other
hand, aimed to provide workers with partial replacement of wages lost during
temporary periods of unemployment due to economic causes. Historically, the
majority of people who file for UI benefits have been men.

2 The advisory council was established under the Emergency Unemployment
Compensation Act of 1991. It consisted of 11 members- 5 appointed by the
President and 3 each by the Senate and the House of Representatives.

As administered in subsequent years (when it became known as Aid to Families
With Dependent Children or AFDC), the welfare program evolved into an open-
ended entitlement program, providing cash assistance to people with
children, usually single parents, who earned little income. Over time, as
more women with children joined the labor force, AFDC recipients with older
children were expected to look for work. More recently, several states
experimented with stricter work requirements (the so- called “work
first” philosophy) and time limits on the receipt of aid. In 1996,
federal legislation known as the Personal Responsibility and Work
Opportunity Reconciliation Act ended AFDC, alternatively providing block
grants to the states as part of a new program called Temporary Assistance
for Needy Families (TANF). The legislation put a maximum 5- year limit on
the availability of federal cash assistance under TANF 3 and required adults
to work or participate in work- related activities after receiving
assistance for 24 months as a condition for continuing to receive benefits.
Between August 1996 and December 1999, the number of TANF families declined
by approximately 2.1 million, and many new workers entered the labor force.

Unlike the Aid to Dependent Children program, the UI program has always
operated as a social insurance program. It is administered as a federal-
state partnership. To finance the program, the states levy and collect
payroll taxes from employers. The funds collected are managed in a trust
fund administered by the federal government. In almost all industries,
federal standards require coverage on all work for employers who pay wages
of $1, 500 or more in any calendar quarter. Today UI coverage is nearly
universal, extending to almost all wage and salaried workers. 4

Employers pay the premiums for the UI program through federal and state
payroll taxes that are assessed on employers but based on employees'
earnings. Employers pay taxes on wages earned by even the lowest- paid
worker. Additionally, if a worker held jobs with two different employers
during the year, the wages from each job are taxed separately.

The federal payroll tax, established by the Federal Unemployment Tax Act
(FUTA), is currently set at 6. 2 percent of the first $7,000 of an
employee's salary. In states with UI programs that meet specified federal
guidelines,

3 In 1999, 21 states had shorter time limits on the receipt of cash
assistance. 4 Self- employed individuals and agricultural workers on small
farms are generally not covered under UI.

employers receive a 5. 4 percent credit toward their FUTA tax payment,
resulting in a net federal tax of 0. 8 percent. These federal taxes finance
the state and federal administrative costs of the UI program, as well as the
federal portion of the Extended Benefit program, advances to states with
insolvent trust funds, and other related federal costs.

The actual rate of the state tax paid by individual employers depends upon
the employer's “experience rating”- a measure related to the
amount of UI benefits collected by a firm's employees. Depending upon the
employer and the state, the state payroll tax may range from 0 to 10
percent. By federal law, state taxes are assessed against at least the first
$7,000 of an employee's salary. However, among the states, the wage base
against which state taxes are assessed varies widely, from $7,000 (in 9
states) to $27,500 in Hawaii. The wage base is less than $11,000 in 32
states, thereby requiring the same tax whether, for example, employees earn
$11,000 per year or $110,000 per year. Revenues from state UI taxes finance
the payment of regular UI benefits and the state portion of the Extended
Benefit program.

Benefit coverage under the UI program is related to an individual's work
history. Generally, state law provides that unemployed workers must fulfill
three general conditions: (1) they must have been “substantially
attached” to the labor market; (2) they must have left their prior job
involuntarily (such as by employer layoff) or have quit their job for
“good cause” only; and (3) they must be currently “able
and available” for work, and, in most states, actively seeking work.
State law provides specific requirements for claimants to meet these general
conditions.

50- Year Decline in Overall, the percentage of the total unemployed
population applying for UI

Application for UI benefits has gradually declined in the past 50 years.
Several factors

generally are cited as contributing to the decline in UI participation,
Benefits Attributed to

although the significance of each is disputed. Three major factors have
Economic and Policy

persisted over most of this period- reduction in manufacturing jobs, Factors

decline in union membership, and increasingly strict state UI eligibility
requirements.

Long- Term Decline in UI Over the past 50 years, the percentage of
unemployed filing for UI benefits

Claims has generally, but gradually, declined. The measure most commonly
used

by the Department of Labor to assess the effect of the UI program- the
standard recipiency rate 5 -shows that while about 50 percent of the
unemployed filed for UI in the 1950s, only about 35 percent of the
unemployed filed for UI in the 1990s. Although this rate has fluctuated
considerably- for example, in 1980 the rate was 44 percent, then dipped to
29 percent in 1984, but by 1991 had increased to 39 percent- it indicates a
general decline over the past 5 decades. In 1999, the recipiency rate was 37
percent. Figure 1 presents the average recipiency rate, by decade, since
1950.

Figure 1: Average Recipiency Rate for the Past 5 Decades

Percentage

60 50

49.2 43.1 41.4

40

33.9 34.9

30 20 10

0 1950- 59

1960- 69 1970- 79 1980- 89 1990- 99 Decade

Source: GAO analysis of Department of Labor data for state UI programs.

5 This rate actually measures the percentage of unemployed who apply for UI
benefits under state UI programs. It is calculated using the number of
claims filed weekly with the state programs divided by the total number of
unemployed as counted in the Bureau of the Census' CPS. By using the total
number of claims for UI benefits, the rate includes those claims that
eventually resulted in benefits as well as those claims that were denied.
For a further discussion of this rate, see app. I.

Of the past 5 decades, the last decade- 1990 through 1999- had the most
stable rate of UI claims, showing the least annual variation. Over this
decade, an average of 35 percent of the unemployed filed for UI benefits-
varying from a high of 39 percent in 1991 to a low of 31 percent in 1993,
then increasing to 37 percent in 1999. Overall, the average recipiency rate
in the 1990s was 1 percentage point higher than that of the 1980s.

Certain Economic and Although there is no agreement about the causes of the
general decline in

Policy Factors Parallel the rate of UI filing, 6 certain factors are
commonly considered significant,

Long- Term UI Decline including (1) the decrease in the number of workers
employed in

manufacturing jobs; (2) the decline of union membership in the workforce;
(3) increasingly tighter state requirements for UI eligibility; (4) federal
taxation of UI benefits beginning in 1979; (5) population shifts, starting
in the 1970s, of workers from northeastern states to southern states, where
unemployed workers are less likely to apply for UI benefits; and (6) changes
in the survey methodology of the CPS during the 1980s that increased the
number of unemployed who were counted (changing the denominator used in
calculating the recipiency rate). 7 Of these factors, the first three affect
the entire period of decline.

Over the past 50 years, the number of workers in manufacturing jobs has
declined in the United States, as has the number of workers who are union
members. Studies suggest that the steady decline in workers in manufacturing
jobs and in union membership has adversely affected the overall
participation in the UI program. 8 According to these studies, both the
manufacturing industry and unions traditionally have encouraged labor
practices that are treated favorably in UI programs. For example, union
members are more likely to be laid off than fired- a practice that makes
workers eligible for UI benefits. Manufacturing firms tend to have layoffs
of large numbers of employees who are handled as a group by UI program

6 For a general discussion of various factors involved in the decline during
the past 30 years, see Department of Labor Unemployment Insurance Occasional
Paper 99- 7, Analysis of Unemployment Insurance Recipiency Rates(
Washington, D. C.: Employment and Training

Administration, June 1999). 7 In its comments, the Department of Labor noted
that the increasing number of part- time and multiple job holders is another
possible factor in the general decline in the rate of UI filing.

8 Department of Labor, Analysis of Unemployment Insurance Recipiency Rates.

officials. Further, both manufacturing workers and union members are more
apt to be better informed about UI benefits.

In the past 5 decades, many states have tightened their UI regulations,
increasing limitations on eligibility for UI benefits and thereby decreasing
the participation in the UI program. In general, in order to demonstrate
that a person is an active member of the labor force, states have a series
of tests dealing with a claimant's recent work history, his or her reasons
for termination from the last job, and evidence that the claimant is still
available for work. For example, most states require that in order to
establish that a person worked a sufficient amount of time to qualify for UI
benefits, he or she must have earned a minimum amount of wages over a year's
time (a so- called “base period”). Over the years, many states
have increased these earnings amounts, thereby limiting who can be eligible
for UI benefits. 9

Other limitations affect program participation as well. For instance, when
the UI program was first established, people who quit their jobs for
compelling personal reasons, such as pressing family obligations like lack
of child care, were not disqualified from receiving UI benefits.
Increasingly, however, states have enacted laws that specifically limit the
generally acceptable reasons (“ good cause”) for quitting a job
to those related to work or to the employer. 10 The number of states with
such statutory restrictions grew from 16 in 1948 to 28 in 1979, and by 1995,
38 states restricted “good cause” for quitting to work- related
circumstances. Under these restrictions, states generally allow a worker to
collect UI benefits if a worker quit because of actions taken by an
employer- if, for example, an employer requires the employee to work a night
shift even though the employee had been hired specifically to work only
during daytime hours. On the other hand, most states disqualify a claimant
for UI if he or she quit a job because of a temporary lack of child care.

9 A number of studies have presented evidence on state trends to increase
minimumearnings requirements, such as Unemployment Insurance: Program's
Ability to Meet Objectives Jeopardized( GAO/ HRD- 93- 107, Sept. 28, 1993).

10 Although most states have this general restriction, states also have
created statutory exceptions to it. For example, Texas will not disqualify
someone who had to leave work to care for a seriously ill child, provided a
doctor certifies that the child is seriously ill.

