Child Labor: Labor Can Strengthen Its Efforts to Protect Children
Who Work (27-SEP-02, GAO-02-880).				 
                                                                 
In 2001, almost 40 percent of all 16- and 17-year-olds in the	 
United States and many 14- and 15-year-olds worked at some time  
in the year. Children in the United States are often encouraged  
to work, and many people believe that children benefit from early
work experiences by developing independence, confidence, and	 
responsibility. However, the public also wants to ensure that the
work experiences of young people enhance, rather than harm, their
future opportunities. The number and characteristics of working  
children have changed little over the past decade. According to  
Bureau of Labor Statistics data, as in 1990, as many as 3.7	 
million children aged 15 to 17 worked in 2001. The number of	 
children who die each year from work-related injuries has changed
little since 1992, but the number of children who incurred	 
nonfatal injuries while working is more difficult to determine	 
because data from different sources provide different estimates  
of the number of injuries and trends over time. The Department of
Labor devotes many resources to ensuring compliance with the	 
child labor provisions of the Fair Labor Standards Act, including
conducting nationwide campaigns designed to increase public	 
awareness of the provisions, but its compliance efforts suffer	 
from limitations that may prevent adequate enforcement of the	 
law.								 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-02-880 					        
    ACCNO:   A05196						        
  TITLE:     Child Labor: Labor Can Strengthen Its Efforts to Protect 
Children Who Work						 
     DATE:   09/27/2002 
  SUBJECT:   Children						 
	     Data collection					 
	     Labor law						 
	     Labor statistics					 
	     Occupational safety				 
	     Statistical data					 
	     BLS Current Population Survey			 
	     DOL Operation Child Watch				 
	     National Longitudinal Survey of Youth		 

                                                                 
Child Labor: Labor Can Strengthen Its Efforts to Protect Children
Who Work (27-SEP-02, GAO-02-880).				 
                                                                 
In 2001, almost 40 percent of all 16- and 17-year-olds in the	 
United States and many 14- and 15-year-olds worked at some time  
in the year. Children in the United States are often encouraged  
to work, and many people believe that children benefit from early
work experiences by developing independence, confidence, and	 
responsibility. However, the public also wants to ensure that the
work experiences of young people enhance, rather than harm, their
future opportunities. The number and characteristics of working  
children have changed little over the past decade. According to  
Bureau of Labor Statistics data, as in 1990, as many as 3.7	 
million children aged 15 to 17 worked in 2001. The number of	 
children who die each year from work-related injuries has changed
little since 1992, but the number of children who incurred	 
nonfatal injuries while working is more difficult to determine	 
because data from different sources provide different estimates  
of the number of injuries and trends over time. The Department of
Labor devotes many resources to ensuring compliance with the	 
child labor provisions of the Fair Labor Standards Act, including
conducting nationwide campaigns designed to increase public	 
awareness of the provisions, but its compliance efforts suffer	 
from limitations that may prevent adequate enforcement of the	 
law.								 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-02-880 					        
    ACCNO:   A05196						        
  TITLE:     Child Labor: Labor Can Strengthen Its Efforts to Protect 
Children Who Work						 
     DATE:   09/27/2002 
  SUBJECT:   Children						 
	     Data collection					 
	     Labor law						 
	     Labor statistics					 
	     Occupational safety				 
	     Statistical data					 
	     BLS Current Population Survey			 
	     DOL Operation Child Watch				 
	     National Longitudinal Survey of Youth		 

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Product.                                                 **
**                                                              **
** 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-02-880

Report to the Chairman, Subcommittee on Labor, Health and Human Services,
and Education, Committee on Appropriations, U. S. Senate

United States General Accounting Office

GAO

September 2002 CHILD LABOR Labor Can Strengthen Its Efforts to Protect
Children Who Work

GAO- 02- 880

Page i GAO- 02- 880 Child Labor Letter 1

Results in Brief 2 Background 4 Number and Characteristics of Children Who
Work and the

Percentage of Illegally Employed Children Have Not Changed Since 1990 10
Many Children Are Illegally Employed; Illegal Employment Varies

by Age and Gender 15 Fatalities Have Changed Little; Extent of Change in
Other WorkRelated Injuries to Children Is Unclear 20 Labor*s Enforcement
of the Child Labor Provisions of FLSA Has

Limitations 33 Conclusions 46 Recommendations 47 Agency Comments and Our
Evaluation 48

Appendix I Scope and Methodology 51

Estimating the Number of Children Employed Illegally in the United States
51 Using NLSY to Estimate the Employment of 12- to 14- year- olds in

the United States 56 Computing Fatality and Injury Rates for Working
Children 57

Appendix II Recent Legislative Proposals on Child Labor 59 Appendix III
Detailed Information on Children in the

United States Who Work 60 Appendix IV Comments from The Department of
Labor 66

GAO Comments 80

Appendix V GAO Contacts and Staff Acknowledgments 87

GAO Contacts 87 Staff Acknowledgments 87 Contents

Page ii GAO- 02- 880 Child Labor Tables

Table 1: Child Labor Provisions for Nonagricultural Jobs 6 Table 2:
Hazardous Occupations Orders for Nonagricultural Work 7 Table 3:
Percentage of Children Employed Illegally in 1990, 1996,

and 2001 15 Table 4: Percentage of Children Employed in 1996 and 1997, CPS

compared to NLSY97 20 Table 5: Number and Percentage of Work- Related
Fatalities for

Children, by Age, 1992 to 2000 22 Table 6: Number and Percentage of Work-
Related Fatalities and

Fatality Rates for Children, by Race, 1992 to 2000 22 Table 7: Fatality
Rates per 100,000 Hours Worked for Children, by

Industry, 1992 to 2000 24 Table 8: Injury Rates for Children per 100,000
Hours Worked, by

Industry, 1992 to 2000 30 Table 9: Number of Investigations in Which Child
Labor Violations

are Found Each Year, Number of Children Found Illegally Employed, and
Penalties Assessed, Fiscal Years 1990 to 2001 37 Table 10: WHD*s Budget,
Number of Investigators, Total WHD

Investigator Hours, and Hours Devoted to Child Labor 38 Table 11:
Investigations Completed by WHD, Fiscal Year 2001 43 Table 12: Variables
for Which Logistic Regressions Were

Conducted 53 Table 13: Legislation Introduced in the 107th Congress to

Strengthen the Child Labor Provisions of FLSA 59 Table 14: Number of
Children Who Worked in an Average School

Month, by Age, 1990 to 2001 60 Table 15: Number of Children Who Worked in
an Average Summer

Month, by Age, 1990 to 2001 60 Table 16: Percentage of Children Employed,
by Age, 2001 61 Table 17: Number of Children Who Worked in an Average
School

Month, by Gender, 1990 to 2001 61 Table 18: Number of Children Who Worked
in an Average Summer

Month, by Gender, 1990 to 2001 62 Table 19: Number of Children Who Worked
in an Average School

Month, by Race, 1990 to 2001 62 Table 20: Number of Children Who Worked in
an Average Summer

Month, by Race, 1990 to 2001 63 Table 21: Number of Employed Children Who
Worked Without Pay

in a Family Business or Farm, 1994 to 2001 63

Page iii GAO- 02- 880 Child Labor

Table 22: Number of Illegally Employed Children in Average School and
Summer Months, by Age, 1990, 1996, and 2001 64 Table 23: Number of
Illegally Employed Children in Average School

and Summer Months, by Gender, 1990, 1996, and 2001 64 Table 24: Number of
Illegally Employed Children in Average School

and Summer Months, by Race, 1990, 1996, and 2001 65

Figures

Figure 1: Number of Children in the United States Aged 15 to 17 Who
Worked, 1990 to 2001 11 Figure 2: Industries in Which Children Worked,
2001 12 Figure 3: Percentage of Illegally Employed Children, by Industry,

2001 16 Figure 4: Percentage of Children Employed in Prohibited

Occupations, by Occupation, 2001 17 Figure 5: Work- Related Fatality Rates
for Children Aged 15 to 17

per 100,000 Hours Worked, 1992 to 2000 21 Figure 6: Work- Related
Fatalities for Children, by Industry, 1992 to

2000 23 Figure 7: Causes of Work- Related Fatalities for Children, 1992 to

2000 25 Figure 8: Number of Work- Related Injuries to Children, 1992 to

2000 27 Figure 9: Work- Related Injuries to Children, by Industry, 1992 to

2000 29 Figure 10: Causes of Work- Related Injuries to Children, 1992 to

2000 31 Figure 11: Types of Work- Related Injuries to Children, 1992 to
2000 32 Figure 12: Injuries to Children by Number of Days Away from

Work, 1992 to 2000 33 Figure 13: Bookmark That is Part of WHD*s New
*YouthRules!*

Child Labor Initiative 40

Page iv GAO- 02- 880 Child Labor Abbreviations

BLS Bureau of Labor Statistics CPS Current Population Survey FLSA Fair
Labor Standards Act GAO General Accounting Office NIOSH National Institute
for Occupational Safety and Health NLSY National Longitudinal Survey of
Youth WHD Wage and Hour Division

Page 1 GAO- 02- 880 Child Labor

September 27, 2002 The Honorable Tom Harkin Chairman, Subcommittee on
Labor, Health and

Human Services, and Education Committee on Appropriations United States
Senate

Dear Mr. Chairman: In 2001, almost 40 percent of all 16- and 17- year-
olds in the United States and many 14- and 15- year- olds worked at some
time in the year. Children in the United States are often encouraged to
work and many people believe that children benefit from early work
experiences by developing independence, confidence, and responsibility.
However, the public also wants to ensure that the work experiences of
young people enhance, rather than harm, their future opportunities. Over
200,000 children are injured on the job each year and about 70 die as a
result of their injuries, according to a recent report by the Department
of Labor. In addition, while work can have a positive effect on academic
achievement, working too many hours may adversely affect children*s
educational attainment. To protect children from such harmful effects, the
Fair Labor Standards Act (FLSA), with provisions on child labor, was
enacted in 1938. These provisions established the minimum age at which
children can work, limited the number of hours and times of day they can
work, and prohibited them from performing work defined as hazardous. It
also made the Secretary of Labor responsible for defining work that is too
hazardous for children and for enforcing the provisions of the law.

You asked us to update information from our 1991 reports on child labor 1
by providing information on (1) how the number and characteristics of
working children in the United States have changed over the past decade,
(2) whether the number and characteristics of work- related injuries to
children have changed over this same time period, and (3) how well Labor
enforces the child labor provisions of FLSA.

1 See U. S. General Accounting Office, Child Labor: Characteristics of
Working Children,

GAO/ HRD- 91- 83BR (Washington, D. C.: June 14, 1991), and U. S. General
Accounting Office,

Labor*s Child Labor Enforcement Efforts: Developments After Operation
Child Watch,

GAO/ T- HRD- 91- 44 (Washington, D. C.: Aug. 7, 1991).

United States General Accounting Office Washington, DC 20548

Page 2 GAO- 02- 880 Child Labor

To obtain this information, we analyzed data on the characteristics of,
and injuries to, children obtained from Labor*s Bureau of Labor Statistics
(BLS) and injury data from the Department of Health and Human Services*
National Institute for Occupational Safety and Health (NIOSH); reviewed
the child labor provisions of FLSA and its implementing regulations,
agency documents, information from Labor*s investigations database, and
individual case files; interviewed Labor officials in Washington, D. C.,
and 11 field offices (5 regions and 6 district offices), officials from
NIOSH, and other experts on child labor; and developed estimates of the
number of children who are employed illegally by comparing BLS data on
working children to the child labor provisions of FLSA. Most of the
estimates in the report were calculated using data obtained from samples
and, therefore, have sampling errors associated with them. All differences
included in the report were tested for statistical significance at the
0.05 level. See app. I for detailed information on our analysis, including
its limitations. We conducted our work from September 2001 to August 2002
in accordance with generally accepted government auditing standards.

The number and characteristics of working children have changed little
over the past decade. According to BLS data, as in 1990, as many as 3.7
million children aged 15 to 17 worked in 2001. Although minority children
and children from families with incomes of less than $25,000 a year were
less likely to work, when they worked, they generally worked more hours
than the average working child. The estimate of the number of illegally
employed children, which we obtained by analyzing BLS data, also has
changed little since 1990. Overall, approximately 4 percent of all 15- to
17- year- olds who worked were employed illegally in 2001, most commonly
in violation of the provisions of the law that restrict the number of
hours 15- year- olds can work. In addition, Hispanic children, children
who were not citizens, and children from families with lower incomes were
more likely to be employed illegally. While the BLS data provide insight
into the characteristics of children who work and the nature of that work,
the data are far from complete. For example, data on work performed by 14-
yearolds are not collected in the main data set on the United States labor
force although, once children reach age 14, they are legally allowed to
work in many types of jobs, such as bagging groceries and ringing up
purchases at a cash register.

The number of children who die each year from work- related injuries has
changed little since 1992, but the number of children who incurred
nonfatal injuries while working is more difficult to determine because
data from different sources provide different estimates of the number of
Results in Brief

Page 3 GAO- 02- 880 Child Labor

injuries and trends over time. From 1992 to 2000, between 62 and 73
children died each year as a result of work- related injuries. Boys and
Hispanic children; children who worked in construction, agriculture, and
transportation and public utilities; and children who worked in a family
business had a greater likelihood of being killed than other working
children. The data on work- related fatalities collected by BLS are fairly
reliable because they are verified using multiple sources. For nonfatal
injuries, however, the two primary sources of nationwide data provide
significantly different estimates of the number of injuries and trends
over the past decade. BLS data collected from employer records show that
almost 13,000 children* less than 1 percent of all working children* were
injured in 1999. However, data compiled by NIOSH from emergency rooms for
the same year show that over 80,000 children sustained a work- related
injury. Moreover, the BLS data show that nonfatal work- related injuries
to children declined significantly from 1992 to 2000* by more than 40
percent* while preliminary data from NIOSH indicate that the number of
nonfatal work- related injuries to children may have increased over the
same period. Although both data sets have limitations, neither BLS nor
NIOSH officials could explain why the trends over time shown by these two
sources differ so dramatically.

Labor devotes many resources to ensuring compliance with the child labor
provisions of FLSA, including conducting nationwide campaigns designed to
increase public awareness of the provisions, but its compliance efforts
suffer from limitations that may prevent adequate enforcement of the law.
First, Labor only recently developed specific goals for improving employer
compliance in the industries in which most children work and continues to
lack specific goals for industries for which children have high rates of
injuries and fatalities. Second, it has not developed adequate methods of
measuring the success of all of its compliance efforts. Third, Labor does
not use all available data to plan its compliance efforts or assess the
efforts of its local offices to ensure compliance with the child labor
provisions of FLSA. Finally, Labor does not provide sufficient guidance
and training to its local offices on how to use their resources most
effectively or help them consistently apply the provisions of the law.

We are making recommendations to the Secretary of Labor aimed at helping
the agency evaluate the effect of its child labor compliance efforts and
ensure that resources are used more effectively, including developing
goals for improving compliance in industries in which children face the
greatest risk of being injured or killed; measures to use in assessing its
success in meeting goals; using existing data to better plan compliance
efforts and assess success in meeting goals; and providing additional

Page 4 GAO- 02- 880 Child Labor

guidance and training to staff. In commenting on a draft of our report,
Labor disagreed with many of our recommendations and conclusions,
including those related to developing additional program goals, improving
methods for measuring program success, and better using data for
management oversight. We continue to believe that additional improvements
can be made in each of these areas. Labor agreed with the remainder of our
recommendations or agreed to take action on them. Labor also provided
technical comments and clarifications, which we incorporated in the report
as appropriate. Our summary evaluation of the agency*s comments is on
pages 48 and 49. Labor*s general comments and our detailed responses are
provided in full in app. IV.