Unemployed LowWage Unemployed low- wage workers were less likely to collect
UI benefits than

Workers Were other unemployed workers in the early 1990s, and the most
recent evidence

suggests that this trend continued throughout the decade. Unemployed Less
Likely to Receive

workers were more apt to receive UI benefits if they worked longer than 35
UI Than Other

weeks, worked full- time rather than part- time, or lived in a state that
Unemployed Workers

tended to have less strict eligibility criteria. However, even when low-
wage workers and other workers shared characteristics that favored UI
receipt-

in the 1990s for example, when they worked more than 35 weeks- low- wage
workers

were less likely to collect UI. In March 1995, almost two- thirds of
unemployed low- wage workers worked immediately before becoming unemployed
in jobs in the retail trade or services industries, 11 industries whose
workers were the least likely to participate in the UI program. In contrast,
only one- third of unemployed higher- wage workers held their last job in
these industries. Although SIPP data were limited to the 4- year period
between 1992 and 1995, 12 other evidence suggests that these patterns
remained throughout the entire decade.

Significantly Lower UI Rate From 1992 to 1995, low- wage workers were twice
as likely to be out of

of Receipt Among LowWage work as higher- wage workers but only half as
likely to receive UI benefits.

Unemployed Workers Table 1 compares the unemployment rates of low- wage
workers with those

of other workers in the early 1990s. During this period, low- wage workers
made up about 50 percent of the unemployed former workers, 13 even though
they were only about 30 percent of the total labor force. Table 2

11 “Services industries,” as used in this report, refers to
business services, personal and entertainment services, medical services,
educational and social services, and professional services.

12 The SIPP is administered in person to participants every 4 months over a
3- year period. The participants who are surveyed for this period are
referred to in total as a “panel.” For example, 1993 SIPP panel
participants first reported data in October 1992 and concluded in December
1995. Our analysis required that we use data covering an extended period of
time. At the time we conducted our research, the only completed SIPP panels
from the 1990s were those beginning in 1990, 1991, 1992, and 1993. The
latest data available from these panels were for December 1995 from the 1993
SIPP panel. As a result, our research with SIPP data was limited to the
period January 1990 through March 1995.

13 Our calculations of “unemployed former workers” specifically
excluded people who did not have a job in the 27- month period before the
month that they were unemployed. These people would include all new entrants
into the labor force, as well as some reentrants. Lowwage workers made up
about 40 percent of the total unemployed population.

shows the rates at which low- wage workers received UI benefits while
unemployed, as compared with the rates of higher- wage workers. 14

Table 1: Unemployment Rates for Low- Wage and Higher- Wage Workers, March
1992-March 1995

Unemployment rate a (percent) Low- wage workers Higher- wage workers Overall
b

March 1992 10.3 4. 6 6. 8 March 1993 10.2 4. 1 6. 9 March 1994 9.2 3. 9 6. 7
March 1995 7.8 3. 2 5. 7 Note: Differences between low- wage workers and
higher- wage workers are statistically significant at the .01 level. a We
calculated the unemployment rate by dividing the number of unemployed
workers by the number

of workers in the labor force. For example, the low- wage unemployment rate
was calculated by dividing the number of low- wage workers unemployed by the
total number of low- wage workers in the labor force. b The overall
unemployment rates we calculated differ from the standard unemployment rates
provided

by the Bureau of Labor Statistics. For these 4 years, standard rates were 7.
5 percent for 1992, 6.9 percent for 1993, 6.1 percent for 1994, and 5.6
percent for 1995. These rates differ because our calculations excluded
workers who were younger than 18 or older than 64, and because there were
technical differences between the database we used for our calculation
(SIPP) and that used for the standard unemployment rates (CPS).

Source: GAO analysis of the 1990, 1991, 1992, and 1993 SIPP panels.

14 The UI rate of receipt we have constructed is not comparable to the
Department of Labor's standard recipiency rate. Our rate of receipt measures
the number of people who have actually received a UI payment as a percentage
of the unemployed in the labor force. The standard recipiency rate, on the
other hand, effectively measures the percentage of unemployed who apply for
UI benefits.

Table 2: UI Rate of Receipt for Low- Wage and Higher- Wage Unemployed
Workers Aged 18- 64, March 1992-March 1995

UI rate of receipt a (percent) Low- wage workers Higher- wage workers
Overall

March 1992 30. 8 62.9 43. 0 March 1993 32. 0 52.1 34. 6 March 1994 21. 4
44.2 26. 6 March 1995 17. 8 40.0 22. 4 Note: Differences between low- wage
workers and higher- wage workers are statistically significant at the .01
level. a We calculated the UI rate of receipt by dividing the number of
unemployed workers who reported UI

as a source of income by the number of workers who were unemployed. Source:
GAO analysis of the 1990, 1991, 1992, and 1993 SIPP panels.

Among unemployed workers who had worked for similar periods of time, low-
wage workers were still less likely to receive UI benefits than higherwage
workers. As shown in table 3, nearly 35 percent of unemployed lowwage
workers who had worked at least 35 weeks during the year collected UI. In
contrast, about 62 percent of unemployed higher- wage workers who had worked
at least the same number of weeks collected UI.

Table 3: UI Rate of Receipt for Low- Wage and Higher- Wage Workers, by
Number of Weeks Worked, Combining SIPP Data for Years 1992-95

UI rate of receipt a (percent) Weeks worked prior to

Low- wage Higher- wage unemployment b workers workers Overall

35 weeks or more 34. 7 61.9 51.1 20 to 35 weeks 27.0 65. 6 41. 0 Less than
20 weeks 13.3 27. 7 17. 7 Note: Differences between low- wage workers and
higher- wage workers who worked 20 to 35 weeks and more than 35 weeks are
statistically significant at the .01 level. Differences between low- wage
workers and higher- wage workers who worked fewer than 20 weeks are
statistically significant at the .05 level. a We calculated the rate of
receipt by dividing the number of unemployed workers who reported UI as a

source of income by the number of workers who were unemployed. b Weeks
worked prior to unemployment is the sum of the number of weeks that the
person worked in

the 12- month period immediately before his or her unemployment. Source: GAO
analysis of the 1990, 1991, 1992, and 1993 SIPP panels.

Even when comparing full- time workers with substantial work histories,
differences remained. Table 4 looks at unemployed people who had worked at
least 35 weeks, grouped into those who had worked full- time and those who
had worked part- time. As can be seen, among the people who had worked full-
time for at least 35 weeks, a considerable difference continues between the
percentages of low- wage and higher- wage unemployed workers who collected
UI benefits.

Table 4: UI Rate of Receipt for Low- Wage and Higher- Wage Full- Time and
Part- Time Workers With at Least 35 Weeks of Employment Prior to
Unemployment, Combining SIPP Data for Years 1992- 95

UI rate of receipt (percent) Employment status

Low- wage workers Higher- wage workers

Full- time a 41. 0 63. 3 Par t- ti me 23. 3 53. 9 Note: Differences between
low- wage workers and higher- wage workers are statistically significant at
the .01 level. a Full- time employment is defined as 35 hours per week or
more.

Source: GAO analysis of the 1990, 1991, 1992, and 1993 SIPP panels.

Although some states had greater participation among the unemployed in their
UI programs- most of these tending to use less strict eligibility criteria
that allow a greater percentage of unemployed to collect benefits- low- wage
unemployed workers continued to be less likely to collect UI benefits than
other unemployed workers, regardless of the states in which they lived. To
group states, we used the Department of Labor standard recipiency rate as a
rough gauge of the relative rates at which the unemployed used the state UI
programs. As can be seen in table 5, even though states with high recipiency
rates were more likely to pay UI benefits, low- wage workers in those states
were still only about half as likely as higher- wage unemployed workers to
collect UI benefits.

Table 5: UI Rate of Receipt by State Recipiency Rate, Combining SIPP Data
for Years 1992-95

UI rate of receipt (percent) State recipiency rate a

Low- wage workers Higher- wage workers Overall

Low 18. 5 38.5 21. 8 High 27. 4 59.9 40. 0 Note: Differences between low-
wage workers and higher- wage workers are statistically significant at the
.01 level. a States were grouped on the basis of our analysis of the
Department of Labor standard recipiency

rates for 1992 through 1995. States with the lowest recipiency rates in all
4 years included Alabama, Arizona, Georgia, Indiana, Louisiana, New
Hampshire, New Mexico, Oklahoma, Texas, and Virginia. States with the
highest recipiency rates included Connecticut, Hawaii, Oregon, Pennsylvania,
Rhode Island, Washington, and Wisconsin.

Source: GAO analysis of 1990, 1991, 1992, and 1993 SIPP panels.

Industry Sector of Previous Overall, low- wage unemployed workers were far
more apt to have worked

Job Affected UI Rate of in retail trade and services and less apt to have
worked in manufacturing,

Receipt mining, or construction than higher- wage unemployed workers. Figure
2

shows the industry sector (based on the worker's last job) for workers who
were unemployed in March 1995. As shown, 64 percent of the low- wage
unemployed workers had been previously engaged in jobs from retail trade and
services, as opposed to 32 percent of higher- wage workers (primarily in the
services industry). On the other hand, while 49 percent of the higherwage
unemployed workers had been employed in manufacturing, construction, or
mining, only 23 percent of the low- wage workers had been employed in these
industries.

Figure 2: Industry Sector of Workers Unemployed in March 1995

Low- Wage Workers a Higher- Wage Workers

Other Other

14% 19%

Mining Services

Services and

27% 25%

Construction

7%

Manufacturing Retail Trade

16%

Mining and

7%

Construction

27%

Retail Trade Manufacturing

37% 22%

Note: Although only the 1995 data are presented here, the distributions are
similar for March 1992, 1993, and 1994. The “other” industry
category includes finance, agriculture, forestry, fishing, transportation,
utilities, wholesale trade, and public administration. a The total for low-
wage workers does not equal 100 percent because of rounding.

Source: GAO analysis of 1993 SIPP panel.

Wide variation exists among industry sectors in the rates at which
unemployed workers collected UI benefits. In general, workers formerly
associated with the retail trade or services industries were far less likely
to receive UI benefits than were workers most recently employed in
manufacturing, construction, or mining. Table 6 compares the rates among
industries for workers unemployed in March 1995. As shown, 16 percent of
former retail employees and 13 percent of former services employees
collected UI benefits, while 39 percent of unemployed manufacturing workers
and 58 percent of unemployed construction and mining workers collected
benefits. Even with these variations among sectors, differences remained in
the rates of UI receipt for unemployed low- wage workers and other workers
in individual industry sectors. Among former services workers, though, both
low- wage and higher- wage workers were far less likely to collect UI than
were higher- wage workers in the other industry sectors.

Table 6: UI Rate of Receipt for Low- Wage and Higher- Wage Workers
Unemployed in March 1995, by Industry

UI rate of receipt (percent) Industry

Low- wage workers Higher- wage workers Overall

Retail trade 12.0 42.6 16.2 Services 8.8 16.2 13.2 Manufacturing 24. 9 51. 1
39. 4 Construction and mining 27. 4 66. 3 57. 9 Note: Differences between
low- wage workers and higher- wage workers in retail trade and the mining
and construction industries are statistically significant at the .05 level.
Differences between low- wage workers and higher- wage workers in the
manufacturing industry are statistically significant at the .1 level.
Differences between low- wage workers and higher- wage workers in the
services industry are not statistically significant.

Source: GAO analysis of 1993 SIPP panel.

Compared with all other industry sectors, the retail trade and services
industries- where most unemployed low- wage workers had held their last job-
had the lowest UI rate of receipt. (See fig. 3.)

Figure 3: Average UI Rate of Receipt by Industry, March 1995

Percentage

70 60

58 53

50 40

39 40 31 33

30

30 20

16 13 10

0 Services

Retail Trade Finance

Agriculture, Forestr

y Transportationand Utilities

Manufacturing WholesaleTrade

Public andConstruction Mining Fishing, Administration

Source: GAO analysis of the 1993 SIPP panel.

Differences Likely Although the available SIPP data for our purposes
extended only to 1995,

Continued Throughout the we concluded on the basis of our analysis of other
data that the rate of UI

Decade receipt for low- wage unemployed workers most likely remained lower
than

that for other unemployed workers through the last half of the decade. This
analysis combined two sets of data that were available for the entire
decade: (1) CPS data showing the percentage of low- wage workers in the
employed labor force and (2) Department of Labor data showing the percentage
of all those collecting UI benefits who were low- wage workers. These two
percentages were stable over the entire time period. From these factors,
together with the likelihood of a higher rate of unemployment for low- wage
workers, we inferred that the UI rate of receipt of low- wage workers
remained lower than that of other workers. Between 1992 and 1995, SIPP data
showed that the unemployment rate of low- wage workers was twice that of
higher- wage workers. Our analysis of these other data showed that as long
as the unemployment rate for low- wage workers

continued to be substantially higher than that for other workers, 15 the
rate of UI receipt for low- wage unemployed workers would still have been
lower than that for other unemployed workers in the last half of the 1990s.
(See app. I for our analysis.)

From other economic factors, it appears likely that the unemployment rate of
low- wage workers remained higher than the unemployment rate (calculated for
all workers) throughout the decade (even though the unemployment rate
declined from 5. 6 percent in 1995 to 4. 2 percent in 1999) and that,
therefore, the rate of UI receipt for low- wage workers remained lower than
that for other workers. For example, low- wage workers were clustered in the
same industries in the later 1990s that they were in during the early 1990s-
about the same percentage (nearly 70 percent) of low- wage workers were
employed in services and retail industries in 1997 as in 1992. In addition,
while many welfare recipients joined the labor force and became employed
during the latter half of the 1990s, many in low- wage jobs, it appears that
they experienced higher than average unemployment rates. According to
Department of Health and Human Services data, about 30 percent of those with
jobs during the late summer 1998 were no longer employed by January 1999. 16
Unemployment rates for former welfare recipients entering the labor force in
1996 and 1997 have been estimated as 35 percent and 33 percent,
respectively. 17 Given these data, we believe that low- wage workers
continued to experience higher than average unemployment rates in the last 5
years of the decade.

15 Our calculations using CPS and Department of Labor data for the 4- year
period 1996 through 1999 indicated that, as long as low- wage workers'
unemployment rate exceeded that for other workers by at least 18 percent,
the low- wage workers' UI rate of receipt would be lower than that for other
workers throughout the rest of the decade.

16 This calculation is based on data collected from states for the purpose
of determining TANF High Performance Bonuses. States may report data from UI
wage records, surveys, or administrative records.

17 Unemployment rates are based on Displacement and Wage Effects of Welfare
Reform, Timothy J. Bartik, W. E. Upjohn Institute for Employment Research
(Jan. 1999). These rates were calculated with March CPS data for female
household heads aged 16 through 44 who had collected welfare benefits in the
previous year.

Economic and Policy Many factors may explain the relatively lower rate of UI
receipt among lowwage

Factors Influencing workers. These factors could include the possibility
that low- wage

workers are more likely to quit work to look for another (perhaps
betterpaying) General UI Decline

job or to be fired for cause than other workers. Both of these Have
Significant

circumstances would generally make claimants ineligible for UI benefits.
Impact on Low- Wage

However, certain major factors commonly cited by experts as contributing to
the general decline in use of the UI program- fewer workers in

Workers manufacturing jobs or with union membership, and tighter state
eligibility

requirements- have particular significance for low- wage workers. Low- Wage
Workers Less

As a group, low- wage workers are much less likely than other workers Likely
to Work in

either to be employed in manufacturing or to be union members. They are
Manufacturing Jobs or to Be

also less likely to be employed in other industries such as construction and
Union Members

mining that, like the manufacturing industry, tend to use layoffs to
terminate employees. Rather, they are likely to work in retail trade or
services, industries that historically have handled job separations
differently (generally, there are fewer employee layoffs) and had less union
membership than industries such as manufacturing. In 1997, about 70 percent
of low- wage workers were employed in retail trade and services, while 18
percent worked in manufacturing, mining, or construction.

Figure 4 compares the employment of low- wage workers with that of other
workers by major industry sectors in 1997. For example, 32 percent of the
low- wage workers held jobs in retail trade during 1997, but only 9.5
percent of other workers were employed in retail trade jobs. Figure 5
compares the union membership of these two groups.

Figure 4: Industry Sector Employment of Low- Wage and Higher- Wage Workers,
1997

Percentage

40

36 35

Low- Wage 35

Higher- Wage

32

30 25 20

20

15

13

10

10 9

7 7

6

5

4 3

3 4 4 4

1 2

0 Trade

Services Mining and

WholesaleTrade Finance

Public Retail Manufacturing

andConstruction Agriculture Forestry

Transportationand Utilities Administration

Note: The services industry combines the following industries: business
services, personal and entertainment services, medical services, educational
and social services, and professional services.

Source: Data are based on Jared Bernstein, Economic Policy Institute, and
Heidi Hartman, Institute for Women's Policy Research, “Defining and
Characterizing the Low- Wage Labor Market,” The Low- Wage Labor
Market, Department of Health and Human Services (Dec. 1999), which analyzed
data from the

1997 CPS Outgoing Rotations.

Figure 5: Union Membership of Low- Wage and Higher- Wage Workers, 1997

Percentage

25 20

18

15 10

6

5 0

Low- Wage Higher- Wage Source: Data are based on Jared Bernstein, Economic
Policy Institute, and Heidi Hartman, Institute for Women's Policy Research,
“Defining and Characterizing the Low- Wage Labor Market,” The
Low- Wage Labor Market, Department of Health and Human Services (Dec. 1999),
which analyzed data from the 1997 CPS Outgoing Rotations.

Certain Eligibility Criteria Certain state eligibility criteria are
particularly challenging to low- wage

and Time Lags in Processing workers, especially to those who have not held
jobs for steady periods of

Wage Records Present time, such as many former welfare recipients.
Unemployed people with

Hurdles for Low- Wage economic and financial characteristics commonly
associated with former

welfare recipients- single parents with dependent children who most often
Workers

have an intermittent work history of low- wage (and frequently part- time)
work- can be particularly vulnerable to these state requirements. These
state criteria include requirements for minimum amounts of earnings as well
as disqualification for benefits if workers leave jobs because of personal
financial circumstances. In addition, the time allotted in many states for
processing wage records may require that a claimant wait between 3 and 6
months before receiving benefits to which he or she is entitled.

Initially, to apply for UI benefits an unemployed person must have had
“substantial attachment to the labor force” in prior work. Most
states 18 use previous earnings- recorded on a quarterly basis in state wage
records- to measure whether a claimant has had sufficient employment
history. For the most part, states require that a claimant have earned a
certain minimum amount over a specified four calendar quarters (the
“base period”). The minimum amount for the base period ranges
from $130 in Hawaii to $3,400 in Florida.

As a practice, the use of earnings to measure employment history treats low-
wage workers differently from higher- paid workers, even if their
participation in the workforce is similar. For example, a worker in Florida
earning the minimum wage of $5.15 per hour must work 660 hours to qualify
for UI, while a worker earning $10. 00 per hour would need to work a little
over one- half as long to qualify for benefits. Although the current state
earnings requirements appear fairly minimal (a full- time worker earning
minimum wage for 40 hours per week would need to work 16.5 weeks to qualify
for UI in Florida), they can have a negative impact on workers with a less
stable job history.