Several key sources provide information on the number and characteristics
of working children. The Current Population Survey (CPS), compiled monthly
by BLS, is the primary source of information on the United States labor
force. CPS provides nationally representative information on the number
and characteristics of working children age 15 and older, 2 including data
on where children work, the types of jobs they hold, and how many hours a
week they work. In addition, it provides demographic information on
children such as age, race, ethnicity, and family income. Because the data
have been collected for over 50 years, they can be used to show how the
number and characteristics of working children have changed over time.
Labor also compiles two additional sources of nationally representative
data on working children. The first is the National Longitudinal Survey of
Youth (NLSY). The most recent NLSY began in 1997 and is referred to as
*NLSY97*; it contains data on one group of approximately 9, 000 children
born from 1980 to 1984. 3 NLSY provides detailed information on the work
experiences of this group of children over time and captures data not
collected in CPS, such as information on children younger than 15 and in-
depth information on children*s work habits, education, and personal
lives. NLSY, however, cannot be used to show how the number and
characteristics of all working children have changed over time because it
only includes information on one group of

2 Although BLS collects data on 15- year- olds who work, it does not
generally publish most of the information since these individuals are not
as significant a part of the labor force as those age 16 and over due to
the dual constraints of child labor laws and mandatory education.

3 The earliest NLSY interviews began in the mid 1960*s under the
predecessor of Labor*s Employment and Training Administration. In 1979,
Labor began a new longitudinal survey of a second group of children called
*NLSY79.* In 1986, BLS took over the collection and analysis of NLSY.
Background

Page 5 GAO- 02- 880 Child Labor

children born from 1980 to 1984. The second set is the National
Agricultural Workers Survey in which data are collected several times each
year from a sample of crop agricultural workers. This survey provides data
on the number and characteristics of children who work in migrant
agriculture, their educational attainment, and their mobility. The
information can be used to supplement data on children in CPS and NLSY but
it is limited. Labor added questions to the survey in fiscal year 2000 to
obtain additional data on children who work in agriculture but has not
been able to obtain data on a sufficiently large number of children
working in crop agriculture to provide information that is statistically
reliable.

The primary sources of data on children who are injured or killed as a
result of work- related injuries come from BLS and NIOSH. BLS reports all
work- related fatalities, including those for children, in its Census of
Fatal Occupational Injuries published each year. BLS identifies these
fatalities through death certificates and reports from state workers*
compensation agencies, medical examiners, the Occupational Safety and
Health Administration, and the news media. BLS collects data on the number
of nonfatal work- related injuries and illnesses in its Survey of
Occupational Injuries and Illnesses from a sample of the injury records
that employers in private industry are required to maintain. NIOSH
collects data on workrelated injuries in its National Electronic Injury
Surveillance System from a sample of emergency room records. 4

The employment of working children is generally covered by FLSA and its
implementing regulations, which limit the types of jobs, number of hours,
and times of day that children younger than 16 years of age can work.
Generally, most children younger than age 14 are prohibited from working
in nonagricultural employment other than casual free- lance jobs such as
babysitting and delivering newspapers. Children who are 14 and 15 years
old may work in many jobs in retail stores, restaurants, and gas stations.
They may not, however, work in any job considered hazardous, including
jobs in manufacturing, mining, construction, transportation, warehousing,
communications, and public utilities. The provisions also prohibit 14- and
15- year- olds from working during school hours and limit the number of
hours and times of day they can work. (See table 1.)

4 NIOSH collects data on work- related injuries to 14- to 17- year- olds
through a collaborative agreement with the Consumer Product Safety
Commission, which maintains the National Electronic Injury Surveillance
System.

Page 6 GAO- 02- 880 Child Labor

Table 1: Child Labor Provisions for Nonagricultural Jobs Age a Occupations
allowed Hours allowed Times of day Other conditions

13 or younger Few* only casual freelance work not covered

under FLSA, such as babysitting or delivering newspapers.

Not applicable. Not applicable. Not applicable. 14 and 15 Occupations in
retail, food

service, and gas stations. b Up to 3 hours a day and 18 hours a week when
school is in session; up to 8 hours a day and 40 hours a week when school
is not in session.

Between 7 am and 7 pm from Labor Day to June 1 and until 9 pm from June 1
through Labor Day.

Prohibited from working in occupations deemed hazardous by the Secretary
of Labor. Also prohibited from working in many occupations, including
occupations in manufacturing, mining, public messenger service, and most
processing occupations, and from performing work other than office or
sales work in construction, transportation, warehousing and storage, and
communications and public utilities. In addition, prohibited from
performing certain activities, such as working on ladders and baking, in
retail, food service, and gas stations. 16 and 17 Occupations other than

those deemed hazardous for children under age 18 by the Secretary of
Labor.

No restrictions. No restrictions. Prohibited from working in occupations
deemed hazardous by the Secretary of Labor.

a Statutory exemptions allow children to work regardless of age in certain
occupations, including acting and delivering newspapers, and in a family-
owned business. b Regulatory exemptions also allow 14- and 15- year- olds
to work in certain occupations, such as

performing sports- attending duties at professional sporting events, or
pursuant to a work experience and career exploration program approved by
Labor. Children 14 and 15 years of age who work in a family- owned
business are not subject to hours or times of day restrictions.

Source: GAO analysis of the child labor provisions of FLSA and
implementing regulations.

FLSA also authorizes the Secretary of Labor to designate certain types of
jobs and equipment as too hazardous for children under the age of 18. Once
children reach age 16, they are only prohibited from working in jobs or
with equipment covered by these Hazardous Occupations Orders, they are not
limited as to the number of hours or times of day they can work. These
hazardous jobs and types of equipment are specified in 17 Hazardous
Occupations Orders originally issued between 1939 and 1963. (See table 2
for a list of the occupations determined to be hazardous by the Secretary
of Labor.) In 2002, NIOSH completed a review of the Hazardous Occupations
Orders for Labor. Its report, issued in July 2002, made several
recommendations for changes to the orders, including establishing new
hazardous orders prohibiting all children younger than

Page 7 GAO- 02- 880 Child Labor

age 18 from working in the construction industry and from working at a
height of 6 feet or higher on ladders, scaffolds, trees, and other
structures. 5 Labor is in the process of reviewing the report and deciding
what actions it will take in response to the recommendations.

Table 2: Hazardous Occupations Orders for Nonagricultural Work Hazardous
Occupations Order number Prohibited occupation

1 Manufacturing and storing of explosives 2 Driving a motor vehicle and
being a helper on a motor

vehicle 3 Coal mining 4 Logging and sawmilling 5 Working with power-
driven woodworking machines 6 Work that involves exposure to radioactive
substances 7 Working with power- driven hoisting machines 8 Working with
power- driven metal- forming, punching,

and shearing machines 9 Mining, other than coal mining 10 Meat packing or
processing (including using powerdriven meat slicing machines) 11 Working
with power- driven bakery machines 12 Working with power- driven paper-
product machines

(including using paper balers and compactors) 13 Manufacturing brick,
tile, and related products 14 Working with power- driven circular saws,
band saws,

and guillotine shears 15 Wrecking, demolition, or ship- breaking 16
Roofing 17 Excavation

Source: Title 29 of the Code of Federal Regulations, subpart E of part 570
(29 CFR 570.50-. 68).

For jobs in agriculture, the child labor provisions are much less
restrictive. Children of any age may work an unlimited number of hours
(outside of school hours) in nonhazardous jobs, either on a farm owned by
their

5 See Department of Health and Human Services, Public Health Service,
Centers for Disease Control and Prevention, National Institute for
Occupational Safety and Health, National Institute for Occupational Safety
and Health (NIOSH) Recommendations to the U. S. Department of Labor for
Changes to Hazardous Orders. (Washington, D. C., 2002).

Page 8 GAO- 02- 880 Child Labor

parents or on a noncommercial farm 6 with the written consent of their
parents. Children aged 14 and 15 are allowed to work an unlimited number
of hours in nonhazardous jobs outside of school hours 7 without parental
consent and, once they reach age 16, they are allowed to work in
agricultural jobs deemed hazardous. 8

The child labor provisions of FLSA do not cover all children. Children who
work for employers whose annual gross volume of sales is less than
$500,000 and whose work cannot be linked to interstate commerce are not
covered under FLSA, although they may be covered under state child labor
laws. 9 In addition, children who are self- employed are not subject to
the child labor provisions of FLSA. Furthermore, although children who
work for their parents are prohibited from working in occupations and
operating equipment listed in the Hazardous Occupations Orders, they are
not subject to other restrictions of FLSA. When children reach age 18,
they are no longer covered under the child labor provisions of FLSA.

In 2001, several legislative proposals were submitted in the House of
Representatives and the Senate that would strengthen the child labor
provisions of FLSA. The proposals include the Children*s Act for
Responsible Employment of 2001, which would, among other things, increase
the maximum penalties for child labor violations and prohibit children
aged 16 and 17 from working in hazardous occupations in agriculture. The
Young American Workers* Bill of Rights would amend the FLSA to require
employers to obtain work permits for all children age 18 and under who are
still in school, require Labor and the Census Bureau to compile data on
child labor from the states, including data on injuries and illnesses, and
add additional restrictions on child labor such as prohibiting children
from making door- to- door sales for profit. (See app. II for a list of
the legislative proposals introduced in the 107th Congress that would
strengthen the child labor provisions of FLSA.)

6 A noncommercial farm is defined as an agricultural employer that did not
use more than 500 days of labor in any calendar quarter in the preceding
year. WHD officials told us this is the working definition of a *family
farm* and that 500 days equates to approximately seven full- time
employees.

7 Children employed by their parents on a farm owned or operated by their
parents may be employed during school hours. 8 The Hazardous Occupations
Orders for agriculture differ from those for general industry.

9 Each state also has its own child labor laws, which often differ from
the child labor provisions of FLSA. When federal and state laws differ,
the stricter standard applies.

Page 9 GAO- 02- 880 Child Labor

Labor*s Wage and Hour Division (WHD) is responsible for enforcing the
child labor provisions of FLSA. WHD*s national office develops the goals
and performance measures for Labor*s child labor compliance efforts and
reports on the results of its efforts in annual performance plans. The
national office is also responsible for providing guidance and training to
WHD*s regional and district offices and for assessing the results of their
child labor compliance efforts. Much of the responsibility for planning
and executing these efforts, however, is left to the discretion of WHD*s
five regional offices and the 50 district offices that report directly to
the regions. The Child Labor Team Leader in the headquarters office is
responsible for coordinating WHD*s child labor compliance efforts,
including disseminating information and guidance to the regional and
district offices, maintaining the section of the WHD Web site with
information on child labor, maintaining the Field Operations Handbook for
investigators, and overseeing special projects. WHD*s child labor
compliance efforts comprise several strategies: enforcement, partnerships,
education and outreach, and public awareness. Its public awareness
strategy is designed to inform the general public about the child labor
provisions of FLSA through activities such as press releases. WHD*s
education and outreach activities* which WHD also refers to as *compliance
assistance** are specifically targeted to groups that can have an impact
on youth employment, such as teachers, parents, and employers. Education
and outreach activities include publishing and distributing materials such
as bookmarks with information on the rules governing the employment of
children under age 18, and maintaining a Web site with information on the
child labor provisions of FLSA and state child labor laws.

WHD*s enforcement actions include on- site investigations of employers and
other activities designed to bring employers into compliance with the law.
When WHD finds violations of the child labor provisions of FLSA during its
investigations, it may assess penalties. The penalties for child labor
violations depend on the severity of the violations, the number of times
the violations occurred, and aggravating factors such as falsification of
records and whether the employer has a record of previous child labor
violations. The penalties range from $275 for a record- keeping violation
to $11,000 for a violation involving a serious injury or death. 10
Information on

10 In 2002, WHD raised the maximum penalty for a child labor violation
from $10,000 to $11,000. This adjustment was required by the Federal Civil
Penalties Inflation Adjustment Act of 1990 (as amended by the Debt
Collection Improvement Act of 1996), which requires federal agencies to
adjust their civil money penalties to account for inflation.

Page 10 GAO- 02- 880 Child Labor

WHD*s investigations, violations, and penalties assessed is tracked in its
investigations database* the Wage and Hour Investigative Support and
Reporting Database, WHISARD.

Over the past decade, according to data tabulated from CPS, the number and
characteristics of working children have not changed. In 2001, working
children were as sizable a part of the United States labor force as they
were in 1990. Most children worked in a variety of occupations
concentrated in a few select industries, primarily retail trade and
services. The percentage of children who worked illegally, either because
they worked in occupations prohibited under the law or more hours than
allowed, also did not change. In 2001, as in 1990, minority children and
children from families with annual incomes below $25,000 were more likely
than other children to work illegally. However, because of limitations of
the data available, we could not determine the full extent and nature of
children*s employment in the United States, such as the number and
characteristics of children younger than age 15 who work and the
percentage of children younger than 15 who are employed illegally.

In 2001, as many as 3.7 million children between the ages of 15 and 17
worked, about 30 percent of all children in this age range. (See fig. 1.)
Between 1990 and 2001, children as a proportion of the total United States
labor force, as well as the percentage of children who worked, remained
relatively stable. The percentage of children who worked in summer months
fluctuated over the decade from a high of 36 percent in 1990 to a low of
30 percent in 2001. Although most children work for an employer, in 2001,
about 52,000 (2 percent) were self- employed and about 10,000 (less than 1
percent) performed unpaid labor in a family business. Number and

Characteristics of Children Who Work and the Percentage of Illegally
Employed Children Have Not Changed Since 1990

Many Children Work, Primarily in Retail Trade and Services

Page 11 GAO- 02- 880 Child Labor

Figure 1: Number of Children in the United States Aged 15 to 17 Who
Worked, 1990 to 2001

Source: GAO analysis of CPS data.

Throughout the decade, children primarily worked in retail trade, in
businesses such as department stores, grocery stores, and restaurants. In
2001, as in 1990, about 60 percent of all working children were employed
in this industry, mostly in eating and drinking places. (See fig. 2.)
Children generally held jobs in sales occupations, such as running a cash
register or clerking at a store, or in services, 11 such as waiting tables
or cleaning hotel rooms.

11 The term *services* is used in the CPS to categorize both an industry
and a type of job (occupation).

Page 12 GAO- 02- 880 Child Labor

Figure 2: Industries in Which Children Worked, 2001

a *Other* includes finance, insurance and real estate (1 percent);
transportation and public utilities (1 percent); and wholesale trade (1
percent).

Source: GAO analysis of CPS data.

Various data sets provide information about children who work in
agriculture. According to CPS data tabulated by GAO, about 4 percent of
all children who worked were employed in agriculture in 2001. Another data
set, the National Agricultural Workers Survey, provides more detailed
information on the characteristics of children who work in crop
agriculture. According to the data in this survey, the characteristics of
children in crop agriculture changed somewhat over the decade. Although
most children working in crop agriculture throughout the decade were boys
aged 16 and 17, the percentage of boys increased, as did the percentage of
children who were foreign- born and newcomers to this

Page 13 GAO- 02- 880 Child Labor

country. 12 Although most children working in agriculture are 16 years of
age and older, the percentage of younger children (14- and 15- year- olds)
who work in agriculture has increased, from 17 percent to 30 percent. In
addition, both the percentage of children who entered the country
illegally to work in agriculture and the percentage of those who were not
accompanied by their parents or other family members rose over the decade.