In table 7, we compare the effect of state earnings and employment
requirements on two unemployed part- time workers who both lost their jobs
in 2000- one earned minimum wage and the other earned $10.00 per hour. The
comparison demonstrates that a part- time, low- wage worker is less likely
to qualify for UI benefits. In fact, in eight states, working 20 hours a
week for 6 months at the minimum wage would be insufficient to qualify an
unemployed worker for benefits.

18 Three states- Michigan, New Jersey, and Ohio- count the number of weeks
worked, requiring a minimum of 20 weeks of work for UI benefits. Washington
requires an unemployed person to have worked for 680 hours during the base
period to apply for UI.

Table 7: Comparison of States' Employment History Requirements for UI
Benefits for Two Unemployed Workers in 2000

Worker 1: prior job paid minimum wage ($ 5. 15 per

Worker 2: prior job paid Work hi st or y

hour) $10.00 per hour

Worked for 20 weeks, 20 Ineligible for UI in 13 states:

Ineligible for UI in 1 state: hours per week AL, CO, FL, IN, KS, ME, NC,

WA a NH, ND, OH, VA, WA, WV

Worked for 26 weeks, 20 Ineligible for UI in 8 states:

Ineligible for UI in 1 state: hours per week FL, IN, ME, NC, ND, NH, VA,

WA a WA

Worked for 40 weeks, 20 Eligible for UI in all states Eligible for UI in all
states hours per week a Washington requires claimants to have worked 680
hours to receive benefits.

Source: GAO analysis of Department of Labor summary of state UI laws.

Next, to be eligible for UI benefits in most states, a person must have
become unemployed involuntarily- that is, the person was either laid off or
quit a job for “good cause.” Generally, if a person leaves a job
for reasons other than good cause, he or she is disqualified from UI
benefits. However, much variation exists among the states about the factual
circumstances that may constitute “good cause.” Even though many
states have laws that restrict good cause to work- related circumstances,
administrative decisions and specific statutory exceptions lead to different
interpretations of “work- related circumstances.”

Certain temporary family crises- such as the sudden loss of child care or
the serious illness of a dependent child- may cause workers in marginal
financial circumstances to quit their jobs. We surveyed the UI directors of
the 50 states 19 about three hypothetical situations involving retail
workers who quit their jobs for compelling personal reasons. In all cases,
it was assumed that the workers were otherwise eligible for UI and that they
were able to work when they applied for benefits.

Table 8 shows that most states would deny benefits to those currently
available for work who had to quit their jobs because child care was

19 Individual state responses to the survey questions discussed in this
report are included in appendix II.

temporarily unavailable. However, if a worker originally hired to work a day
shift was suddenly required to work a night shift and had to quit the job
because child care was not available, only eight states would deny UI
benefits. If an employee had quit to take care of a seriously ill child,
about half of the states would deny benefits.

Table 8: Eligibility for UI After Quitting Job for Compelling Personal
Circumstances Circumstances Eligible for UI Ineligible for UI

Retail worker currently able to work originally quit 17 states 32 states

job because child care was temporarily unavailable. a

Retail worker hired to work a day shift originally quit 41 states 9 states

job because was suddenly required to work the night shift and child care was
not available during the night shift.

Retail worker currently able to work originally quit 24 states 26 states job
to care for a seriously ill child. a One state did not respond to this
example.

Source: GAO survey of state UI directors.

In general, under state laws the unemployed person must also be available
and able to work and, in most states, actively seeking work. Again, states
have different definitions as to who is currently available and seeking
work, often requiring that a claimant search for full- time work. In
addition, some states require that the claimant be available to take a job
for any shift that might be offered. Because many former welfare recipients
work parttime and may be limited in the hours they work because of lack of
child care and limited access to transportation, we surveyed the states on
their requirements related to these issues.

Table 9 shows that three- fifths of the states would not allow benefits to
be paid to an unemployed part- time worker continuing to look only for
parttime work, even though the worker is otherwise eligible for UI. However,
if a person looking for work in retail trade could not work during a night
shift because child care or transportation was not available, most states
would continue to pay UI benefits.

Table 9: Eligibility for UI Benefits for Unemployed Workers Seeking Part-
Time Work or Limited Shifts

Circumstances Eligible for UI Ineligible for UI

Unemployed retail worker previously in part- time job 20 states 30 states is
looking for job with same 30- hour work week. Unemployed retail worker is
not available to work

38 states 10 states evenings or weekends because of lack of child care. a

Unemployed retail worker is not available to work 41 states 6 states

nights because public transportation is generally not available at night. b
a Two states did not respond to this example. b Three states did not respond
to this example.

Source: GAO survey of state UI directors.

Finally, even if the unemployed worker is eligible to receive benefits, the
time it takes to process wage records may cause serious delays before the
worker can collect UI benefits. In most states, a claimant for UI must have
worked in two calendar quarters and have state wage records that show
earnings in each of the quarters. However, the time it takes to add
quarterly employee wage information to the state wage records generally
means that the complete wage records will not be available until the next
quarter after the information is received.

Two factors cause delays in processing state wage records, which are
compiled from quarterly employee wage reports. First, the wage report is not
due to most states until a month after the end of the quarter in which the
wages are earned. For example, the wage report for the last calendar quarter
of the year (ending on December 31) is due to the state January 31. Second,
after the state receives the wage report, it needs time to process it. While
many states require that employers with more than 250 employees file wage
reports on magnetic media, smaller companies often file on paper documents,
which may take 3 to 6 weeks longer to process. Therefore, although some wage
data may be available after the first month of the next quarter (February 1
in the example), all wage data may not be available until the beginning of
the next quarter (April 1).

To allow for these processing delays, most states specify that wages that
count for UI must have been earned within the first four quarters of the
last five completed quarters. These four quarters are called the
“standard base period.” In many states unemployed workers whose
only work was in the most recent 6 months may have to wait between 3 and 6
more months to have their earnings counted toward UI eligibility. For
example, if a worker starts a job in a retail store in October but gets laid
off February 1, 39 states would not apply the worker's total earnings toward
UI eligibility until after July 1. 20

Currently, only 11 states will count the worker's earnings immediately
toward UI eligibility. If a worker does not have sufficient earnings in the
standard base period, most of these states will allow what is known as an
“alternative base period” and count the earnings in the last
four completed quarters (so that the worker's January earnings would be
counted in the second calendar quarter starting in April). In these states,
if the wage records have not yet been processed, state officials most
commonly make a “wage request” of an employer to verify a
claimant's most recent earnings.

Little Change to UI Since welfare reform in 1996, the welfare rolls have
dropped and large

Since Welfare Reform numbers of people have joined the labor force, many in
low- wage jobs. Yet,

most states have made little change to their UI benefit coverage provisions
that would assist low- wage workers. Specifically, states have made few
alterations to eligibility criteria, such as minimum earnings requirements,
and other practices that in their current form may make it more difficult
for low- wage workers to qualify for UI. Recently, however, a group
representing the Department of Labor, state UI directors, and others has
offered proposals to expand benefit coverage for UI claimants that address
some of the issues related to low- wage workers.

20 In California, the worker's complete wage records might not be available
until August 1, 6 months after he or she was first unemployed. However, in
California the worker also would be eligible to receive benefits as of April
1 based solely on one quarter's earnings.

UI Benefit Coverage For the low- wage worker with an unstable job history,
little has changed in

Provisions and Processing state laws in recent years to increase the
likelihood of UI coverage. 21 In

Lags Remain Largely fact, in some states UI benefits for such workers became
less accessible.

Unchanged For example, a former welfare recipient started her first job
October 1 as a

retail clerk paid at $5.15 per hour. After working 26 weeks for 20 hours
each week, she was laid off because of slow sales. During that period, she
earned $2,678 and worked 520 hours. In 1996, she would have been ineligible
for benefits in five states- Indiana, Maine, New Hampshire, North Dakota,
and Virginia- because these states require a claimant to have earned more
than this worker's total wages, and also in Washington because she had not
worked a sufficient number of hours. In 2000, she would be ineligible in
eight states- those listed above plus Florida and North Carolina- because
these states raised their minimum earnings requirements. During the period
1996 through 2000, 19 states increased the total earnings required for UI
eligibility, 1 state lowered its requirement, and the remaining 29 kept the
same minimum earnings level.

If the worker in the previous example resided in a state where she was
eligible for UI, the benefits available to her would most likely be about
the same in 2000 as in 1996. In 12 states, she would receive additional
benefits if she had dependents. The states vary as to both weekly benefit
amounts and how long a claimant may receive the weekly amount. Table 10
illustrates the benefit coverage of UI if this worker filed in 2000 in the
four most populous states.

21 In its comments, the Department of Labor pointed out that although the
states have not made many changes, the national minimum wage increased twice
between 1996 and 2000, rising from $4. 25 to $4. 75 on October 1, 1996, and
increasing to $5. 15 on September 1, 1997. Labor notes that these increases
made it more likely that low- wage workers met state eligibility
requirements.

Table 10: UI Benefits in 2000 for Unemployed Worker Who Had Worked Part-
Time for 26 Weeks at $5.15 per Hour

Weekly Duration of

Tot al benef i t s State amount benefits available

California $58 23 weeks $1, 339 Florida 0 0 0 New York $54 26 weeks $1,393
Texas $54 13. 5 weeks $723 Note: By comparison, the monthly cash benefit
under the TANF welfare program available to a family of three in these four
states in January 2000 amounted to $626 in California, $303 in Florida, $577
in New York, and $201 in Texas.

Source: GAO analysis of Department of Labor summary of state UI laws for
2000.