Children*s work has a decidedly seasonal nature. More children work in
summer months when school is out of session than in school months. 13 In
2001, 30 percent of all children aged 15 to 17 worked in summer months,
compared to 23 percent who worked in school months. Not only do more
children work in the summer, they also work more hours. In 2001, children
worked an average of 21 hours a week in summer months, compared to 16
hours a week in school months. This is similar to the number of hours they
worked in both summer and school months in 1990. Children also
overwhelmingly worked part- time; 87 percent of all working children
worked part- time in 2001.

Although children worked about the same number of hours in 2001 as in
1990, their average hourly earnings increased by 10 percent. In 1990,
children paid an hourly salary earned an average of $5.70 an hour; 14 in
2001, the average hourly salary was $6.36. From 1990 to 2001, children*s
average hourly earnings exceeded the minimum wage and the amount by which
their earnings exceeded the minimum wage increased. By 2001, their average
hourly earnings exceeded the minimum wage of $5.15 an hour by $1.21.

The older a child is, the more likely he or she is to work and the more
hours he or she is likely to work. These differences are even greater in
the summer when more children work. For example, in summer months in 2001,
43 percent of all 17- year- olds and 33 percent of all 16- year- olds

12 See U. S. General Accounting Office, Child Labor In Agriculture:
Changes Needed to Better Protect Health and Educational Opportunities,
GAO/ HEHS- 98- 193 (Washington, D. C.: Aug. 21, 1998).

13 We defined *school months* as September through May and *summer months*
as June through August, although this simplification does not recognize
factors such as schools that continue classes into June or schools that
operate on a year- round basis.

14 This figure is in 2001 dollars. When not adjusted for inflation,
children working in 1990 actually earned $4.21 an hour. Older Children Are
More Likely

than Younger Children to Work

Page 14 GAO- 02- 880 Child Labor

worked, while only 15 percent of all 15- year- olds worked. Moreover, in
2001, 17- year- olds worked an average of 23 hours a week in summer
months, 5 more hours a week than 15- year- olds and 2 more hours a week
than 16- year- olds.

Children of different ages tend to work in different industries. Older
children are more likely to work in the retail trade industry than younger
children. For example, in summer months in 2001, nearly 60 percent of
employed 17- year- olds worked in retail trade, whereas 38 percent of
employed 15- year- olds worked in this industry. On the other hand, 15-
year- olds are more likely than older children to work in agriculture,
possibly because of the looser age restrictions that provide younger
children with opportunities for employment in agriculture that they do not
have in other industries.

In addition to age, family income and race are also related to children*s
employment. Children from families with lower incomes are less likely to
work than those from higher income families and minority children are less
likely to work than white children. About 17 percent of children in
families with annual incomes below $25,000 a year worked in 2001, compared
to 29 percent of children in families with incomes above $75,000 a year.
15 In 2001, 15 percent of black children and 17 percent of Hispanic
children worked, compared to about 30 percent of white children. 16

Despite the fact that children from lower income families are less likely
to work, when they do work, they tend to work more hours. Children in
families with annual incomes below $25,000 worked an average of 21 hours a
week in 2001, 5 more hours a week than children whose families had annual
incomes of $75,000 or more.

15 We divided annual family incomes into four quartiles: $0 to less than
$25,000; $25, 000 to less than $50,000, $50, 000 to less than $75,000, and
$75, 000 or more. 16 BLS includes Hispanics in CPS data by asking
respondents to categorize children of all races into the ethnicity of
*Hispanic* or *non- Hispanic.* In order to simplify the discussion about
differences among races, we grouped children into one of four races:
*White,* *Black,* *Hispanic* and *Other.* In this grouping, Hispanic
children were those whose race was identified as white, black or other,
but whose ethnicity was identified as Hispanic. Children who identified
themselves as black, white, or other and identified themselves as non-
Hispanic were coded as the race they initially selected. Therefore, in
this report, all children categorized as white, black, or other refer only
to non- Hispanic children and children categorized as Hispanic include
those of all races. These categorizations are not equivalent to those used
by BLS. Poor and Minority Children Are

Less Likely to Work

Page 15 GAO- 02- 880 Child Labor

In addition, while all minority children were less likely to work than
white children, Hispanic children who were employed worked more hours than
other children. In 2001, Hispanic children worked an average of 24 hours a
week, at least 5 more hours a week than other children.

As in 1990, we estimated that as many as 4 percent of all 15- to 17- year-
olds who worked in 2001 worked illegally, either because they worked more
hours than allowed under the law or because they worked in prohibited
hazardous occupations. 17 Because the child labor provisions of FLSA for
15- year- olds are more restrictive, they are more likely to work
illegally than 16- or 17- year- olds. For example, in 2001, over 21
percent of all employed 15- year- olds worked illegally in school months,
compared to 1 percent of all 16- and 17- year- olds. Although the overall
percentage of children who worked illegally in both summer months and
school months remained constant from 1990 to 2001, the percentage of 15-
year- olds who worked illegally in school months increased over the
period. (See table 3.)

Table 3: Percentage of Children Employed Illegally in 1990, 1996, and 2001
1990 1996 2001 School months

15- year- olds 17.6% 15.2% 21.4% 16- year- olds 0. 8 0.9 1.4 17- year-
olds 1. 8 2.0 1.4

Total 3.9% 3.2% 4.0% Summer months

15- year- olds 8. 4% 8.0% 8.6% 16- year- olds 1. 6 1.1 1.0 17- year- olds
2. 6 1.6 1.4

Total 3.1% 2.5% 2.4%

Source: GAO estimates of children employed in violation of FLSA in 1990,
1996, and 2001.

Most of the children who worked illegally were 15- year- olds who worked
more hours than allowed under the child labor provisions of FLSA. Of the

17 Because CPS does not capture data on children who work in occupations
that are illegal under criminal statutes, such as drug dealing and
prostitution, we could not include them in our estimates. We estimated
illegal employment by comparing information reported in CPS to the
restrictions in FLSA and its implementing regulations on the number of
hours that 15- year- olds can work and the industries and occupations in
which 15-, 16-, and 17- year- olds are prohibited from working. See app. I
for detailed information on our estimates of illegal employment and
limitations of the estimates. Many Children Are

Illegally Employed; Illegal Employment Varies by Age and Gender

Page 16 GAO- 02- 880 Child Labor

15- year- olds who worked illegally, nearly 80 percent worked an excessive
number of hours* on average, 10 hours more than the maximum number of
hours allowed. Because of the hours restrictions for 15- year- olds in
school months, they were much more likely to work illegally in school
months than in summer months.

Children who worked illegally not only worked more than the allowed number
of hours, they also worked in prohibited industries and occupations.
Although most children worked in the retail trade industry during the past
decade, a substantial percentage of the illegal employment was found in
other industries (see fig. 3). For example, in 2001, although only 3
percent of all children worked in manufacturing, this industry accounted
for 14 percent of all illegally employed children. Similarly, although
only 3 percent of all working children worked in construction, the
construction industry accounted for 16 percent of all illegally employed
children.

Figure 3: Percentage of Illegally Employed Children, by Industry, 2001

a *Other* includes the forestry and fisheries (2 percent); wholesale trade
(1 percent); finance, insurance and real estate (0.4 percent) and public
administration (0.4 percent) industries. Source: GAO estimates of children
who were employed in violation of FLSA in 2001.

Page 17 GAO- 02- 880 Child Labor

Moreover, although most children worked in sales and services occupations,
children who were employed in prohibited occupations were most commonly
employed illegally in transportation- related occupations. In 2001,
according to our estimates, over 40 percent of all children who were
employed illegally because they worked in prohibited occupations worked as
truck drivers. (See fig. 4.)

Figure 4: Percentage of Children Employed in Prohibited Occupations, by
Occupation, 2001

a *Other* includes roofers (7 percent), butchers (7 percent), and forestry
and logging workers (4 percent). Source: GAO estimates of children who
were employed in violation of FLSA in 2001.

We determined that a child*s gender, race, and citizenship status were
related to the likelihood that he or she worked in a prohibited
occupation, but a child*s annual family income was not related. In 2001,
boys were more likely than girls to work in prohibited occupations. In
addition, black children were less likely than children of other races to
work in prohibited occupations, and children who were not citizens were
more likely than those who were citizens to work in prohibited
occupations. There was no correlation, however, between family income and
children who worked in hazardous occupations. In other words, children
from families with lower annual incomes were not more or less likely to
work in prohibited occupations than other children.

Page 18 GAO- 02- 880 Child Labor

Although children from families with lower incomes were not more likely to
work in prohibited occupations, they were more likely to work more hours
than allowed. In school months in 2001, 32 percent of all employed 15-
year- olds with family incomes less than $25, 000 a year worked more hours
than allowed under the law. In contrast, only 9 percent of all employed
15- year- olds with family incomes over $75, 000 a year worked more hours
than allowed.

In addition to income, race is also associated with the likelihood that a
child will work more hours than allowed; Hispanic children are more likely
to work too many hours than other children. In 2001, 42 percent of
Hispanic 15- year- olds worked more hours than allowed during school
months, compared to 16 percent of whites and 18 percent of blacks.

The region of the country in which children live as well as the type of
area in which they live (metropolitan or nonmetropolitan) are also
associated with illegal employment. A higher proportion of employed 15-
year- olds who lived in the South and West worked more hours than allowed
than those who lived in the Northeast and Midwest. 18 In school months in
2001, 15- year- olds in the South and West were almost 1.5 times more
likely than their counterparts in the Northeast and Midwest to work more
hours than allowed. Additionally, 15- year- olds who lived in metropolitan
areas were 1.4 times more likely to work more hours than allowed under the
law than those who lived in nonmetropolitan areas.

Although data on children who work has been collected in CPS for 50 years,
and therefore provides useful information on trends in child labor,
limitations of the data affect our ability to accurately describe all
children in the United States who work. 19 One limitation is that data on
children younger than age 15 are not collected. In 1989, Labor stopped
collecting data on 14- year- olds in CPS, although these children are
allowed under FLSA and its implementing regulations to work in many jobs.
20 Because CPS does not gather data on 14- year- olds, the estimates
presented

18 The BLS data divide the country into four regions: Midwest, Northeast,
South, and West. See app. I for a list of all states in each region. 19
Additional limitations affecting our analysis are highlighted in app. I.

20 BLS officials told us that the decision to stop collecting data on 14-
year- olds who work was made because of budget constraints and because
they received so few responses for children of this age. Data Limitations
Affect

Child Labor Estimates

Page 19 GAO- 02- 880 Child Labor

in this report paint an incomplete picture of the employment patterns for
younger children.

Another limitation of CPS is the method used to capture information on
working children. While some data are collected directly from children who
work, most of the information in CPS on working children is provided by an
adult member of the household, in most cases a parent. Because the adult
answering the questions may not be aware of the full extent of the child*s
activities, information about the child*s employment may be underreported
or omitted, or the adult may incorrectly identify the industry,
occupation, or hours worked.

Another BLS survey, NLSY, collects information on children differently. In
NLSY, children are directly asked by interviewers to describe their jobs
as well as their activities at work. Because children are asked directly
about their work and are asked more detailed questions about employment
throughout the year, including using a calendar to prompt the child to
account for each period of employment or unemployment during the year, the
percentage of children who work reported in NLSY is much higher than in
CPS. For example, NLSY97 showed that about 24 percent of all 15- year-
olds worked in a particular week during the summer of 1996, 21 while CPS
data showed that 18 percent of 15- year- olds worked in that same period.
In addition, NLSY collects data on younger children who work. For example,
NLSY97 showed that 13 percent of all 14- year- olds worked for an employer
in a particular week during the summer of 1996; CPS does not capture data
on 14- year- olds who work. (See table 4.) NLSY97 also reported that as
many as 44 percent of all 13- year- olds and 34 percent of all 12- year-
olds received income from freelance jobs at some point in 1997. 22

21 The reference period is the week of the 12 th of the month. See app. I
for more information. 22 This type of work as defined in NLSY includes
freelance jobs such as babysitting and mowing lawns.

Page 20 GAO- 02- 880 Child Labor

Table 4: Percentage of Children Employed in 1996 and 1997, CPS compared to
NLSY97 Jan.- May, 1996 June- Aug., 1996 Sept.- Dec., 1996 Jan.- Apr., 1997
CPS NLSY CPS NLSY CPS NLSY CPS NLSY

14- year- olds a 7.8 a 13.2 a 10.4 a 6.4 15- year- olds 8.5 17.1 18.2 23.5
10.0 16.3 8. 9 14.8 16- year- olds 24.6 b 36.9 b 27.5 35.2 23.4 32.9

a CPS does not capture data on children younger than 15 years of age who
work. b Sample sizes for this period were too small provide statistically
reliable data.

Source: Bureau of Labor Statistics, Report on the Youth Labor Force, May
2000 (Revised November 2000) and GAO estimate of employed 14- year- olds
from NLSY97 data.

The number and characteristics of children who die each year as a result
of a work- related injury have changed little over the past decade. From
1992 to 2000, 23 the number of fatalities and the fatality rates for
working children remained fairly constant. It is difficult, however, to
determine whether the number of work- related injuries to children has
changed, because the two primary sources of data on nonfatal injuries to
working children* BLS and NIOSH* provide significantly different estimates
of the number of children injured over the decade. Data from one of the
two sources, however, indicates that the characteristics of children
injured have changed little over the decade.

The number and characteristics of work- related fatalities and fatality
rates for working children remained relatively constant from 1992 to 2000.
Each year, according to data collected by BLS on work- related fatal
injuries, 24 between 62 and 73 children died from injuries sustained while
working, a total of 613 children over the 9- year period. In addition, the
fatality rate for children aged 15 to 17* the number of deaths per 100,000
hours worked* was fairly constant, ranging from approximately 0.006 deaths
per 100,000

23 Data on fatal work- related injuries for 1990 and 1991 collected by BLS
are not comparable to data for later years. Prior to 1992, BLS fatality
data were collected in the Survey of Occupational Injuries and Illnesses,
a sample survey that differs in scope and methodology from the current
Census of Fatal Occupational Injuries. In addition, information on the age
of injured workers was not collected in the survey prior to 1992.

24 To ensure the accuracy of the data in the Census of Fatal Occupational
Injuries, BLS collects occupational fatality information from a number of
different sources, such as death certificates, reports from state workers*
compensation offices, fatality investigation reports, and news reports.
Fatalities Have

Changed Little; Extent of Change in Other Work- Related Injuries to
Children Is Unclear

Number and Characteristics of WorkRelated Fatalities and Fatality Rates
for Working Children Changed Little over the Decade

Page 21 GAO- 02- 880 Child Labor

hours worked in 1998 to about 0.010 deaths per 100,000 hours worked in
1992. 25 (See fig. 5.)

Figure 5: Work- Related Fatality Rates for Children Aged 15 to 17 per
100,000 Hours Worked, 1992 to 2000

Source: GAO estimates of fatality rates calculated from BLS*s Census of
Fatal Occupational Injuries and CPS data.

Demographic data for children who died as a result of a work- related
injury show that most of the children killed were boys 16 years of age or
older. Although girls were as likely as boys to work from 1992 to 2000,
boys were almost eight times more likely to die as a result of a
workrelated injury than girls. In addition, although most of the children
killed (about 60 percent) were aged 16 and 17, a substantial number (20
percent) were 13 years of age or younger. (See table 5.)

25 We calculated fatality rates by dividing the number of deaths stated in
BLS*s Census of Fatal Occupational Injuries by the number of hours
children worked during the same time period as shown in CPS. See app. I
for detailed information on our calculations.

Page 22 GAO- 02- 880 Child Labor

Table 5: Number and Percentage of Work- Related Fatalities for Children,
by Age, 1992 to 2000

Number Percentage

12 and younger 85 14% 13- year- olds 38 6 14- year- olds 42 7 15- year-
olds 78 13 16- year- olds 145 24 17- year- olds 225 37

Total 613 100% a a Percentages shown total more than 100 percent because
of rounding.

Source: BLS*s Census of Fatal Occupational Injuries.

Most of the children younger than age 15 who died as the result of a
workrelated injury were employed in agriculture, and about 50 percent of
the children who died as a result of injuries incurred while working in
agriculture were 14 years old or younger. Moreover, even though most
children who died from a work- related injury were white (75 percent),
Hispanic children had a fatality rate that was twice as high as the rate
for white children. (See table 6.)

Table 6: Number and Percentage of Work- Related Fatalities and Fatality
Rates for Children, by Race, 1992 to 2000 White Hispanic Black Total

Number of Children 460 90 32 613 Percentage of Children a 75% 15% 5% 100%
Fatality Rate per 100,000 Hours Worked b 0.008 0.016 0.006 0.008

a About 5 percent of the fatalities were for children not identified with
one of the races listed above or for whom race was not reported. b The
fatality rates shown are only for children age 15 and older because BLS
does not collect data in

CPS on hours worked for children younger than age 15. Source: GAO
estimates calculated from BLS*s Census of Fatal Occupational Injuries and
CPS data.

Fatality data also show that certain industries and types of businesses
pose a greater danger to working children. During the decade, over 40
percent of the children killed as a result of work- related injuries
worked in agriculture, primarily crop production. Retail trade and
construction accounted for 20 percent and 14 percent of all fatalities,
respectively. (See fig. 6.) In addition, although children who worked in
family businesses accounted for less than 1 percent of all working
children, they accounted for about one- third of all fatalities. Moreover,
children who worked for small employers* those with 10 employees or

Page 23 GAO- 02- 880 Child Labor

less* accounted for nearly two- thirds of the fatalities for which
employer size was reported. About 90 percent of the children killed who
worked for an agricultural employer for which establishment size was
reported worked for employers with 10 or fewer employees. 26

Figure 6: Work- Related Fatalities for Children, by Industry, 1992 to 2000

Notes: Industry data for 13 of the 613 work- related fatalities (2
percent) were not available. Percentages total more than 100 percent
because of rounding. a *Other* includes government (4 percent),
transportation and public utilities

(3 percent), wholesale trade (2 percent), and forestry and fishing (2
percent). Source: BLS*s Census of Fatal Occupational Injuries.

26 This information is from unpublished estimates obtained from BLS.
Although BLS collects data on establishment size as categorized by the
number of employees, these data were only available for about 64 percent
of the fatalities to workers under age 18 in the Census of Fatal
Occupational Injuries for 1992 to 2000. The distribution of fatalities by
establishment size could differ significantly if establishment size were
known for all fatalities. Of the 255 children killed as a result of work-
related injuries that occurred in agriculture, data on establishment size
were only available for 155 children (61 percent).

Page 24 GAO- 02- 880 Child Labor

Data on fatality rates, however, show that the number of hours children
work in each industry needs to be considered as well as the number of
deaths. For example, although many of the children killed worked in retail
trade, the fatality rate was 0.003 per 100,000 hours worked, much lower
than the rate in construction, 0.050, in which 14 percent of the
fatalities occurred. (See table 7.) For transportation and public
utilities, the fatality rate was also relatively high, 0.027, although
this industry only accounted for 3 percent of the fatalities.

Table 7: Fatality Rates per 100,000 Hours Worked for Children, by
Industry, 1992 to 2000

Industry Fatality Rate

Agriculture, forestry, and fishing 0. 050 Construction 0.050
Transportation and public utilities a 0.027 Wholesale trade 0.016
Manufacturing 0.012 Retail Trade 0. 003 Services 0.003

a *Transportation and public utilities* includes transportation industries
such as railroads, taxicab services, and warehousing and storage;
communications industries such as radio, television, and cable services;
and utilities industries such as electric light and power, gas, and
sanitary services.

Source: GAO estimates calculated using BLS*s Census of Fatal Occupational
Injuries and CPS data.

Children die as a result of many different types of work- related
accidents. From 1992 to 2000, 44 percent of all children killed as a
result of a workrelated injury died as a result of transportation- related
accidents, including highway collisions and nonhighway incidents, such as
a fall from a moving vehicle. Other common events included being caught or
compressed by equipment or being struck by a falling or flying object.
(See fig. 7.) For example, in 2000, a 16- year- old boy who worked in a
supermarket was crushed to death by a cardboard box compactor, and another
16- year- old boy was crushed when the forklift he was operating flipped
over and landed on his chest. 27

27 Descriptions of fatal injuries were obtained from Fatality Assessment
and Control Evaluation (FACE) reports compiled by NIOSH. The Census of
Fatal Occupational Injuries does not report information on individual
fatalities.

Page 25 GAO- 02- 880 Child Labor

Figure 7: Causes of Work- Related Fatalities for Children, 1992 to 2000

Source: BLS*s Census of Fatal Occupational Injuries.

Overall, the characteristics of children who died as a result of
workrelated injuries and the characteristics of their fatalities remained
relatively constant from 1992 to 2000, although transportation- related
fatalities appeared to increase. In both 1992 and 2000, 60 percent of the
children who died as a result of work- related injuries were 16 and 17
years old, and 75 percent of children killed for whom race was available
were white. 28 In addition, throughout the decade, the percentage of
children killed who worked in agriculture generally ranged from 33 percent
to 44 percent, although it rose to 58 percent in 1998. 29 While most
characteristics changed little over the decade, the number of children who

28 Race was not available for 8 percent of the children who died as a
result of work- related injuries when data were broken out on an annual
basis for 1992 to 2000, either because it was not reported or because it
did not meet BLS*s publication guidelines designed to protect individuals*
anonymity.

29 Industry was not available for 12 percent of the fatalities when data
were broken out on an annual basis for 1992 to 2000, either because it was
not reported or because it did not meet BLS*s publication guidelines.

Page 26 GAO- 02- 880 Child Labor

died as the result of a work- related transportation incident increased
from 37 percent in 1992 to 52 percent in 2000. 30

The two primary sources of nationwide data on work- related injuries, one
collected by BLS from employer records and the other collected by NIOSH
from emergency room records, differ substantially in their estimates of
the number of working children injured each year, the types of injuries
they sustained, and the trends over time. For 1999, BLS reported that
almost 13,000 children were injured on the job, while NIOSH estimated that
over 80,000 children were injured on the job that year. 31 BLS*s estimate
came from records that employers are required to maintain for all work-
related injuries serious enough to cause children to miss at least one day
of work. NIOSH*s estimate was based on records of injuries treated in
emergency rooms. 32 In addition, different types of injuries were more
prevalent among those reported by employers than those treated in
emergency rooms. BLS data for 1992 to 2000 33 showed that sprains,
strains, and tears were the most common injuries, while NIOSH data
indicated that lacerations were the most common injury.

The BLS and NIOSH data also, over time, indicate different trends in the
numbers of work- related injuries to children. BLS data show that injuries
have decreased by more than 40 percent, from about 22,000 in 1992 to about
13,000 in 1999. However, data from NIOSH suggest that injuries may have
actually increased during the same period, from about 64,000 in 1992 to
about 80,000 in 1999. (See fig. 8.)

30 Cause of injury was not available for 5 percent of the fatalities when
data were broken out on an annual basis for 1992 to 2000, either because
it was not reported or because it did not meet BLS*s publication
guidelines.

31 BLS*s estimates of work- related injuries include data on all children
younger than age 18, although injuries to children younger than age 15
represent less than 1 percent of all injuries. The NIOSH estimates
generally are for children aged 14 to 17, although the estimates for 1998
and 1999 did not include injuries to 14- year- olds. NIOSH officials told
us that the number of work- related injuries for 14- year- olds is very
small.

32 Although NIOSH increased its estimate to over 200,000 based on a study
published in 1988 that indicated only about one- third of all occupational
injuries are treated in emergency rooms, NIOSH officials told us that they
do not know whether the study accurately represents the proportion of
injuries treated in emergency rooms.

33 Data on nonfatal work- related injuries for 1990 and 1991 collected by
BLS are not comparable to data for later years because they were combined
with fatal injuries prior to 1992. Two Data Sources Show

Different Numbers and Trends for Nonfatal Injuries

Page 27 GAO- 02- 880 Child Labor

Figure 8: Number of Work- Related Injuries to Children, 1992 to 2000

a Data were not available from NIOSH for all years. Source: BLS*s Survey
of Occupational Injuries and Illnesses and NIOSH*s National Electronic
Injury Surveillance System.

Due to limitations of both data sources, it is difficult to determine the
true extent of injuries to children who work. Both sets of data
underreport injuries to children. BLS captures only injuries serious
enough to require at least one missed day of work; therefore, data on many
injuries are not captured. Moreover, because most children work part-
time, many children who are injured may not miss a full day of work
because they were not scheduled to work after being injured. In addition,
BLS does not collect data on children who are self- employed; those who
work in federal, state, or local government; those who work for
agricultural employers with fewer than 11 employees; and household
workers. NIOSH data also underreport injuries because children who are
injured may not inform hospital staff that their injuries are work-
related and hospital staff may omit details of injuries that connect them
to work. As a result, some injuries treated in emergency rooms may not be
accurately counted as being work- related. Although BLS and NIOSH
officials recognized that the data they collect underreport the number of
work- related injuries to

Page 28 GAO- 02- 880 Child Labor

children, and recognized the limitations in their data, they emphasized
that their data capture information on different types of injuries and can
be used to complement each other. For example, certain types of injuries
such as lacerations may be treated in emergency rooms but not necessitate
one or more days away from work, while other types of injuries that may
result in missed work days such as sprains and strains may be more
commonly treated in physicians* offices or outpatient clinics than in
emergency rooms. Officials were not, however, able to explain why the
trends indicated by the two sources differed so greatly.

Although BLS data from employer records on work- related injuries to
children are not complete, they provide information on the types of
injuries sustained by children over the past decade. In general, the
characteristics of occupational injuries to children and the
characteristics of children who were injured as indicated by the BLS data
changed little from 1992 to 2000. Demographic data for children injured
show that 60 percent of the injuries occurred to boys, even though they
accounted for only about half of all working children. Moreover, during
this period, children aged 16 and over sustained the vast majority of all
work- related injuries. BLS data also show that, throughout the decade,
most children (84 percent) were injured while working in the two
industries in which children are most likely to work* retail trade and
services. Within these industries, in 2000, children who worked in eating
and drinking places, food stores, general merchandise stores, and health
services had the largest numbers of injuries. The injury rates per hours
worked, however, were lower for the retail trade and services industries
than the rates for children who worked in wholesale trade, transportation
and public utilities, and manufacturing. Characteristics of Injuries

and of Children Injured Changed Little from 1992 to 2000

Page 29 GAO- 02- 880 Child Labor

Figure 9: Work- Related Injuries to Children, by Industry, 1992 to 2000

Source: BLS*s Survey of Occupation Injuries and Illnesses.

However, as with fatalities, the likelihood of injury needs to be
considered in addition to the actual number of injuries. We found that the
risk of injury was highest in some industries with the smallest number of
injuries to children. For example, although many fewer children who worked
in the wholesale trade, transportation and public utilities, and
manufacturing industries sustained a work- related injury than those who
worked in retail trade and in services, children working in wholesale
trade, transportation and public utilities, and manufacturing had higher
rates of injury than those working in retail trade or services. (See table
8.)

Page 30 GAO- 02- 880 Child Labor

Table 8: Injury Rates for Children per 100,000 Hours Worked, by Industry,
1992 to 2000

Industry Injury rate Number of injuries

Wholesale trade 5.38 3,963 Transportation and public utilities 5. 01 2,566
Manufacturing 3.90 7,981 Retail trade 2. 89 94,379 Construction 2.68 4,282
Services 2.03 28,956 Agriculture, forestry, and fishing a 1.38 3,507

a Because most agricultural employers have fewer than 11 employees, and
because these employers are excluded from BLS*s Survey of Occupational
Illnesses and Injuries, injuries to children who work in agriculture are
underrepresented.

Source: GAO estimates of injury rates calculated from BLS*s Survey of
Occupational Injuries and Illnesses and CPS data. Although BLS*s injury
data include injuries to children younger than 15 years of age, CPS data
do not include hours worked by children younger than 15. However, this
discrepancy did not affect our estimates because injuries to children
younger than 15 represent less than 1 percent of all injuries. See details
of our calculations of injury rates in app. I.

Injuries are caused by a variety of factors, but most frequently from
coming into contact with an object or equipment. (See fig. 10.) For
example, in 1998 a 17- year- old worker fractured his hand when a piece of
metal slipped off a power- driven machine he was operating and landed on
his hand. Other common accidents were overexertion, falls, and contact
with hot objects. The causes of work- related nonfatal injuries were
fairly constant over this period.

Page 31 GAO- 02- 880 Child Labor

Figure 10: Causes of Work- Related Injuries to Children, 1992 to 2000

a Bodily reaction and exertion includes injuries caused by bending,
climbing, reaching, twisting, and slipping or tripping without a fall.
Source: BLS*s Survey of Occupational Injuries and Illnesses.

The nature of work- related injuries to children also did not change from
1992 to 2000. Children injured on the job most frequently sustained
strains and tears, cuts and lacerations, and bruises and contusions. (See
fig. 11.)

Page 32 GAO- 02- 880 Child Labor

Figure 11: Types of Work- Related Injuries to Children, 1992 to 2000

Note: Figures in the chart total more than 100 percent due to rounding. a
*Other* includes over 50 different categories of injuries, including
concussions, chemical burns, and

puncture wounds; each category represents less than 1 percent of all
injuries. Source: BLS*s Survey of Occupational Injuries and Illnesses.

In contrast to the decrease in the number of injuries indicated by the BLS
data, the severity of injuries to children (as defined by median number of
days away from work) reported by BLS has remained relatively constant
throughout the decade. About 65 percent of the injuries from 1992 to 2000
required children to miss 5 or fewer days of work, while 20 percent of the
injuries required them to miss more than 10 days. (See fig. 12.)

Page 33 GAO- 02- 880 Child Labor

Figure 12: Injuries to Children by Number of Days Away from Work, 1992 to
2000

Source: BLS*s Survey of Occupation Injuries and Illnesses.

Labor has devoted substantial resources to ensuring compliance with the
child labor provisions of FLSA over the past decade, and has continuously
indicated that child labor is one of the agency*s highest priorities, but
its efforts to improve employer compliance suffer from limitations that
hamper its enforcement of the law. First, while Labor has recently begun
to identify specific, measurable goals for the industries in which
children are most likely to work, it continues to lack goals for those
industries where children face significant risks. Second, it has not
developed methods of measuring the success of many of its child labor
compliance efforts. Third, Labor does not use all available data to plan
its future efforts or measure its progress in improving employer
compliance and ensuring that all working children are adequately protected
under the law. Finally, Labor does not provide sufficient guidance and
training to its regional and district offices on how to use their
resources most effectively or help them consistently apply the child labor
provisions of FLSA. Labor*s Enforcement

of the Child Labor Provisions of FLSA Has Limitations

Page 34 GAO- 02- 880 Child Labor

Since 1990, Labor has developed better goals for increasing employer
compliance with the child labor provisions of FLSA. 34 Over the decade,
its goals for improving compliance have moved from general statements
about conducting investigations and education and outreach activities
designed to increase child labor compliance to more specific goals that
focus on improving compliance rates in certain low- wage industries such
as agriculture and low- wage businesses such as nursing homes and
residential care facilities* rates established through investigations of
employers in those industries. Labor recently established specific goals
for fiscal year 2003 for the industries in which children are most likely
to work.