Since 1996, there also has been very little movement among the states to
adjust for the time lag in reporting wages if it affects when an unemployed
low- wage worker can be eligible for benefits. Thus, even if the unemployed
worker in our example was eligible to receive benefits, the time it would
take to have her wages count would most likely cause delay before she could
apply for, and collect, UI benefits. In 1996, nine states had provisions 22
to allow recent wages to count, even if the wages were earned outside the
normal base period, if a claimant needed the earnings to qualify for UI. In
2000, two more states (North Carolina and Wisconsin) had similar provisions.
23 Among the remaining states, however, our survey of UI state directors
indicated that it is unlikely much change will occur in the near future. Of
the 39 states without provisions to count recent earnings, only one state
director said that his state (Alaska) was likely to adopt such a provision,
and state directors from 29 states said that their states were either very
unlikely or unlikely to adopt this change.

22 Seven states (Maine, Massachusetts, New Jersey, Ohio, Rhode Island,
Vermont, and Washington) had an alternative base period, allowing a claimant
to use wages from the last four completed calendar quarters. Two states (New
York and Michigan) used the preceding 52 weeks as a standard base period and
had no lag period.

23 New York also altered its base period. As of April 1, 1999, its standard
base period changed from the preceding 52 weeks to the first 4 of the last 5
completed quarters, with an alternative base period of the last four
completed quarters ending with the week preceding the filing of a valid
original claim.

Proposals for UI Reform In the past 5 years, the Advisory Council on
Unemployment Compensation and a stakeholder workgroup that includes state UI
directors, union

representatives, business representatives, and Department of Labor officials
have made proposals that would expand the availability of the UI program for
low- wage workers, among other reforms. According to Labor, the changing U.
S. economy and its labor force have led to the current movement for reform.
The UI program was designed over 60 years ago and worked well for a certain
type of worker within the U. S. economy at that time. Since then, the U. S.
economy and the composition of its labor force have changed, while the UI
program has been slow to adapt to these changes. Labor noted that this has
resulted in a larger portion of the labor force more closely resembling a
category of worker that UI was not designed to assist. More recently, the
reform of the welfare system has further increased the number of workers in
this category.

In 1995, the advisory council made a series of proposals regarding lowwage
workers as part of a larger set of recommendations about the needs of
today's labor market. Subsequently, the Department of Labor organized a
dialogue with state, employer, and union representatives to continue the
debate on possible UI reform. As a result of this dialogue, a stakeholder
workgroup of federal, state, and private sector officials recently proposed
reforms for the UI program. Reform proposals applicable to low- wage workers
from these two groups include the following:

Shorten the lag time in qualifying earnings for UI eligibility. The advisory
council recommended that all states use a “moveable” base period
to consider earnings necessary to qualify a claimant for benefits. Under
this proposal, the minimum earnings requirement could be met by earnings
from the last four completed quarters, rather than from the first four of
the last five completed quarters. Although initially the stakeholder
workgroup considered a proposal to provide incentive funding to the states
for “alternative” base periods similar to the advisory council's
moveable base period, ultimately it suggested that states try to use
technology advances to process the UI reports faster and, where at all
possible, to use the latest wage earnings available for all claimants. Set
minimum standards for UI earnings requirements. The advisory

council recommended that all states set their laws so that required base
period earnings do not exceed 800 times the state's minimum wage. In its
dialogue, the Department of Labor asked for comments on a proposal that
would set the minimum earnings requirements to 400 times the minimum wage
(this figure was selected so that someone who had

worked for 20 weeks for 20 hours at minimum wage would be eligible for
benefits in every state). However, the final proposals from the stakeholder
workgroup did not include any recommendation on this issue. Do not
disqualify claimants seeking part- time work. Both the advisory

council and the stakeholder workgroup proposed that states should not reject
claimants simply because they are looking for part- time, rather than full-
time, work. Do not disqualify claimants who quit a job to care for a
dependent.

Although neither group ultimately recommended this proposal, the Department
of Labor originally offered it for comment. The proposal would have provided
financial incentives to states to pay UI benefits to claimants who had to
quit their jobs to care temporarily for a child or other family member.

State objections to these proposals focus on the expansion of benefits, and
the states argue that (1) the costs of the proposals are burdensome and (2)
the proposals violate the traditional roles of the federal and state
governments in the operation of the UI program. Regarding the first issue,
the states point to, for example, the extra costs of obtaining the most
recent earnings records for UI claimants. In response to proposed state
legislation, Texas estimated the extra administrative costs at $153,000
annually for making special requests to employers for recent wage
information. The proposal from the stakeholder workgroup would eliminate the
requirement that states make these special requests, instead calling for
federal funding of improved technology to accelerate state processing of UI
wage records. However, the largest cost cited by states relates to the
increased number of claimants receiving UI benefits. If alternative base
provisions were implemented, Texas estimated that the annual costs to the
unemployment insurance trust fund would be $24 million per year in benefits
paid to potential claimants; California officials estimated their costs at
$33 million per year.

From the standpoint of the states, the second objection- changes to the
traditional roles of the state and federal governments- raises more
difficult problems. While the federal government has imposed some specific
requirements, these requirements are viewed as minor conditions only; for
example, UI claimants cannot be denied benefits if they refuse work as a
union strikebreaker. In contrast, the proposals discussed here- for example,
the earnings requirements or allowable reasons for quitting a job- pertain
to issues that state officials consider integral to the operation

of the state's UI program that, until now, have been generally under the
control of the state.

Conclusion Despite interest in ensuring that the UI program is meeting the
needs of low- wage workers, little action has been taken at the state or
federal levels

to expand UI availability to this group. In part, this reflects the
difficulty of addressing the cost implications of expanded eligibility and
balancing states' autonomy in operating their UI programs. Yet, as a safety
net, the UI program continues to offer only minimal protection for low- wage
workers. Even though employers in many states pay the same UI payroll taxes
for employees earning minimum wage as they pay for employees earning far
more than that amount, low- wage workers are much less likely than
higherwage workers to be included in the UI safety net. In the event of an
economic downturn, many low- wage workers may find that, unlike higherwage
workers, they will be unable to qualify for UI benefits. While the situation
deserves attention on its own merits, the sweeping changes in national
welfare policy heighten its importance. A UI program that supports all
workers who lose their jobs through no fault of their own during times of
economic hardship can play an important role in helping many former welfare
recipients maintain their places in the labor force and out of the welfare
system.

Agency Comments In its review of a draft of this report, the Department of
Labor generally agreed with our findings and conclusion. It made three major
comments:

(1) that the changing U. S. economy and its labor force have led to the
current movement for UI reform; (2) that nonmonetary eligibility criteria
such as voluntarily quitting a job may explain some of the differences
between the UI rate of receipt for low- wage and other workers; and (3) that
the increases in the national minimum wage between 1996 and 2000 may have
made some unemployed low- wage workers eligible for UI. We concur with these
comments and have modified our report as appropriate. Labor also made
technical comments, which we have included in our report where appropriate.
(Labor's comments appear in app. III.)

As agreed with your offices, unless you publicly announce its contents
earlier, we plan no further distribution of this report until 30 days after
its issue date. At that time, we will send copies to the Honorable Alexis M.
Herman, Secretary of Labor; the Honorable Donna E. Shalala, Secretary of

Health and Human Services; appropriate congressional committees; and other
interested parties. We will also make copies available upon request.

Please call me at (202) 512- 7215 or Gale Harris at (202) 512- 7235 if you
or your staffs have any questions about this report. Other GAO contacts and
staff acknowledgments are listed in appendix IV.

Sigurd R. Nilsen Director, Education, Workforce, and

Income Security Issues

Appendi Appendi xes xI

Scope and Methodology We used a variety of data sources to examine the role
of unemployment insurance (UI) as part of the safety net for low- wage
workers. To show the general trends of UI participation among all
unemployed, we summarized data compiled by the Department of Labor. To
measure the use of UI by low- wage workers as opposed to other workers, we
used data from the Survey of Income and Program Participation (SIPP), a
survey conducted by the Bureau of the Census. To determine the specific
eligibility criteria used currently in state UI programs, we surveyed the
directors of these programs. Finally, to assess whether states have changed
their policies and practices to better ensure that low- wage workers are
included in the UI safety net, we reviewed data from the Department of Labor
as well as data from a national survey of UI directors, and we visited the
four most populous states to talk with state officials about their UI
system. We performed our work between January 2000 and September 2000 in
accordance with generally accepted government auditing standards.

Estimating the To compare low- wage workers' experience with UI with that of
other

Unemployment Rate workers, we estimated the unemployment rates and the UI
rates of receipt

for the two groups of workers. To do this, we needed information on (1) the
and UI Rate of Receipt

employment status of individuals; (2) specific characteristics of the of
Low- Wage Workers

employed population; (3) specific characteristics of the unemployed
population; and (4) detailed information on unemployed people who collected
UI. We talked with experts at the Department of Labor and the Bureau of the
Census and reviewed academic research and other related literature to
determine what data sources could be used for our study.

We considered four data sources with information on the use of the UI
program nationwide- SIPP, the Current Population Survey (CPS), the Benefit
Accuracy Measurement program (BAM), and general Department of Labor UI
administrative data. SIPP is a longitudinal survey that collects information
on labor force participation and income sources over a 3- year period. CPS,
a national survey conducted by the Bureau of the Census for the Bureau of
Labor Statistics, is a longitudinal survey that collects data on employment
status and other demographic characteristics over a 1- year period. BAM is a
Department of Labor program that collects information in order to evaluate
the accuracy of state UI payments, and it includes specific data on
demographic characteristics of people who collect UI benefits. Labor also
maintains other administrative databases that collect information related to
unemployment and the UI program. Figure 6 compares various data elements
that are available among these four sources.