WHD*s national goals in the early part of the decade focused on increasing
the number of its child labor enforcement activities. For example, Labor*s
goals for fiscal years 1990 and 1992 focused on its plans to dramatically
increase the number of child labor investigations through nationwide
efforts it called *Operation Child Watch.* These high- visibility efforts
involved sending out hundreds of investigators for 1- and 2- day periods
to conduct child labor investigations. As a result of these
investigations, WHD found thousands of child labor violations and assessed
millions of dollars in penalties. Operation Child Watch also gave WHD a
sense of where it could expect to find child labor violations in the
future.

In fiscal year 2000, Labor considered setting specific goals for
industries in which many children worked and were likely to be injured*
the grocery and restaurant industries* but decided not to based on the
results it received from a survey of employers. That year, WHD conducted a
*survey* of employers designed to establish baseline compliance rates for
employers in these industries. 35 In the survey, WHD conducted

34 This improvement is in line with the move towards the greater
efficiency, effectiveness, and accountability established by the
Government Performance and Results Act of 1993, P. L. 103- 62, which
required agencies to set goals and develop performance measures to use in
assessing their progress in meeting the goals by fiscal year 1999 and to
begin reporting on their efforts in fiscal year 2000.

35 WHD refers to these investigations as the *National Child Labor
Survey.* In its initial sample, WHD investigated a total of 200 employers
that employed 3, 890 children under age 18: 68 supermarkets, 65 full-
service restaurants, and 67 fast food restaurants. As part of the survey,
WHD also investigated 119 employers in these industries with previous
child labor violations: 39 supermarkets, 40 full- service restaurants, and
40 fast food restaurants. Goals Have Only Recently

Been Developed for Industries in Which Children Are Most Likely to Work

Page 35 GAO- 02- 880 Child Labor

investigations of statistically valid samples of supermarkets, 36 full-
service restaurants, and fast food restaurants that employ children and
found that the overall compliance rates for these industries were 82
percent, 78 percent, and 70 percent, respectively. It also conducted
investigations of employers with prior child labor violations and found
that, although the compliance rates for two of the three industries were
still lower than the overall compliance rates, the rates had improved from
0 percent previously to 72 percent of supermarkets with prior violations,
52 percent of full- service restaurants, and 72 percent of fast food
restaurants. Because WHD considered these rates to be sufficiently high,
particularly compared to other industries, it decided not to set goals for
these industries for improving compliance with the child labor provisions
of FLSA.

WHD made this decision despite the conclusions reached in its analysis of
the results of the survey. The analysis stated that, while child labor
compliance rates were relatively higher in these three industries than
overall FLSA compliance rates, it would be a *serious mistake* to compare
the findings of the survey with compliance rates found in its surveys of
other industries because all of the requirements of FLSA were not
evaluated in the child labor compliance survey. In addition, the analysis
stated that there were *still enormous child labor compliance issues that
these industries need to address,* with at least 52,000 and as many as
220,000 children illegally employed in the three industries. The analysis
also stated that the compliance rate in some segments of these industries
was lower than the overall compliance rates. For example, it noted that
although most children employed in full service restaurants, 89 percent,
were employed in compliance with the child labor provisions of FLSA, the
compliance rate was much lower, 53 percent, for 14- and 15- year- olds
working in these restaurants.

More recently, WHD developed specific goals for improving overall employer
compliance with FLSA in some low- wage industries, including compliance
with the child labor provisions. For example, WHD*s fiscal year 2002
performance plan established a goal of 75 percent compliance with the
minimum wage, overtime, and child labor provisions of FLSA for employers
in the long- term health care industry (nursing homes and residential care
facilities).

36 Supermarkets were defined as retail establishments primarily engaged in
selling a general line of food. Convenience stores, food marts in gas
stations, and delicatessens were not included in the survey.

Page 36 GAO- 02- 880 Child Labor

In September 2002, in its comments on a draft of this report, Labor
reported that WHD had developed specific, measurable goals in its draft
fiscal year 2003 annual performance plan. The draft plan includes goals
for grocery stores, fast food restaurants, and full- service restaurants,
industries in which children are most likely to work. WHD did not,
however, establish goals for industries in which we found that children
were likely to be employed illegally, such as manufacturing and
construction, or for industries in which children have the highest rates
of fatalities and nonfatal injuries, such as construction, wholesale
trade, and transportation and public utilities.

WHD has not developed adequate methods of measuring the success of all of
its child labor compliance efforts. Over the decade, WHD*s measures of
success have generally consisted of counting the number of its child labor
compliance activities. For example, WHD measured the success of Operation
Child Watch in the early 1990s by citing the large number of violations
and penalties produced. More recently, WHD headquarters officials told us
they look at trends in the number of child labor violations found each
year, and noted that the numbers have dropped dramatically since the early
1990s. (See table 9.) Methods of Measuring the

Success of All Efforts to Improve Compliance Have Not Been Developed

Page 37 GAO- 02- 880 Child Labor

Table 9: Number of Investigations in Which Child Labor Violations are
Found Each Year, Number of Children Found Illegally Employed, and
Penalties Assessed, Fiscal Years 1990 to 2001

FY 90 FY 91 FY 92 F 93 FY 94 FY 95 FY 96 FY 97 FY 98 FY 99 FY 00 FY 01

Number of investigations with child labor violations a 5,889 3,627 3,923
2,358 2,086 1,651 1,820 1,141 1,237 3,117 1,886 2,103 Number of illegally
employed children 39,790 27,655 19,443 10,375 8, 764 6,331 7,873 5,270
5,588 6,766 7,874 9,918 Average number of children per investigation 6. 8
7.6 5. 0 4.4 4. 2 3.8 4. 3 4.6 4. 5 2.2 4. 2 4.7 Penalties assessed (in
millions) $8.5 $12.7 $14.4 $8.2 $6.5 $13.5 b $9.2 $4.9 $4.0 $4.7 $6.2 $6.7
Average penalties per investigation $1,435 $3,508 $3,677 $3,484 $3,101
$8,153 $5,071 $4,319 $3,260 $1,492 $3,276 $3,173

a Child labor violations were found from child labor investigations and
from investigations conducted for other purposes, such as investigations
initiated by complaints about payment of minimum wages or overtime. b The
total penalties for fiscal year 1995 were significantly larger than other
years because they

included one case for which the penalties assessed totaled $5.5 million.
Source: WHD*s investigations database.

It is not clear, however, what factors led to the decrease in the number
of violations. It could have resulted from an increased rate of compliance
among employers, the decrease in the number of child labor investigations
conducted by WHD, or other factors. For example, since Operation Child
Watch, the number of investigator hours devoted to child labor
investigations has declined from a high of 11 percent to 7 percent in
2001, with a low of 5 percent in 1998. (See table 10.)

Page 38 GAO- 02- 880 Child Labor

Table 10: WHD*s Budget, Number of Investigators, Total WHD Investigator
Hours, and Hours Devoted to Child Labor FY 90 FY 91 FY 92 FY 93 FY 94 FY
95 FY 96 FY 97 FY 98 FY 99 FY 00 FY 01

WHD*s budget (in millions) $90.8 $91.3 $94.4 $95.0 $97.1 $100.7 $99.8
$177.9 $121.2 $129.6 $141.7 $152.4 Number of investigators 938 865 835 804
800 809 781 942 942 938 949 945 Total hours (in thousands) 1,265 1,109
1,089 999 887 813 714 741 910 982 968 999 Child labor hours (%) 143

(11%) 83 (7%) 105

(10%) 74 (7%) 73

(8%) 60 (7%) 60

(8%) 41 (6%) 44

(5%) 58 (6%) 71

(7%) 74 (7%) Source: WHD*s investigations database.

While WHD headquarters officials use various factors to measure the
success of their enforcement efforts, they told us that they do not know
how to develop methods of measuring the success of their education and
outreach activities. Officials told us they could tell their activities
generate attention if WHD receives a lot of media inquires and that they
count the number of times individuals access information from WHD*s Web
site. However, WHD does not link its education and outreach activities
with changes in employer behavior. While we recognize the inherent
difficulties in assessing the impact on employer behavior of WHD*s
education and outreach activities, particularly the difficulty of
establishing a cause and effect relationship between changes in compliance
and WHD*s activities, such measurement is important in distributing WHD*s
limited resources. Moreover, we found that some district offices were
trying to develop such methods. For example, officials at one district
office told us that they felt it was important to obtain at least an
indication of success by conducting investigations of employers after
their education and outreach activities. They also said they were working
with researchers at the University of Tennessee to develop methods of
measuring the success of all of their child labor compliance activities.

Despite having difficulty in developing methods of measuring the
effectiveness of its education and outreach activities, WHD headquarters
officials told us that they devote a lot of their resources* both at the
national and local level* to education and outreach and other compliance
assistance activities. Since Operation Child Watch, most of Labor*s
national child labor campaigns have focused on providing education and
outreach and other compliance assistance to employers and on increasing
public awareness of the child labor provisions of FLSA. For example, from
1996 to 2001, WHD conducted its *Work Safe This Summer* initiative
designed to reach children beginning summer jobs. As part of this national

Page 39 GAO- 02- 880 Child Labor

initiative, WHD provided pamphlets and posters with information on the
child labor provisions of FLSA, a Web site with safety tips and
information about the law, and a public service announcement displayed on
movie screens in 22 states and printed on 37 million shopping bags
distributed by K- Mart. Labor*s newest national child labor initiative,
*YouthRules!* announced by the Secretary in May 2002, is aimed at
*educating young workers, parents and employers about workplace hazards,
proper hours of work and ensuring that students* education should be
everyone*s first priority.* The initiative includes partnerships that WHD
has formed with all levels of government as well as businesses, unions,
and advocacy groups to provide training and educational materials, such as
bookmarks that list some of the rules regarding jobs in which children
ages 14 through 17 can work. (See fig. 13.)

Page 40 GAO- 02- 880 Child Labor

Figure 13: Bookmark That is Part of WHD*s New *YouthRules!* Child Labor
Initiative

Front Back Source: Labor*s Wage and Hour Division.

Page 41 GAO- 02- 880 Child Labor

WHD does not use all of the data available to plan its future child labor
compliance efforts or assess their success. WHD does not routinely use
data from BLS to identify the industries and occupations in which children
work, or data from BLS and NIOSH to determine the industries or
occupations in which children are most likely to sustain work- related
injuries. WHD also does not use information from its investigations
database or reports on local child labor initiatives to assess the results
of the regional and district offices* child labor compliance efforts or
hold them accountable for ensuring that the child labor provisions of FLSA
are enforced adequately nationwide.

WHD does not routinely obtain information from BLS on the industries and
occupations in which children work* legally or illegally* or those in
which they are most likely to be injured or killed to use in targeting its
compliance efforts. As a result, WHD may not target sufficient resources
to these areas. For example, we estimated that a significant proportion of
illegally employed children worked in construction and manufacturing*
industries in which WHD targets few of its child labor compliance efforts.
We also found that children were more likely to be injured or killed while
working in several industries in which WHD conducts few child labor
investigations* construction, wholesale trade, and transportation and
public utilities. WHD officials told us that they do not use data from BLS
on the industries and occupations in which children work because the
information is not easily accessible and because the data do not provide
information at the local level that could be used to help plan local
compliance efforts. However, CPS data is available at the regional level
and we developed state- level information by combining several years of
data. For example, by combining CPS data for 4 years (1998 to 2001) and
comparing it to the provisions of FLSA, we estimated the number of
children illegally employed in each state. (See app. I for detailed
information on these estimates.)

WHD also does not utilize information from its investigations database to
determine what types of strategies are most successful in detecting child
labor violations in its investigations of employers. WHD conducts child
labor investigations 37 either as a result of a receiving a complaint
about a possible child labor violation or from a decision by a WHD office
to

37 In this report, we refer to investigations that focus specifically on
child labor, whether initiated by WHD or as the result of a complaint, as
*child labor investigations* although WHD does not use this term. All
Available Data Are Not

Used to Plan or Assess the Success of Child Labor Compliance Efforts

Page 42 GAO- 02- 880 Child Labor

conduct investigations of employers that focus specifically on children
they employ (* self- initiated* investigations). 38 WHD officials told us
that, although they review every complaint they receive about a possible
child labor violation, they do not receive many complaints about child
labor. 39 They also said that WHD*s national office does not require all
district offices to conduct self- initiated child labor investigations. In
fact, since Operation Child Watch, the only national initiative that
required all district offices to conduct child labor investigations was
the child labor compliance survey WHD conducted of the grocery and
restaurant industries in fiscal year 2000. WHD officials also told us that
it is not necessary for the national office to require all of its district
offices to conduct self- initiated child labor investigations because
investigators look for child labor violations during all investigations,
not just those that focus on child labor. However, in reviewing data we
requested from WHD*s investigations database, we found that, although
self- initiated child labor investigations accounted for only 5 percent of
all investigations completed in fiscal year 2001, these investigations
accounted for almost 40 percent of all cases in which child labor
violations were found. (See table 11.) In other words, WHD would have
missed a large proportion of the cases with child labor violations if some
district offices had not conducted selfinitiated child labor
investigations in the absence of a national directive to do so.

38 WHD refers to these investigations as *directed* investigations. In
this report, we use the term *self- initiated* investigations. A self-
initiated investigation can be initiated by WHD*s national, regional, or
district offices.

39 Most complaints WHD receives are from employees who complain about
possible minimum wage or overtime violations of FLSA.

Page 43 GAO- 02- 880 Child Labor

Table 11: Investigations Completed by WHD, Fiscal Year 2001 Investigations
completed Investigations that resulted in child labor violations

Number Percentage Number Percentage

Self- initiated child labor investigations 2,011 5.3% 826 39.3%
Investigations initiated as the result of a child labor complaint

694 1.8 431 20.5 Subtotal- child labor investigations 2,705 7.1 1,257 59.8
Other investigations a 35,350 92.9 846 40.2

Total investigations 38,055 100% 2,103 100%

a *Other investigations* include self- initiated investigations of areas
other than child labor, such as investigations of the agriculture and
garment industries, and investigations initiated as a result of complaints
about topics other than child labor, such as potential minimum wage and
overtime violations.

Source: WHD*s investigations database.

WHD headquarters officials also do not use data from its investigations
database to ensure that district offices nationwide are providing an
adequate level of protection to all children who work by ensuring that all
district offices nationwide conduct at least some minimum number of child
labor investigations. From the data we obtained for fiscal year 2001, we
found that the number of investigations conducted by each district office
and by WHD*s five regions varied widely and that a few district offices
accounted for a large number of the child labor investigations and cases
in which child labor violations were found. For example, in fiscal year
2001, we found that the total number of child labor investigations (both
those initiated by complaints and self- initiated investigations)
conducted by the district offices in each region varied widely, from a
total of 942 child labor investigations in the northeast region to a total
of 227 investigations in the southwest region. We also found that the
number of investigations conducted by each of the district offices varied
widely; 8 of the 50 district offices accounted for 43 percent of all child
labor investigations completed in fiscal year 2001, and 9 of the 50
district offices completed fewer than 10 self- initiated child labor
investigations in fiscal year 2001. WHD headquarters officials were not
able to explain these regional and district variances.

Page 44 GAO- 02- 880 Child Labor

Although WHD*s regional and district offices conduct many local child
labor initiatives, 40 WHD headquarters officials do not use the quarterly
reports prepared by these offices on their local initiatives to ensure
that, nationwide, local child labor initiatives provide adequate
protection for all children who work. WHD headquarters officials told us
that all regional and district offices should have some local child labor
initiatives, unless they have a reasonable justification not to, such as
other higher- priority compliance efforts, and that, in recent years,
every district office has had some local child labor initiatives. However,
the reports we reviewed for fiscal years 1999 through 2002 showed that, in
some years, many districts had no child labor initiatives and the number
of child labor initiatives varied significantly from one region to
another. For example, in fiscal year 2000, almost all of the district
offices in the northeast region had at least one child labor initiative
and two regions had regionwide initiatives that required all of their
district offices to conduct child labor investigations. Some of the
district offices in the southwest region, however, had no local child
labor initiatives planned for fiscal years 1999 through 2002, although our
estimates of illegally employed children showed greater numbers of
children in the South to be illegally employed than in the Northeast.

While WHD leaves many of the decisions about how to deploy enforcement
resources to its regional and district offices, it does not ensure that
these offices receive sufficient guidance and training to properly target
or carry out their child labor compliance efforts. In addition, some
regional and district offices do not use data on previous child labor
violations to target their compliance efforts because WHD has not provided
all staff with adequate training on how to obtain reports from its
investigations database.

WHD*s national office does not provide regional and district offices with
adequate guidance on how to target their child labor compliance efforts.
WHD*s national office has not provided criteria to the regional and
district offices to use in targeting their child labor compliance efforts.
Instead, district offices generally rely on the anecdotal knowledge they
have gained from previous investigations to plan their future compliance
efforts.

40 WHD*s child labor *initiatives* refer to any activities aimed at
increasing compliance with the child labor provisions of FLSA, including
investigations of employers, providing educational materials to employers
and children, and meeting with school officials responsible for issuing
work permits for children. These activities may be initiated by WHD*s
national, regional, or district offices. Sufficient Guidance and

Training Has Not Been Provided to Regional and District Offices

Page 45 GAO- 02- 880 Child Labor

WHD*s national office also does not provide regional or district offices
with data from BLS on the industries and occupations in which children
work, legally or illegally, and are most likely to be injured. Officials
in one district office we visited told us that, because of their past
enforcement efforts, illegal child labor was not a problem in their area
and thus, they did not devote a lot of their enforcement resources to
targeting child labor violations. However, our analysis showed that a
large percentage of children working in this state worked in those
industries where children were most likely to be injured. In addition, our
estimates of the number of illegally employed children in each state
showed that a larger proportion of 15- year- olds in that state were
likely to be employed illegally than those in any other state we visited.
While these estimates do not prove that illegal child labor was a problem
in the area of the state covered by that district office, it is an
indication that potential violations exist undetected.

WHD*s national office also does not provide specific guidance to regional
and district offices on when to assess penalties for child labor
violations. As a result, district office practices for penalty assessment
and collection vary significantly. In reviewing investigations with child
labor violations completed in fiscal years 2000 and 2001 at the district
offices we visited, we found that district officials sometimes reduced the
penalties for child labor violations or did not assess any penalties for
child labor violations while in other, similar cases, they assessed
penalties. For example, one district office we visited generally did not
waive penalties for child labor violations it found, but another district
office waived many of the penalties for the child labor violations it
found, including a violation of a Hazardous Occupations Order involving a
child who was working with dangerous equipment. We also found that the
average percentage of the penalties collected varied significantly among
the district offices we visited* from a low of 49 percent of the original
amount assessed to a high of 82 percent for investigations completed in
fiscal years 2000 and 2001.

Finally, WHD does provide adequate training to all of its regional and
district offices on how to obtain reports from the investigations database
to use in targeting their child labor compliance efforts. Two of the five
regional child labor coordinators and officials at two of the district
offices we visited indicated that they had not been trained on how to
obtain reports from WHD*s database. Officials at the district offices we
visited told us they had staff members with the skills needed to obtain
reports from the investigations database that enabled them to identify
where they had found child labor violations in the past to use in
targeting their future compliance efforts. Officials at two of these
offices, however, said that they had only one person with these skills and
that, if they lost this person,

Page 46 GAO- 02- 880 Child Labor

they would not be able to obtain the reports they need from the
investigations database.

By allowing and encouraging children to work, the nation acknowledges that
children can derive many benefits from working, such as independence,
confidence, and responsibility. However, work may also have negative
consequences for children*s physical, emotional, and educational
development. In recognition of this, Labor devotes significant attention
to enforcing the child labor provisions of FLSA and ensuring safe and
productive work experiences for children. Over the last decade, its
efforts have ranged from conducting investigations to carrying out
nationwide education and outreach campaigns.

Despite the importance that WHD places on ensuring the safety of children
who work, it faces difficulty in showing that its efforts have made a
difference in improving the working conditions of children. WHD*s ability
to effectively enforce the child labor provisions of FLSA is dependent, in
part, upon the establishment of a sound performance management system that
provides program managers with the type of information they need to
evaluate the effectiveness of their child labor compliance efforts and
ensure that their limited resources are used in the most effective manner.
WHD has taken some steps towards developing such a system, as reflected in
the program goals it has established and refined in recent years. However,
WHD continues to lack goals for industries that have high rates of
injuries and fatalities. Further, WHD*s performance management system does
not allow managers to fully assess the impact of its child labor
compliance efforts. As a result, WHD lacks a sound basis for determining,
among other things, the extent to which it should devote resources to
child labor investigations versus its education and outreach and other
compliance assistance activities. Its resource allocation decisions are
also hindered because WHD does not routinely use all of the data available
from BLS and NIOSH to target its child labor compliance efforts to the
places that are most dangerous for children and in which they are most
likely to work illegally. Similarly, by not fully utilizing information
from its own investigations database to evaluate the child labor
compliance efforts of its regional and district offices, WHD is not able
to target its local efforts most effectively or hold local offices
accountable at the national level for enforcing the child labor provisions
of FLSA and ensuring that all children who work are adequately protected.
Finally, even if WHD were to use all existing data as effectively as
possible, it would still not be able to use the data to better target its
child labor compliance efforts for children younger than 15 because BLS*s
CPS data do not include information on Conclusions

Page 47 GAO- 02- 880 Child Labor

14- year- olds who work and NLSY97 does not have information on children
born after 1984. We recognize, however, that collecting these data will
require additional resources.

Regional and district offices also need sufficient guidance and training
to effectively carry out their mission. Because WHD*s national office
leaves many decisions about the appropriate amount of resources to devote
to child labor compliance efforts to its regional and district offices,
but does not provide them with criteria for targeting their efforts or
data on industries and occupations in which children work and are most
likely to be injured, they may not be using their limited resources as
efficiently as possible. Additionally, because district offices do not
receive specific guidance from WHD*s national office about when to assess
penalties for child labor violations, nationwide, they do not consistently
assess penalties for these violations. Finally, because staff in WHD*s
district offices do not receive adequate training on how to use WHD*s
investigations database, they are not able to effectively use all
available information to help them assess their past child labor
compliance efforts and better target their future efforts.

To strengthen WHD*s ability to evaluate the effectiveness of its child
labor compliance efforts and ensure that limited resources are used in the
most effective manner, the Secretary of Labor should

 establish additional specific, measurable goals for WHD*s child labor
compliance efforts for industries in which children are most likely to be
injured or killed;

 develop methods of measuring the success of WHD*s child labor compliance
efforts, including its education and outreach activities;

 routinely obtain data from BLS and NIOSH on the industries, occupations,
and locations in which children work* both legally and illegally* and
sustain work- related injuries and use them to target WHD*s child labor
compliance efforts;

 routinely obtain and review data from its investigations database on the
number and types of investigations conducted by WHD*s district offices and
child labor violations found and use these data to (1) ensure that WHD*s
resources are deployed in the most effective manner and (2) hold regional
and local offices accountable at the national level for ensuring
Recommendations

Page 48 GAO- 02- 880 Child Labor

that all children nationwide are protected under the child labor
provisions of FLSA;

 consider enhancing the data collected on children who work by expanding
CPS to include 14- year olds or beginning additional cohorts of NLSY at
regular intervals, such as every five years.

To provide WHD*s regional and district offices with the information they
need to properly plan and implement their child labor compliance efforts,
the Secretary of Labor should

 provide better guidance to WHD*s regional and district offices on how to
improve employer compliance and specific guidance on when to assess
penalties for child labor violations.

 provide training to all WHD staff on how to obtain information from the
investigations database.

We provided a draft of this report to Labor for comment. Overall, Labor
disagreed with many of the conclusions and recommendations in the draft.
Labor*s comments and our specific responses are included in app. IV.

Labor disagreed with our recommendation to establish specific, measurable
goals for industries in which most children work and in which they are
most likely to be injured or killed. The agency noted that WHD*s draft
fiscal year 2003 performance plan includes specific goals for the
industries that employ the most children. One goal, for example, is to
decrease the percentage of grocery stores and full- service restaurants
with repeat child labor violations by 2 percent. We think these are
positive steps and have revised our report to reflect these efforts.
However, the draft performance plan does not include goals for other
industries because Labor does not believe there are sufficient numbers of
children working in these industries to justify setting such goals. We
believe that this approach reduces attention to some of the industries
that pose significant risks to children. For example, there are no
specific goals for construction, which has high fatality and injury rates.

Labor also disagreed with the conclusions supporting our recommendation to
develop methods of measuring the success of its child labor compliance
efforts. In this regard, Labor noted that it had recently decided to
repeat its compliance survey of grocery stores, full- service restaurants,
and fast food restaurants in fiscal year 2004. While we are Agency
Comments

and Our Evaluation

Page 49 GAO- 02- 880 Child Labor

encouraged by this step, Labor has not developed methods of measuring the
success of other child labor compliance efforts, particularly its
education and outreach activities, which, admittedly, are difficult to
evaluate. In addition, it has not conducted child labor compliance surveys
of other industries.

Finally, Labor took issue with our recommendation to routinely obtain data
from BLS and NIOSH on the industries, occupations, and locations in which
children work and sustain injuries and use them to target WHD*s child
labor compliance efforts. Labor maintains that it does routinely use such
data. However, we found that WHD has no procedures for routinely
collecting data from BLS and NIOSH to plan its child labor enforcement
efforts. Moreover, the district offices we visited confirmed that they do
not receive this information to use in planning their local child labor
efforts.

Labor agreed with the remainder of our recommendations or agreed to take
action on them. For example, it agreed to obtain and review data from the
WHISARD database to better hold local offices accountable, explore the
possibilities of collecting additional data through CPS or NLSY, provide
better guidance to WHD*s regional and district offices concerning penalty
assessment, and provide training to WHD staff on how to obtain information
from the investigations database.

Labor also provided technical comments and clarifications, which we
incorporated in the report as appropriate.

We are sending copies of this report to the Secretary of Labor, the Wage
and Hour Administrator, the Commissioner of the Bureau of Labor
Statistics, and other interested parties. Copies will be made available to
others upon request. This report is also available at no charge on GAO*s
Web site at http:// www. gao. gov.

Page 50 GAO- 02- 880 Child Labor

If you have any questions concerning this report please contact me at
(202) 512- 7215 or Lori Rectanus at (202) 512- 9847. Other major
contributors are listed in appendix. V.

Sincerely yours, Robert E. Robertson, Director, Education, Workforce

and Income Security Issues

Appendix I: Scope and Methodology Page 51 GAO- 02- 880 Child Labor

This appendix discusses in more detail our scope and methodology for (1)
estimating the number of working children in the United States who are
employed illegally, (2) using data from the National Longitudinal Survey
of Youth (NLSY) to estimate the number of 12- to 14- year- olds in the
United States who work, (3) computing fatality and injury rates for
working children, and (4) using data on occupational injuries to calculate
relative risk. Most of the estimates in the report were calculated using
data obtained from samples and, therefore, have sampling errors associated
with them. All differences included in the report were tested for
statistical significance at the 0.05 level.

To estimate the number of children who are employed illegally in the
United States* children employed in violation of the child labor
provisions of the Fair Labor Standards Act (FLSA)* we compared data
obtained from the Bureau of Labor Statistics*s (BLS) Current Population
Survey (CPS) on children in the United States who work to the child labor
provisions of FLSA. Because CPS does not capture data on children in the
United States who work in occupations that are illegal under criminal
statutes, such as drug dealing and prostitution, we could not include them
in our estimates. In developing our estimates, we worked with Douglas
Kruse, Ph. D., a professor at Rutgers University*s School of Management
and Labor Relations and a research associate with the National Bureau of
Economic Research who updated our initial estimate of illegal child labor
in the 1991 GAO report on child labor using CPS data from January 1995 to
September 1997. 1 We compared CPS data for 1990, 1996, and 2001 to the
child labor provisions of FLSA by defining the base population and
identifying illegal employment by occupation and hours. We further refined
this analysis through the use of logistic regression for illegal
occupations and crosstabulations for hours.

We used data from 36 monthly surveys from CPS for the following years:
1990, 1996 and 2001. For each year, we created data sets with children
aged 15 to 17. Each year held roughly 70, 000 observations.

We defined the base population by using the child labor provisions of FLSA
as a guide. Because FLSA does not apply to children who are

1 See *Illegal Child Labor in the United States: Prevalence and
Characteristics,* Industrial and Labor Relations Review, vol. 54, no. 1
(2000). Appendix I: Scope and Methodology

Estimating the Number of Children Employed Illegally in the United States

Defining the Base Population

Appendix I: Scope and Methodology Page 52 GAO- 02- 880 Child Labor

self- employed and the provisions are much less stringent for children who
work in agriculture than for those who work in nonagricultural jobs, we
eliminated children who were self- employed and those who worked in
agriculture from the data set. Key questions concerning child employment
in a family business were not added to CPS until 1994. Therefore, in order
to compare results for 2001 with results for 1990, we included children
who worked in a family business in the 2001 base population, even though
these children are not covered by FLSA. (See app. III for information on
the number of children in the United States who worked in a family
business in 2001.)

To obtain numbers of children who worked illegally, we averaged the
results of the nine school months to obtain the number of children who
worked illegally in school months (defined in our report as September
through May) and the three summer months to obtain the number of children
who worked illegally in summer months (defined as June through August).

To simplify the discussion on racial differences, we combined race and
ethnicity categories. CPS treats ethnicity separately from race. As a
result, a child can be identified in CPS as both *white* and *Hispanic,*
or as both *black* and *Hispanic.* To separately determine the effect of
illegal employment on whites, blacks, and Hispanics, we coded children
whose race was identified as *white* or *black* and whose ethnicity was
identified as *Hispanic,* as *Hispanic* only. For children whose race was
identified as *black* or *white,* and whose ethnicity was identified as
*non- Hispanic,* we included them in the category of race identified (i.
e., as *black* or *white* only). This resulted in four racial categories:
white, black, Hispanic, and other. For 1996 and 2001, *Other* included
Asians and Native Americans. For 1990, *Other* included Asians, Native
Americans and *Other,* which contained anyone who identified a race that
did not fall within the other categories listed.

Hazardous Occupations, 15- to 17- year- olds

Using methods developed in our previous report and refined by Douglas
Kruse, we defined a list of 100 occupational codes that correspond to
activities prohibited by the Hazardous Occupations Orders. Any child in
the base population with one of these codes was included in our estimates
of children employed illegally. Illegal Employment:

Occupations

Appendix I: Scope and Methodology Page 53 GAO- 02- 880 Child Labor
Prohibited Occupations, 15- year- olds

We used a similar method to identify occupations specifically prohibited
for 15- year- olds. 2 In this instance, 216 occupational codes matched the
descriptions of activities prohibited for 15- year- olds to children under
16. Any child identified in the base population with one of these codes
was included in our estimates of children employed illegally.

Regression Results

Due to the small sample sizes of children identified in these groups,
cross tabulations were not statistically significant at the 0.05 level. In
order to obtain information on the likelihood of children who were most
likely to work in illegal occupations, we conducted a logistic regression.

To determine demographic differences in the likelihood of working
illegally, the following variables were included in the model: sex, race,
urbanicity (metropolitan or non- metropolitan area), income, citizenship,
and region. Differences that were found to be statistically significant
are indicated in table 12 below.