Figure 6: Comparison of Data Elements Available Among Sources With
Nationwide Information on UI Participation

Note: Shaded areas indicate data were not available. Source: GAO analysis of
SIPP, CPS, BAM, and other Department of Labor UI data.

From our review, we determined that SIPP was the only data source that would
allow us to estimate what portions of the unemployed population were low-
wage and higher- wage and the extent to which each group received UI. SIPP
is a survey administered in person to participants every 4 months over a 3-
year period. During the 3- year period, the same set of questions is asked
of the same individuals, allowing for analysis of an individual's labor
force experience over the entire time. The respondents who are surveyed for
this period are referred to in total as a “panel.” For example,
participants in the panel included in the 1993 SIPP first reported

data beginning in October 1992, and they continued to report data at 4month
intervals through December 1995.

Our data analysis required that we use SIPP data that covered an entire 3-
year period. At the time we conducted our research, the only completed SIPP
panels with data from the 1990s were those started in fiscal years 1990,
1991, 1992, and 1993. The latest data available from these panels were for
December 1995 from the 1993 SIPP panel. As a result, our research using SIPP
data was limited to the period January 1990 through March 1995. 1

General Methodology for To estimate the unemployment rates and the UI rates
of receipt for lowwage

SIPP Analysis and higher- wage workers using the 1990 through 1993 SIPP
panels,

we took the following steps:

Step 1: We created a sample from each SIPP panel of 18- to 64- yearolds who
were not self- employed.

We limited our sample in each SIPP panel to those between the ages of 18 and
64. We also excluded those who were self- employed and for whom there were
incomplete data during the 3- year period. Our sample included data on the
wages and salaries of respondents as well as the number of hours and weeks
worked. 2

Step 2: We used March of the last year of each SIPP panel to determine
employment status.

Because of the design of our analysis, we chose to focus on the employment
status in one month, March, from the last year of each panel. 3 Because
seasonal employment can greatly affect employment status at certain times of
the year (such as summer and winter), we examined data

1 Although the SIPP database contained data through December 1995, because
of the design of our analysis (described more fully later in this app.), we
used data through March 1995. 2 Our final unweighted samples were 17, 088
cases for 1992, 12, 266 cases for 1993, 15, 953 cases for 1994, and 15,448
cases for 1995. 3 Therefore, from the 1990 SIPP panel, we focused on March
1992; from the 1991 SIPP panel, we focused on March 1993; from the 1992 SIPP
panel, we used data from March 1994; and from the 1993 SIPP panel, we used
data from March 1995.

from March, a month less likely to be affected by seasonal employment. SIPP
records data on a monthly basis. Since it is possible for an individual to
be both employed and unemployed in the same month, to address this issue, we
consulted with officials at the Bureau of the Census. 4 We considered a
person as employed if he or she had a job for the entire month or, if the
person missed work during this period, it was not because he or she was laid
off and he or she spent no time looking for a job. 5 We considered a person
as unemployed if he or she was out of work for the entire month or, if the
person did work for part of the month, he or she spent the rest of the time
laid off or looking for another job. 6 We did not include those who were out
of the labor force. 7

Step 3: We “looked back” 27 months for the most recent job.

To determine the wage level of an unemployed person, we identified the most
recent job held by that person. To do this, we reviewed the work history to
determine whether the person had had a job during the period covered by the
SIPP panel. Starting in March of the last panel year (the month we used to
determine unemployment), we looked back on a monthby- month basis to
determine the most recent month in which that person was identified as
employed. By using March of the last panel year we could, where necessary,
look back for a period of 27 months to identify prior employment. In some
cases, a respondent did not have a job during the entire 27- month period.
If a respondent had not held a job at all during this time, we excluded the
person from our sample.

4 See Paul Ryscavage, The Survey of Income and Program Participation:
Measuring Spells of Unemployment and Their Outcomes, No. 84 (Washington, D.
C.: Bureau of the Census, Dec. 28, 1988). 5 We specifically defined employed
on the basis of the following SIPP categories: (1) with a job entire month,
worked all weeks; (2) with a job entire month, missed 1 or more weeks, no
time on layoff; (3) with a job 1 or more weeks, no time spent looking for
work and no time on layoff.

6 We specifically defined unemployed on the basis of the following SIPP
categories: (1) no job during the month, spent entire month looking for work
or on layoff; (2) no job during the month, spent 1 or more weeks looking for
work or on layoff; (3) with a job 1 or more weeks, spent 1 or more weeks
looking for work or on layoff; and (4) with a job entire month, missed 1 or
more weeks of work, spent time on layoff.

7 We define labor force as either (1) employed people or (2) unemployed
people who are seeking employment.

Step 4: We divided our sample into low- wage and higher- wage workers.

We divided both the employed and unemployed populations into either low-
wage or higher- wage workers on the basis of data from the current job if
the person was employed or from the person's most recent job if unemployed.
We defined low- wage as earning $8 per hour or less, based on 1999 dollars.
For our analysis, we adjusted the rate for inflation. We determined the $8
level by dividing the annual income for a family of four at the federal
poverty level by 2, 080 hours (full- year, full- time employment). 8 We
determined each person's hourly earnings in one of two ways: (1) if the data
included an hourly wage, we used the reported hourly wage in the most recent
month that the person was employed or (2) if there was no reported hourly
wage, we constructed an hourly wage using data from the most recent month
that the person was employed (reported monthly salary divided by the number
of weeks worked in the month multiplied by the number of hours usually
worked during the week). In some cases, a respondent who had had a job could
not be classified as either low- wage or higher- wage because of missing
wage or salary data. Table 11 shows what percentage of our SIPP sample had
missing wage or salary data.

Table 11: Classification of Jobs Held by Unemployed Workers in the 27- Month
Period

Numbers in percentages

Year Low- wage job Higher- wage job Missing wage or salary data

1992 44. 5 51.4 4. 1 1993 49. 6 47.9 2. 6 1994 50. 5 44.5 5. 0 1995 49. 6
44.4 6. 0 Source: GAO analysis of SIPP data.

8 The $8 per hour threshold for defining “low- wage” has been
used in other studies. See, for example, Economic Policy Institute, The
State of Working America 1998- 1999( Ithaca, N. Y.: Cornell University
Press, 1999) and Jared Bernstein and Heidi Hartman, “Defining and
Characterizing the Low- Wage Labor Market,” The Low- Wage Labor
Market( Washington D. C.: Department of Health and Human Services, Dec.
1999).

Step 5: We determined whether the unemployed worker reported UI asa source
of income.

We then identified whether each respondent had received UI benefits while
out of work. If an unemployed former worker reported UI as a source of
income in March of the last panel year, we classified the person as
receiving UI. Otherwise, we classified the person as not receiving UI.

Step 6: We identified additional characteristics of the unemployed
population.

We further analyzed the unemployed population by identifying the industries
in which they had worked, whether they had worked full- time, and how long
they had worked before becoming unemployed. To identify the kinds of
industries low- wage and higher- wage unemployed workers had worked, we used
the industry code of the most recent job. Our work presents data for nine
industry groups: (1) retail trade; (2) manufacturing; (3) finance; (4)
mining and construction; (5) agriculture, fishing, and forestry; (6)
wholesale trade; (7) transportation and utilities; (8) public
administration; and (9) services. We developed these groups by combining
detailed CPS industry codes. For example, the services sector combines five
types of service industries: business services, personal and entertainment
services, medical services, education and social services, and professional
services.

Next, we identified whether the unemployed former worker had been employed
full- time or part- time. We defined full- time employment as working 35 or
more hours each week and part- time employment as working fewer than 35
hours per week.

To determine the number of weeks worked by the person in the year
immediately before he or she became unemployed, we created a subsample from
our original SIPP sample. The subsample included only those who were
unemployed in March of the last panel year and who had had a job during the
15 months prior to unemployment (as compared with the 27- month look- back
period for our original sample). After identifying the most recent job in
the 15- month period, we counted how many weeks each person had worked
during the 12 months before becoming unemployed. We classified the
unemployed low- wage and higher- wage workers into three categories: (1)
those who had worked more than 35 weeks during the year, (2) those who had
worked between 20 and 35 weeks during the year,

and (3) those who had worked fewer than 20 weeks during the year. To
increase the size of this subsample, we combined all four SIPP panels.

Calculating the We calculated the overall unemployment rate by dividing the
number of

Unemployment Rate unemployed by the total number of people in the labor
force, as follows:

The unemployment rates for low- wage and higher- wage workers were
calculated on the basis of their representation in the labor force. For
example, the unemployment rate for low- wage workers was calculated by
dividing the number of unemployed low- wage workers by the number of low-
wage workers in the labor force.

In calculating an unemployment rate for higher- wage workers, we included
the cases with missing wage or salary data in such a way as to calculate the
most conservative estimates, that is, estimates that would minimize any
differences between the low- wage and higher- wage unemployment rates. To
calculate this rate, we assumed that all those with missing wage data were
actually higher- wage unemployed workers. With that assumption, the group
with missing wage data was added to the higher- wage unemployed group when
we calculated the higher- wage unemployment rate.

The unemployment rate we calculated using the SIPP data differs for a
variety of reasons from the standard unemployment rate published by the

Bureau of Labor Statistics from CPS data. These two rates differ in part
because each rate measures different populations. Specifically, our rate
includes only those in the labor force who are between the ages of 18 and
64, whereas the standard rate includes all those who are 16 years old and
older. Also, our analysis excludes those who are self- employed, while the
standard rate includes self- employed workers. Another key contrast results
from technical differences between the two databases used to calculate these
rates. SIPP, for example, records data for each month, but the CPS records
data for a 1- week period. Therefore, while employment status in SIPP
measures whether or not a person was employed during the entire month, the
CPS measures whether the person was employed during the week that included
the 12th of the month.