Table 12: Variables for Which Logistic Regressions Were Conducted
Demographic characteristics Odds ratio

Urban vs. rural 0.79 Non- citizen vs. citizen 1.75 a Female vs. male 0.21
a Hispanic vs. white 1.14 Black vs. white 0.39 a Other vs. white 0.97
Lowest income quartile vs. highest income quartile 1.40 Second income
quartile vs. highest income quartile 1.15 Third income quartile vs.
highest income quartile 1.10 Midwest vs. West 0.94 Northeast vs. West 0.82
South vs. West 0.90 a The differences between these characteristics were
statistically significant at the 0.05 level.

Source: GAO.

2 Although both 14- and 15- year- olds are prohibited from working in
these occupations, we could not include 14- year- olds in our estimates of
illegal employment because BLS does not capture information on children
younger than 15 who work in CPS.

Appendix I: Scope and Methodology Page 54 GAO- 02- 880 Child Labor

We used the CPS variable that describes the number of hours actually
worked each week to determine whether 15- year- olds 3 worked more hours
than allowed each week under FLSA. Children who worked over 18 hours in a
week in school months, from September to May, were included in our
estimates of illegal employment. Children who worked over 40 hours a week
in summer months, June through August, were also included in our
estimates. We further refined this analysis of the number of 15- year-
olds who worked more hours than allowed under the law by conducting cross
tabulations for race, industry, occupation, sex, geographic region,
metropolitan area, and income.

To estimate levels of illegal employment in each region, we used data for
the four regions defined in CPS: Midwest, Northeast, South, and West. The
states included in each region are

Midwest: Illinois, Iowa, Kansas, Michigan, Minnesota, Missouri, Nebraska,
North Dakota, Ohio, South Dakota, and Wisconsin;

Northeast: Connecticut, Maine, Massachusetts, New Hampshire, New Jersey,
New York, Pennsylvania, Rhode Island, and Vermont;

South: Alabama, Arkansas, Delaware, District of Columbia, Florida,
Georgia, Kentucky, Louisiana, Maryland, Mississippi, North Carolina,
Oklahoma, South Carolina, Tennessee, Texas, Virginia, and West Virginia;
and

West: Alaska, Arizona, California, Colorado, Hawaii, Idaho, Montana,
Nevada, New Mexico, Oregon, Utah, Washington, and Wyoming.

To estimate illegal employment in each state, we combined CPS data for 4
years (48 monthly surveys) from 1998 to 2001. Because observations at the
state level were relatively small, combining data for 4 years increased
the sample size and allowed us to make estimates about the state
population.

3 As with prohibited occupations, although both 14- and 15- year- olds are
restricted from working more than 18 hours a week during school months and
from working more than 40 hours a week during the summer, we could not
include 14- year- olds in our estimates because data on them are not
captured in CPS. Illegal Employment: Hours

for 15- year- olds Regional and State- Level Results

Appendix I: Scope and Methodology Page 55 GAO- 02- 880 Child Labor

The specific nature of child labor laws makes it difficult to reconcile
them with generalized data sources. FLSA and implementing regulations and
guidance describe activities that children are prohibited from performing.
Although CPS data on occupations are fairly specific, they do not list the
actual activities children perform on the job. Therefore, it is not
possible to determine whether each child who works in a prohibited
occupation category is actually performing prohibited activities. As a
result, our estimates of illegal employment are both under- and
overstated. The estimates are understated* probably by a significant
amount* because many activities prohibited, particularly the many types of
equipment children are prohibited from using detailed in the Hazardous
Occupations Orders, cannot be gleaned by reviewing the occupational
categories in CPS. For example, all children younger than 18 are
prohibited from using meat slicers. However, although CPS shows that many
children work in *food services,* it does not indicate how many of them
work with meat slicers. The estimates are also overstated because they
include some activities that are allowed in occupations that are
prohibited. For example, because work as a public messenger is expressly
prohibited for15- yearolds, we included children who worked in the
occupation category *mail carrier* in our estimates of illegal employment.
However, these children could have been delivering mail within an office
or sorting mail in a mailroom, activities that are allowed. It was not,
however, possible for us to determine the extent to which our estimates of
illegal employment are under- or overstated as result of these
limitations.

In addition, our estimates of the number of 15- year- olds are understated
because CPS does not capture information necessary to determine the number
of 15- year- olds who work more hours than allowed on a daily basis or at
prohibited times of day. 4 CPS does not collect information on either the
length of the workday, nor its starting and ending times. As with
prohibited occupations, however, it was not possible to determine the
extent to which our estimates are understated as a result of these
limitations. However, we believe that the amount of this understatement is
large because many of the child labor violations found by WHD in its
investigations of employers are violations of the restrictions on number
of hours and times of day that children are allowed to work.

4 Under FLSA, 14- and 15- year- olds are prohibited from working more than
3 hours on a school day or more than 8 hours on a non- school day, and are
prohibited from working before 7 am all year and after 7 pm during school
months and after 9 pm during summer months (defined by WHD as *between
June 1 and Labor day*). Limitations of Our

Estimates of Illegal Employment

Appendix I: Scope and Methodology Page 56 GAO- 02- 880 Child Labor

Finally, our estimates are overstated to the extent that they include
children whose employment is not covered under FLSA, either because their
employers do not meet the threshold for enterprise coverage under FLSA, 5
or because the child does not work in interstate commerce. CPS does not
collect information on the characteristics of employers for which children
work, such as sales volume or other proxies for annual dollar volume of
sales, or information that could be used to determine whether children
work in interstate commerce. As with the other limitations, however, the
extent to which our estimates are overstated because not all children work
in employment covered under FLSA cannot be determined.

We used data from NLSY97, the survey of children born from 1980 to 1984,
to estimate the number of 12- to 14- year- olds who work. In NLSY,
children are asked whether or not they worked in each of the 52 weeks in a
given year. We coded children noted in NLSY97 as being born in 1982 as 14
years old in 1996 and children born in 1983 as 14 years old in 1997. To
determine employment for these children in 1996 and 1997, we identified
those weeks that contained the 12 th day of the month since the reference
period for NLSY is the 12 th week of a particular month. We determined the
frequency of employed children in each of the 16 weeks and averaged the
result over the total number of months in each period. We performed these
calculations for three periods in 1996 and one period in 1997 as shown in
table 4 in this report.

Most employment questions in NLSY deal with events that take place after
the child*s 14th birthday. As a result, we used a different method to
estimate employment for children younger than 14 years of age. One of the
variables in NLSY97 identified children who indicated that they had
received income from a job in the previous year (1996). We coded children
born in 1983 and 1984 as being 13 years of age and 12 years of age,
respectively. We coded these children who said they had received income
from a job as having been employed at some point in the previous year.

5 In order for a business to be covered by FLSA*s child labor provisions,
its annual dollar volume of sales or other business must be at least
$500,000, and it must have employees engaged in interstate commerce. Labor
can also establish coverage for a child on an individual basis if the
child is involved in interstate commerce. Using NLSY to

Estimate the Employment of 12- to 14- year- olds in the United States

Appendix I: Scope and Methodology Page 57 GAO- 02- 880 Child Labor

To calculate fatality and injury rates for children, we used data from
three BLS sources: (1) data on fatal work- related injuries from the
Census of Fatal Occupational Injuries, (2) data on nonfatal injuries from
the Survey of Occupational Injuries and Illnesses, and (3) data on the
hours children work from CPS.

We used the Census of Fatal Occupational Injuries and Survey of
Occupational Injuries and Illnesses to identify the number and
characteristics of 15- to 17- year- olds who died as a result of a work-
related injury or sustained a nonfatal injury from 1992 to 2000. Because
the number of fatalities in each year was relatively low, we aggregated
the data over the decade to provide statistically reliable information. We
did not aggregate the data on nonfatal injuries over the decade because
the number of injuries reported was large.

We calculated fatality and injury rates by using the total number of hours
children worked as indicated in CPS. We, and some other researchers,
consider this method more appropriate than using the number of children
who work to compute fatality and injury rates because many children work
part- time. Using the number of hours children work provides a better
measure of exposure to injury and leads to a more accurate assessment of
risk than using the number of children.

We used data from 108 monthly surveys to obtain the total number of hours
worked by 15- to 17- year- olds between 1992 and 2000. In computing the
rates for nonfatal injuries, we excluded children who worked for the
government or were self- employed because they are excluded from the data
for nonfatal injuries. We totaled the hours worked each week by all
children for each year and totaled the hours worked each week for children
of different sexes, races, and those who worked in certain industries.
Because the data sets for fatal and nonfatal injuries use different
industry groupings than CPS, we combined industries so that they would
reflect the same industry groupings. Because we were not able to obtain
SOII data that only included 15- to 17- year- olds by industry, we
calculated industry rates by dividing the total number of injuries to
children under 18 by the total hours worked by 15- to 17- year- olds in
each industry. Since CPS data do not include hours worked by children
younger than 15, we were unable to include the hours worked by these
children in our calculations. However, since children under 15 make up
less than 1 percent of all injuries to children, this discrepancy did not
affect our estimates. Computing Fatality

and Injury Rates for Working Children

Appendix I: Scope and Methodology Page 58 GAO- 02- 880 Child Labor

To calculate the fatality rate for each category, we divided the total
number of injuries by the total number of hours children worked and
multiplied the result by 100,000. This figure represents the number of
injuries per hundred thousand hours worked.

Appendix II: Recent Legislative Proposals on Child Labor

Page 59 GAO- 02- 880 Child Labor

Several legislative proposals were introduced in the 107th Congress to
strengthen the child labor provisions of FLSA. In July 2002, the Senate
passed the Traveling Sales Crew Protection Act, legislation designed to
ensure that child employees of traveling sales crews are protected under
FLSA. The law includes provisions that prohibit children under age 18 from
working in traveling sales work where they remain away from home for more
than 24 hours. Other legislative proposals introduced in the 107th
Congress are listed in table 13 below.

Table 13: Legislation Introduced in the 107th Congress to Strengthen the
Child Labor Provisions of FLSA Bill Number Name/ Introduced by Summary

S. 869 CARE Act of 2002 Senator Harkin

Would amend FLSA to increase the maximum penalty for child labor
violations from $10,000 to $15,000 and impose a minimum penalty of $500.
It would limit the exemptions for agriculture to those children employed
on family farms outside of school hours, and all children would be
prohibited from working in hazardous occupations. It also would impose
additional record- keeping requirements on Labor to track the types of
industries and occupations in which children are employed and compile
information on child labor violations, and require employers to report to
the state work- related injuries and illnesses incurred by children.
Finally, it requires Labor to work more closely with certain organizations
with similar responsibilities. H. R. 2239 CARE Act of 2002

Representative Roybal- Allard Similar to S. 869. Also would require Labor
to hire 100 additional inspectors

specifically to enforce compliance with child labor laws and increase the
budget for the Labor Solicitor*s office by 10% for increased prosecution
of child labor law violations, subject to the availability of
appropriations. H. R. 961 Young American Workers Bill

of Rights Representative Lantos

Would amend FLSA to require employers to obtain work permits for all
children age 18 or under who are still in school. Would also require Labor
and the Census Bureau to compile data on child labor from the states,
including data on child labor violations, and produce an annual report.
Contains additional restrictions on child labor, such as prohibiting
children under age 16 from making door- to- door sales for profit and
using fryers and other cooking equipment, and increases the civil and
criminal penalties for child labor violations. Also would require Labor to
establish an Advisory Committee for Child Labor, and to publish and
disseminate the names and addresses of individuals who willfully violate
child labor laws. H. R. 1869 Amy Robinson Memorial Act

Representative Frost Would require employers to give notice to the parents
or guardians of children under

age 18 or of individuals with disabilities employed under special
certificates issued by Labor if they also employ individuals with a
criminal record for violence. H. R. 4768 Garment Consumer*s

Right- to- Know Act of 2002 Representative Velazquez

Would amend FLSA by providing greater access to information about garment
workers in order to prevent or halt sweatshop conditions, including those
involving child labor violations. H. R. 4667 Child Modeling Exploitation

Prevention Act of 2002 Representative Foley

Would amend FLSA to prohibit employers from hiring minors under age 17 to
work in exploitive child modeling which involves the marketing of images
of children rather than the marketing of products or services.

Source: GAO*s analysis of legislative proposals.

Appendix II: Recent Legislative Proposals on Child Labor

Appendix III: Detailed Information on Children in the United States Who
Work

Page 60 GAO- 02- 880 Child Labor

The number and characteristics of children who worked from 1990 to 2001 as
tabulated from the Current Population Survey (CPS) data we obtained from
the Bureau of Labor Statistics (BLS) are detailed in the tables below.

Table 14: Number of Children Who Worked in an Average School Month, by
Age, 1990 to 2001

(In thousands)

15- year- olds 16- year- olds 17- year- olds Total

1990 427 884 1,349 2,660

1991 361 796 1,233 2,389

1992 339 759 1,182 2,280

1993 310 796 1,238 2,344

1994 367 935 1,391 2,693

1995 374 959 1,413 2,746

1996 357 1,015 1,454 2,826

1997 357 979 1,500 2,836

1998 384 1,039 1,591 3,014

1999 400 1,063 1,579 3,043

2000 407 1,051 1,590 3,048

2001 381 960 1,490 2,831

Source: GAO analysis of CPS data.

Table 15: Number of Children Who Worked in an Average Summer Month, by
Age, 1990 to 2001

(In thousands)

15- year- olds 16- year- olds 17- year- olds Total

1990 726 1,295 1,643 3,664

1991 641 1,164 1,573 3,378

1992 655 1,161 1,584 3,400

1993 620 1,255 1,562 3,437

1994 680 1,349 1,795 3,824

1995 737 1,419 1,807 3,963

1996 703 1,436 1,811 3,949

1997 680 1,392 1,834 3,906

1998 700 1,409 1,928 4,037

1999 695 1,397 1,953 4,044

2000 721 1,432 1,873 4,026

2001 595 1,348 1,717 3,661

Source: GAO analysis of CPS data.

Appendix III: Detailed Information on Children in the United States Who
Work

Appendix III: Detailed Information on Children in the United States Who
Work

Page 61 GAO- 02- 880 Child Labor

Table 16: Percentage of Children Employed, by Age, 2001

(In thousands)

All children Employed children Percentage School months

15- year- olds 4, 023 381 9.5% 16- year olds 3,991 960 24 17- year olds
4,062 1,490 36.7

Total 12,076 2, 831 23.4% Summer months

15- year olds 4,022 595 14.8% 16- year olds 4,061 1,348 33.2 17- year olds
3,989 1,717 43.1

Total 12,071 3, 661 30.3%

Source: GAO analysis of CPS Data.

Table 17: Number of Children Who Worked in an Average School Month, by
Gender, 1990 to 2001

(In thousands)

Male Female

1990 1, 354 1,306 1991 1, 205 1,184 1992 1, 164 1,116 1993 1, 199 1,145
1994 1, 377 1,316 1995 1, 376 1,370 1996 1, 407 1,419 1997 1, 444 1,391
1998 1, 518 1,497 1999 1, 565 1,477 2000 1, 524 1,524 2001 1, 394 1,436

Source: GAO analysis of CPS data.

Appendix III: Detailed Information on Children in the United States Who
Work

Page 62 GAO- 02- 880 Child Labor

Table 18: Number of Children Who Worked in an Average Summer Month, by
Gender, 1990 to 2001

(In thousands)

Male Female

1990 1, 956 1,708 1991 1, 848 1,250 1992 1, 807 1,278 1993 1, 878 1,560
1994 2, 007 1,817 1995 2, 108 1,855 1996 2, 019 1,930 1997 2, 068 1,838
1998 2, 107 1,930 1999 2, 203 1,841 2000 2, 120 1,906 2001 1, 875 1,785

Source: GAO analysis of CPS data.