Calculating the UI Rate of We calculated the overall UI rate of receipt by
dividing the number of

Receipt unemployed workers who had collected UI benefits during March by the

total number of unemployed workers in that month. In calculating the UI
rates of receipt, we included the respondents with missing wage data in such
a way as to present the most conservative estimates; that is, we allocated
the missing wage data so that our estimates would minimize the difference
between the receipt rates for the two groups. Therefore, to calculate the UI
rate of receipt for low- wage workers, we assumed that all workers with
missing wage data who were paid UI benefits were low- wage unemployed
workers. Conversely, to calculate the higher- wage UI rate of receipt, we
assumed that all those with missing wage data who were not paid UI benefits
were higher- wage unemployed workers.

The UI rate of receipt we constructed using SIPP data is not comparable to
the Department of Labor's standard UI recipiency rate. Our UI rate of
receipt measures the number of people who have actually received a UI check
as a percentage of the unemployed in the labor force. Labor's standard
recipiency rate, on the other hand, measures the number of people who file a
claim for UI as a percentage of the unemployed in the labor force. 9 By
measuring the number of people who file a claim for benefits, Labor's rate
includes those who eventually receive UI as well as those who do not.

Because the UI program differs from state to state, we analyzed whether
there were differences in the UI rates of receipt for the low- wage and
higher- wage unemployed that lived in different states. For purposes of
comparison, we created two groups of states- those with high standard
recipiency rates and those with low standard recipiency rates based on
Labor's standard recipiency rates from 1992 through 1995. States that
consistently were in the 15 states with the highest standard UI recipiency
rates were in one group and states that were consistently the lowest 15 were
in the other. To increase the size of the sample for each of the two groups
of states, we combined all four SIPP panels.

For each unemployment rate and UI rate of receipt presented in the analysis,
we tested whether the differences between low- wage and higherwage workers
were statistically significant. To do this, we compared the sampling errors
for the two estimates and, if the sampling errors for the low- wage and
higher- wage workers did not overlap, we concluded that each difference is
statistically significant. Unless otherwise noted, statistical significance
was tested at the .01 level, thereby allowing us to conclude that there is
only a 1 percent chance that the difference between the rates is due to
sampling error.

Data Analysis for the Rest of Although SIPP data are unavailable for the
latter half of the 1990s, other

the 1990s relevant data are available from the CPS and from the Department
of

Labor's BAM database. From the CPS data, it is possible to calculate the
proportion of low- wage (or higher- wage) workers in the employed labor
force. From the BAM data we can determine the proportion of low- wage (or
higher- wage) workers who received UI, compared with the total

9 Labor's overall standard recipiency rate was 34 percent in 1992, 31
percent in 1993, 33 percent in 1994, and 35 percent in 1995.

number of unemployed workers receiving UI. Table 12 compares these data with
our calculations, using the SIPP data for the years 1992 through 1995.

Table 12: Data From CPS, BAM, and SIPP for Low- Wage Workers in the Employed
Labor Force and Workers Paid UI, 1992- 99

Low- wage workers as a Low- wage workers paid UI as a percentage of all
employed

percentage of all unemployed workers

workers paid UI CPS a

SIPP b BAM c

SIPP b Year

(A) (B) (C) (D)

1992 28. 4 24. 9 27. 8 27. 4 1993 30. 0 26. 1 25. 5 37. 1 1994 30. 7 28. 6
29. 3 31. 8 1995 30. 6 27. 7 27. 3 27. 9 1996 30. 3 N/ A 30.3 N/ A 1997 28.
6 N/ A 30.6 N/ A 1998 27. 4 N/ A 30.8 N/ A 1999 26. 8 N/ A 27.1 N/ A Note:
N/ A = not available. a Calculation from the Economic Policy Institute.

b As calculated by GAO. c As calculated by the Department of Labor.

As shown in column A of table 12, the CPS percentage of low- wage workers in
the employed workforce was roughly 30 percent for the entire period. In
column C, the BAM percentage of those paid UI who were low- wage workers was
also roughly 30 percent during the same time. Although the table refers only
to low- wage workers, the percentages for higher- wage workers can be
computed as the inverse of the low- wage percentage (that is, 100 minus the
low- wage percentage). Thus, the CPS percentage for higher- wage workers is
roughly 70 percent, and this percentage is about the same as the BAM
percentage for higher- wage workers- also about 70 percent.

Figure 7 demonstrates that the data from the CPS and BAM can be used to
infer how the rate of UI receipt and the unemployment rate relate across
wage groups between 1996 and 1999. It can be shown that, given the observed
pattern of data from CPS and BAM, it is unlikely that the rate of

UI receipt for low- wage workers would exceed that of higher- wage workers
during this period. Essentially, for this to have occurred, the unemployment
rates across the wage groups would have to have converged dramatically
during this period- in sharp contrast to the experience of earlier years.
Whereas between 1992 and 1995 the unemployment rate for low- wage workers
exceeded that of higher- wage workers by well over 100 percent, as long as
the unemployment rate for low- wage workers exceeded that of higher- wage
workers by as little as 18 percent throughout the rest of the decade, the
rate of receipt for low- wage workers remained less than that of higher-
wage workers for each year between 1996 through 1999.

In figure 7, RR (lw) equals the UI rate of receipt for low- wage workers and
UR (lw) equals the unemployment rate of low- wage workers. For higherwage
workers, the UI rate of receipt is RR (hw) and the unemployment rate is UR
(hw).

Figure 7: Formulation Demonstrating the Use of CPS and BAM Data to Infer How
the Rate of UI Receipt and the Unemployment Rate Relate Across Wage Groups
for 1996 Through 1999

Considering the highest value W had taken from 1996 through 1999 was 1.18 in
1998, RR( lw) is less than RR (hw) throughout the decade as long as UR (lw)
is greater than UR( hw) by at least 18 percent.

Eligibility Criteria for To ascertain how individual states would treat
specific circumstances that

State UI Programs might be particularly applicable to low- wage workers, and
to former

welfare recipients, we sent a questionnaire to state UI directors. For the
most part, the questionnaire presented hypothetical situations related to
unemployed workers, and asked the UI directors to determine whether the
unemployed worker could qualify for UI benefits in their state program. We
received responses from all 50 states.

Status of State UI To assess whether state UI programs had changed in the 4-
year period

Programs since welfare reforms, we reviewed the Department of Labor's
compilation

of legislative changes in state UI laws from all 50 states. To examine the
current status of the state UI programs, we included questions about the
operation of UI programs in our survey of the state UI directors. We also
visited four states- California, Texas, New York, and Florida- to talk with
officials about the state UI program. We chose these states not only because
they were the most populous states but also because they presented contrasts
in how they manage their UI programs.

State Responses to GAO Questionnaire on UI

Appendi xII

Eligibility In June 2000, we sent a questionnaire to the UI directors for
the 50 states. In this appendix, we present the questions that we discuss in
our report and data on how each state responded.

Question 2: Does your state have an alternative base period?

Alternative base period No alternative base period

11 states: MA, ME, MI, NC, NJ, NY, OH, RI, VT, WA, WI 39 states: AK, AL AR,
AZ, CA, CO, CT, DE, FL, GA, HI, IA, ID, IL, IN, KS, KY, LA, MD, MN, MO, MS,
MT, ND, NE, NH, NM, NV, OK, OR, PA, SC, SD, TN, TX, UT, VA, WV, WY

Question 4: (If the state had no alternative base period) In your opinion,
how likely or unlikely is it that your state will adopt an alternative base
period in the near future?

Very unlikely Unlikely As likely as unlikely Likely Don't know

9 states: AZ, DE, KY, LA, 20 states: AL, AR, CO,

7 states: CA, CT, IA, IN, 1 state: AK 2 states: ND, SC MD, MO, NE, NH, UT
FL, GA, HI, ID, IL, KS,

MN, NM, VA MS, MT, NV, OK, OR, PA, SD, TN, TX, WV, WY

Questions 5, 6, and 7 dealt with personal circumstances for quitting a job.
For each question, the respondent was asked to assume that (a) the worker
has quit a job as a stockroom clerk with a retail chain store; (b) the
worker's reasons for leaving the job are compelling and not the fault of the
worker; (c) the worker is otherwise eligible to receive UI benefits.

Question 5: The worker quits because child care suddenly becomes unavailable
and the employer cannot reschedule the worker's hours. Subsequently, child
care becomes available. The worker files for UI.

Eligible after waiting for a Immediately eligible fixed period Ineligible No
response

13 states: AK, AR, AZ, CA, HI, 4 states: MD, NE, WA, WY 32 states: AL, CO,
CT, DE, FL,

1 state: WV IA, KS, MA, NY, OR, PA, RI, VA GA, ID, IL, IN, KY, LA, ME, MI,

MN, MO, MS, MT, NC, ND, NH, NJ, NM, NV, OH, OK, SC, SD, TN, TX, UT, VT, WI

Question 6: The worker was originally hired to work during the day shift.
However, the employer requires that the worker change work hours to a night
shift, and the worker quits because child care is unavailable during the
night shift.

Eligible Ineligible

41 states: AK, AL, AR, AZ, CA, CT, DE, FL, HI, IA, IL, IN, KS, KY, 9 states:
CO, GA, ID, ME, MI, NV, OH, OR, VA LA, MA, MD, MN, MO, MS, MT, NC, ND, NE,
NH, NJ, NM, NY, OK, PA, RI, SC, SD, TN, TX, UT, VT, WA, WI, WV, WY

Question 7: The worker quits to care for a sick child (physician certifies
need). Subsequently, the worker is available to resume work. The worker
files for UI.

Immediately eligible Eligible after waiting a fixed period Ineligible

22 states: AK, AR, AZ, CA, CT, HI, IA, IL, 2 states: NC, NE 26 states: AL,
CO, DE, FL, GA, ID, IN, KY, KS, MA, MD, ME, NY, OK, OR, PA, RI, TX,

LA, MI, MN, MO, MS, MT, ND, NH, NJ, NM, UT, VA, WA, WI

NV, OH, SC, SD, TN, VT, WV, WY

Questions 8, 9, and 10 dealt with the requirement that a worker be
“able and available” for work. For each question, it is assumed
that the worker's previous job was a stockroom clerk in a retail chain
store, that the worker was laid off this job, and that the worker is
otherwise eligible to receive UI.

Question 8: In her prior job, the worker was employed part- time, for 30
hours a week. When she applies for UI and is asked whether she is available
for work, she indicates that she is looking for work with the same hours as
those of her previous job and is not able to work more hours than
previously.

Eligible Ineligible

20states: AR, CO, DE, FL, HI, IA, KS, LA, MA, MN, ND, NE, NV, 30states: AK,
AL, AZ, CA, CT, GA, ID, IL, IN, KY, MD, ME, MI, MO, NY, OK, PA, SD, TN, VT,
WY MS, MT, NC, NH, NJ, NM, OH, OR, RI, SC, TX, UT, VA, WA, WI,

WV

Question 9: The worker is available to work full- time during the weekdays.
However, she cannot work evenings or weekends because of the lack of
affordable child care at that time. As a result, the worker is unable to
take jobs that require that she work evenings or weekends.

Eligible Ineligible No response

38 states: AK, AR, AZ, CA, CT, DE, FL, GA, 10 states: AL, CO, ID, KY, ME,
MI, MT, OH,

2 states: VA, WV HI, IA, IL, IN, KS, LA, MA, MD, MN, MO, MS,

OR, TX NC, ND, NE, NH, NJ, NM, NV, NY, OK, PA, RI, SC, SD, TN, UT, VT, WA,
WI, WY

Question 10: The worker is available to work full- time but always relies on
public transportation to get to work. The worker is therefore unable to take
jobs that require work during a night shift because public transportation is
generally not available during night hours.

Eligible Ineligible No response

41 states: AK, AR, AZ, CA, CT, DE, FL, GA, 6 states: AL, CO, ID, ME, OH, TX
3 states: VA, WV, WY

HI, IA, IL, IN, KS, KY, LA, MA, MD, MI, MN, MO, MS, MT, NC, ND, NE, NH, NJ,
NM, NV, NY, OK, OR, PA, RI, SC, SD, TN, UT, VT, WA, WI

Appendi xI II Comments From the Department of Labor

Appendi xI V

GAO Contacts and Staff Acknowledgments GAO Contacts Gale C. Harris, (202)
512- 7235 Nancy M. Peters, (202) 512- 9065 Staff Acknowledgments In
addition, Michelle C. Verbrugge, Richard Kelley, Grant M. Mallie, Joan K.

Vogel, and Andrew M. Davenport made key contributions to this report.

Related GAO Products Welfare Reform: States' Implementation and Effects on
the Workforce Development System( GAO/ T- HEHS- 99- 190, Sept. 9, 1999).

Welfare Reform: States' Implementation Progress and Information on Former
Recipients( GAO/ T- HEHS- 99- 116, May 27, 1999).

Welfare Reform: Information on Former Recipients' Status( GAO/ HEHS- 9948,
Apr. 28, 1999).

Welfare Reform: States' Experiences in Providing Employment Assistance to
TANF Clients( GAO/ HEHS- 99- 22, Feb. 26, 1999).

Welfare Reform: States Are Restructuring Programs to Reduce Welfare
Dependence( GAO/ HEHS- 98- 109, June 17, 1998).

Unemployment Insurance: Program's Ability to Meet Objectives Jeopardized(
GAO/ HRD- 93- 107, Sept. 28, 1993).

(205510) Lett er

GAO United States General Accounting Office

Page 1 GAO- 01- 181 Unemployment Insurance

Contents

Contents Page 2 GAO- 01- 181 Unemployment Insurance

Page 3 GAO- 01- 181 Unemployment Insurance United States General Accounting
Office

Washington, D. C. 20548 Page 3 GAO- 01- 181 Unemployment Insurance

Page 4 GAO- 01- 181 Unemployment Insurance

Page 5 GAO- 01- 181 Unemployment Insurance

Page 6 GAO- 01- 181 Unemployment Insurance

Page 7 GAO- 01- 181 Unemployment Insurance

Page 8 GAO- 01- 181 Unemployment Insurance

Page 9 GAO- 01- 181 Unemployment Insurance

Page 10 GAO- 01- 181 Unemployment Insurance

Page 11 GAO- 01- 181 Unemployment Insurance

Page 12 GAO- 01- 181 Unemployment Insurance

Page 13 GAO- 01- 181 Unemployment Insurance

Page 14 GAO- 01- 181 Unemployment Insurance

Page 15 GAO- 01- 181 Unemployment Insurance

Page 16 GAO- 01- 181 Unemployment Insurance

Page 17 GAO- 01- 181 Unemployment Insurance

Page 18 GAO- 01- 181 Unemployment Insurance

Page 19 GAO- 01- 181 Unemployment Insurance

Page 20 GAO- 01- 181 Unemployment Insurance

Page 21 GAO- 01- 181 Unemployment Insurance

Page 22 GAO- 01- 181 Unemployment Insurance

Page 23 GAO- 01- 181 Unemployment Insurance

Page 24 GAO- 01- 181 Unemployment Insurance

Page 25 GAO- 01- 181 Unemployment Insurance

Page 26 GAO- 01- 181 Unemployment Insurance

Page 27 GAO- 01- 181 Unemployment Insurance

Page 28 GAO- 01- 181 Unemployment Insurance

Page 29 GAO- 01- 181 Unemployment Insurance

Page 30 GAO- 01- 181 Unemployment Insurance

Page 31 GAO- 01- 181 Unemployment Insurance

Page 32 GAO- 01- 181 Unemployment Insurance

Page 33 GAO- 01- 181 Unemployment Insurance

Page 34 GAO- 01- 181 Unemployment Insurance

Page 35 GAO- 01- 181 Unemployment Insurance

Page 36 GAO- 01- 181 Unemployment Insurance

Appendix I

Appendix I Scope and Methodology

Page 37 GAO- 01- 181 Unemployment Insurance

Appendix I Scope and Methodology

Page 38 GAO- 01- 181 Unemployment Insurance

Appendix I Scope and Methodology

Page 39 GAO- 01- 181 Unemployment Insurance

Appendix I Scope and Methodology

Page 40 GAO- 01- 181 Unemployment Insurance

Appendix I Scope and Methodology

Page 41 GAO- 01- 181 Unemployment Insurance

Appendix I Scope and Methodology

Page 42 GAO- 01- 181 Unemployment Insurance

Appendix I Scope and Methodology

Page 43 GAO- 01- 181 Unemployment Insurance

Appendix I Scope and Methodology

Page 44 GAO- 01- 181 Unemployment Insurance

Appendix I Scope and Methodology

Page 45 GAO- 01- 181 Unemployment Insurance

Appendix I Scope and Methodology

Page 46 GAO- 01- 181 Unemployment Insurance

Appendix I Scope and Methodology

Page 47 GAO- 01- 181 Unemployment Insurance

Appendix I Scope and Methodology

Page 48 GAO- 01- 181 Unemployment Insurance

Appendix I Scope and Methodology

Page 49 GAO- 01- 181 Unemployment Insurance

Page 50 GAO- 01- 181 Unemployment Insurance

Appendix II

Appendix II State Responses to GAO Questionnaire on UI Eligibility

Page 51 GAO- 01- 181 Unemployment Insurance

Appendix II State Responses to GAO Questionnaire on UI Eligibility

Page 52 GAO- 01- 181 Unemployment Insurance

Page 53 GAO- 01- 181 Unemployment Insurance

Appendix III

Appendix III Comments From the Department of Labor

Page 54 GAO- 01- 181 Unemployment Insurance

Page 55 GAO- 01- 181 Unemployment Insurance

Appendix IV

Page 56 GAO- 01- 181 Unemployment Insurance

Ordering Information The first copy of each GAO report is free. Additional
copies of reports are $2 each. A check or money order should be made out to

the Superintendent of Documents. VISA and MasterCard credit cards are
accepted, also.

Orders for 100 or more copies to be mailed to a single address are
discounted 25 percent.

Orders by mail: U. S. General Accounting Office P. O. Box 37050 Washington,
DC 20013

Orders by visiting: Room 1100 700 4th St. NW (corner of 4th and G Sts. NW)
U. S. General Accounting Office Washington, DC

Orders by phone: (202) 512- 6000 fax: (202) 512- 6061 TDD (202) 512- 2537

Each day, GAO issues a list of newly available reports and testimony. To
receive facsimile copies of the daily list or any list from the past 30
days, please call (202) 512- 6000 using a touchtone phone. A recorded menu
will provide information on how to obtain these lists.

Orders by Internet: For information on how to access GAO reports on the
Internet, send an e- mail message with “info” in the body to:
info@ www. gao. gov or visit GAO's World Wide Web home page at: http:// www.
gao. gov

To Report Fraud, Waste, or Abuse in Federal Programs

Contact one: Web site: http:// www. gao. gov/ fraudnet/ fraudnet. htm e-
mail: fraudnet@ gao. gov 1- 800- 424- 5454 (automated answering system)

United States General Accounting Office Washington, D. C. 20548- 0001

Official Business Penalty for Private Use $300

Address Correction Requested Bulk Rate

Postage & Fees Paid GAO Permit No. GI00
*** End of document. ***