Table 19: Number of Children Who Worked in an Average School Month, by
Race, 1990 to 2001

(In thousands)

White Black Hispanic Other

1990 2, 173 199 210 77 1991 1, 996 151 167 74 1992 1, 907 144 162 70 1993
1, 960 163 157 64 1994 2, 223 219 170 50 1995 2, 291 211 157 50 1996 2,
310 231 200 86 1997 2, 303 226 222 84 1998 2, 439 244 241 90 1999 2, 412
249 259 122 2000 2, 383 264 282 119 2001 2, 220 239 272 100

Source: GAO analysis of CPS data.

Appendix III: Detailed Information on Children in the United States Who
Work

Page 63 GAO- 02- 880 Child Labor

Table 20: Number of Children Who Worked in an Average Summer Month, by
Race, 1990 to 2001

(In thousands)

White Black Hispanic Other

1990 2, 968 295 287 114 1991 2, 741 293 232 112 1992 2, 752 305 241 102
1993 2, 843 281 207 106 1994 3, 071 375 263 74 1995 3, 220 359 280 57 1996
3, 210 354 256 129 1997 3, 111 350 326 119 1998 3, 227 386 315 109 1999 3,
166 374 350 155 2000 3, 191 377 341 118 2001 2, 885 369 304 103

Source: GAO analysis of CPS data.

Table 21: Number of Employed Children Who Worked Without Pay in a Family
Business or Farm, 1994 to 2001

(In thousands)

School months Summer months

1994 37 48 1995 39 43 1996 34 42 1997 30 47 1998 24 33 1999 23 30 2000 23
41 2001 20 27

Source: GAO analysis of CPS data.

Appendix III: Detailed Information on Children in the United States Who
Work

Page 64 GAO- 02- 880 Child Labor

Table 22: Number of Illegally Employed Children in Average School and
Summer Months, by Age, 1990, 1996, and 2001

(In thousands)

1990 1996 2001 School months

15- year olds 71 50 78 16- year olds 7 9 13 17- year olds 23 29 21

Total 102 88 112

Summer months

15- year olds 58 51 48 16- year olds 20 16 13 17- year olds 33 28 24

Total 111 95 86 Source: GAO analysis of CPS Data.

Table 23: Number of Illegally Employed Children in Average School and
Summer Months, by Gender, 1990, 1996, and 2001

(In thousands)

1990 1996 2001 School months

Hours a Male 29 17 29 Female 29 20 33 Reg 3 occupations b Male 10 11 10

Female 1 1 2 Hazardous orders c Male 30 35 33

Female 4 4 5

Total 102 88 112 Summer months

Hours a Male 7 7 11 Female 9 2 2 Reg 3 occupations b Male 56 32 33

Female 3 6 12 Hazardous orders c Male 29 40 22

Female 7 8 7

Total 111 95 86

a Worked more hours than allowed under the child labor provisions of FLSA.
b Worked in occupations in which 15- year- olds are prohibited by
regulation (Reg 3) from working. c Worked in occupations in which 15-,
16-, and 17- year- olds are prohibited in the Hazardous Occupations Orders
from working. Source: GAO analysis of CPS Data.

Appendix III: Detailed Information on Children in the United States Who
Work

Page 65 GAO- 02- 880 Child Labor

Table 24: Number of Illegally Employed Children in Average School and
Summer Months, by Race, 1990, 1996, and 2001

(In thousands)

1990 1996 2001 School months

Hours a White 45 3 7 Black 5 1 .3 Hispanic 7 32 53 Other .8 d 2 Reg 3
occupations b White 10 11 10

Black .4 .3 d Hispanic .3 .1 2 Other .02 .5 .4 Hazardous orders c White 31
36 31

Black .5 .4 .4 Hispanic 3 2 5 Other d .5 2

Total 102 88 112 Summer months

Hours a White 13 d 8 Black d d 1 Hispanic 1 9 4 Other 1 d d Reg 3
occupations b White 26 29 22

Black 3 2 2 Hispanic 7 4 4 Other d 3 1 Hazardous orders c White 47 45 38

Black 2 .8 5 Hispanic 8 2 .7 Other 2 d .5

Total 110 95 86

a Worked more hours than allowed under the child labor provisions of FLSA.
b Worked in occupations in which 15- year- olds are prohibited by
regulation (Reg 3) from working. c Worked in occupations in which 15-,
16-, and 17- year- olds are prohibited in the Hazardous Occupations Orders
from working. d No observations for this category present in the data set.

Source: GAO analysis of CPS data.

Appendix IV: Comments from The Department of Labor

Page 66 GAO- 02- 880 Child Labor

Appendix IV: Comments from The Department of Labor

Note: GAO comments supplementing those in the report text appear at the
end of this appendix.

See comment 1.

Appendix IV: Comments from The Department of Labor

Page 67 GAO- 02- 880 Child Labor

See comment 5. See comment 4.

See comment 3. See comment 2.

Appendix IV: Comments from The Department of Labor

Page 68 GAO- 02- 880 Child Labor

See comment 8. See comment 7.

See comment 3. See comment 6.

Appendix IV: Comments from The Department of Labor

Page 69 GAO- 02- 880 Child Labor

See comment 11. See comment 10.

See comment 9.

Appendix IV: Comments from The Department of Labor

Page 70 GAO- 02- 880 Child Labor

See comment 15. See comment 14. See comment 13.

See comment 12.

Appendix IV: Comments from The Department of Labor

Page 71 GAO- 02- 880 Child Labor

See comment 16. See comment 10.

Appendix IV: Comments from The Department of Labor

Page 72 GAO- 02- 880 Child Labor

See comment 22. See comment 21. See comment 20.

See comment 19. See comment 18.

See comment 17.

Appendix IV: Comments from The Department of Labor

Page 73 GAO- 02- 880 Child Labor

See comment 24. See comment 23.

Appendix IV: Comments from The Department of Labor

Page 74 GAO- 02- 880 Child Labor

See comment 25.

Appendix IV: Comments from The Department of Labor

Page 75 GAO- 02- 880 Child Labor

See comment 27. See comment 26.

Appendix IV: Comments from The Department of Labor

Page 76 GAO- 02- 880 Child Labor

See comment 27.

Appendix IV: Comments from The Department of Labor

Page 77 GAO- 02- 880 Child Labor

See comment 29. See comment 28.

Appendix IV: Comments from The Department of Labor

Page 78 GAO- 02- 880 Child Labor

Appendix IV: Comments from The Department of Labor

Page 79 GAO- 02- 880 Child Labor

Appendix IV: Comments from The Department of Labor

Page 80 GAO- 02- 880 Child Labor

1. Our draft report acknowledged the difficulty of estimating the number
of children employed in violation of the child labor provisions of FLSA
and the limitations of using CPS data to provide such estimates. While
recognizing these limitations, we continue to believe that our analysis
provides the best estimates that can be made with available data. In fact,
we used the same methods as those used to develop estimates cited by BLS
and WHD in several of their publications. For example, in BLS*s November
2000 Report on the Youth Labor Force and WHD*s May 2000 materials prepared
for its *Spring Into Safety* child labor campaign, they cited estimates of
the number of children illegally employed prepared by Douglas Kruse, the
researcher with whom we worked in developing our estimates. WHD stated
that his 1997 study provided *the best currently available estimate of the
number of children employed in violation of Federal or State child labor
laws.*

2. Labor cites the drop in workplace injuries indicated by the BLS data
from the Survey of Occupational Injuries and Illnesses as evidence that
*the workplace is becoming safer for young workers,* but ignores NIOSH
data from emergency room records that indicates that workplace injuries to
children are on the rise.

3. Although we highlighted many of Labor*s efforts to ensure compliance
with the child labor provisions of FLSA, our report was not intended to
describe all of the agency*s child labor enforcement efforts over the past
ten years but rather, as requested, to determine how well Labor enforces
these provisions and to make recommendations for improvement.

4. WHD is correct in its statement that the number of investigator hours
devoted to child labor enforcement in fiscal year 2001, 74,000 hours,
represents the largest investment of investigator time by WHD in the last
5 years. However, our report, which focuses on WHD*s efforts since 1990,
shows that this is lower than the resources dedicated to child labor
enforcement in the first half of the decade. For example, in fiscal years
1990, 1991, and 1992, WHD investigators devoted 143,000 hours, 83,000
hours, and 105,000 hours, respectively, to child labor enforcement.

5. We did not provide information on the number of hours spent by WHD
investigators on activities other than investigations because, although
the investigations database, WHISARD, can track this information, we
found, and WHD officials confirmed, that it did not contain complete
information on the time investigators spend on non- enforcement activities
such as education and outreach. GAO Comments

Appendix IV: Comments from The Department of Labor

Page 81 GAO- 02- 880 Child Labor

6. We continue to believe that the trends in the number of child labor
violations found and the amount of penalties assessed* including the fact
that the numbers have declined significantly since 1990* are not valid
indicators of WHD*s commitment to child labor compliance, nor are they
evidence of the success of its efforts to ensure compliance with the child
labor provisions of FLSA. As noted in our report, because we do not know
what factors led to the changes in the number of violations, it is unclear
whether the increase in the number of violations found by WHD in fiscal
year 2002 compared to fiscal year 2001 indicates a growing problem with
child labor, improvements in WHD*s efforts to identify violations, or
other factors.

7. This statement is misleading. Our draft report recognized that, in
2002, the maximum penalty for a child labor violation was raised from
$10,000 to $11,000. We did not recognize WHD*s efforts to publish a final
rule on the 1999 regulatory proposal to update some of the hazardous
occupations orders or issue new regulations because these actions have not
been completed.

8. This statement is also misleading. Our draft report acknowledged that
NIOSH completed a review of the hazardous occupations orders for Labor,
which made several recommendations for changes to the orders. We also
noted that Labor was in the process of reviewing the report and deciding
the actions it would take in response to NIOSH*s recommendations.

9. We believe that our report provides a comprehensive and balanced
picture of the effectiveness of Labor*s efforts to ensure compliance with
the child labor provisions of FLSA and contains important information on
how these efforts could be improved.

10. The information in our draft report was based on our review of WHD*s
fiscal year 2001 and 2002 annual performance plans. Neither of these plans
contained specific, measurable goals for the industries in which most
children work or in which children are most likely to sustain a serious
injury. However, WHD*s draft performance plan for fiscal year 2003, which
we recently obtained, appears to be a step in the right direction in terms
of setting more specific, measurable goals, although it does not contain
goals for some of the industries with high injury

Appendix IV: Comments from The Department of Labor

Page 82 GAO- 02- 880 Child Labor

and fatality rates. 1 Accordingly, we have revised the information in our
report and the associated recommendations.

11. While the referenced employer compliance survey is an important step
in setting goals for improving employer compliance, it is only one of the
steps required. The baseline data become the starting point for measuring
progress in improving compliance. However, after completing the survey in
fiscal year 2000, WHD chose not to set specific goals for improving
compliance in the grocery and restaurant industries in its fiscal year
2001 and 2002 performance plans. It made this decision even though the
rate of non- compliance for fast food restaurants was 30 percent, and the
non- compliance rate for full service restaurants with previous child
labor violations was almost 50 percent.

12. We do not believe that the most important indicator of compliance is
the percentage of children who are employed illegally. Although we agree
it is an indicator that should be reviewed, a few companies that employ
large numbers of children can have a disproportionate effect on the
numbers. Therefore, we believe the most important indicator of compliance
is the percentage of employers who are in compliance with the child labor
provisions of FLSA, particularly because employers are responsible for
maintaining compliance with the provisions of the law.

13. As noted previously, prior to fiscal year 2003, WHD*s annual
performance plans did not contain specific, measurable goals for improving
compliance with the child labor provisions of FLSA for many of the
industries in which most children work or are likely to sustain work-
related injuries. See GAO comment 10.

14. As Labor*s quotation indicates, the goals in WHD*s fiscal year 2002
performance plan did not establish specific goals for the industries in
which most children work or are likely to sustain work- related injuries.
See GAO comment 10.

1 Although we asked WHD officials about the status of the fiscal year 2003
performance plan and, at our exit conference in August 2002, discussed our
recommendation that they develop more specific, measurable goals for
improving child labor compliance, they did not inform us that they had
completed a draft of WHD*s fiscal year 2003 plan in July 2002 which
contained more specific, measurable goals or provide us with a copy of the
plan at that time.

Appendix IV: Comments from The Department of Labor

Page 83 GAO- 02- 880 Child Labor

15. As noted previously, we have revised the report to reflect the goals
contained in WHD*s draft performance plan for fiscal year 2003. See GAO
comment 10.

16. The fact that WHD conducts investigations in these industries does not
negate the need for WHD to set goals for increasing compliance with the
child labor provisions of FLSA for these industries. Because WHD conducts
many of its investigations in response to complaints about possible
violations, the resource allocations indicated in Labor*s charts may
reflect WHD*s responses to complaints as much as the results of its
efforts to target child labor violations in these industries.

17. We disagree that our analysis of fatality rates is a *flawed
interpretation of the BLS and NIOSH data* and that is more appropriate to
use the actual numbers of children employed, injured, and killed to
determine risk. While we understand the importance of reviewing the number
of children who work in each industry and the numbers of children injured
or killed, we believe that injury and fatality rates are important
indicators that WHD should use in allocating its resources. As noted in an
article by a researcher at BLS about fatality rates, *although [fatality]
counts are informative in identifying worker groups that experience large
numbers of fatalities, they do not by themselves measure risk. To quantify
risk, the data on workplace fatalities must be associated with a measure
of worker exposure to risk, such as employment or hours worked. The number
of hours worked is preferable because different workers spend variable
hours on the job in a given time period (e. g. year), and therefore have
different lengths of exposure to workplace hazards.* 2 We believe it is
particularly important to use the number of hours worked, rather than
simply using the number of children employed, to calculate injury and
fatality rates because so many children (87 percent) work part time.

18. For the transportation and public utilities industry, Labor mentions
only fatality rates and the number of fatalities. In addition to high
fatality rates for this industry, we found that the injury rate was high
compared to other industries. Therefore, we believe that it is important
for WHD to focus some of its resources on reviewing child labor in this
industry.

2 See *A Relative Risk Analysis of Workplace Fatalities* by John W. Ruser,
a Research Economist with BLS*s Compensation Research and Program
Development Group, Office of Compensation and Working Conditions.

Appendix IV: Comments from The Department of Labor

Page 84 GAO- 02- 880 Child Labor

19. It is not clear that establishing a goal for improving compliance in
the transportation and public utilities industry would be unrealistic and
counterproductive. Given the limitations of the estimates of illegal
employment, we question the agency*s estimate that only 4,400 youth are
employed illegally in this industry. The statistics cited by Labor on the
number of violations found as evidence that violations are not prevalent
in this industry are also questionable because it is unclear whether WHD
would have found more child labor violations if it had targeted employers
in this industry.

20. Labor cites only the fatality rates and number of deaths in the
wholesale trade industry, although we found that the injury rates for
children working in this industry were also relatively high. While the
number of children working in this industry may be small, we believe the
injury rates to be sufficiently high to warrant further WHD attention. WHD
should look to the extensive experience of its staff cited on page 10 of
Labor*s comments to identify the areas of the country in which children
are most likely to work in these industries.

21. As with the transportation and public utilities and the wholesale
trade industries, Labor only cites the fatality rates for the construction
industry. Again, we noted that the injury rate for children who work in
this industry was relatively high. In addition, we question how WHD
determined that *many of these specialty trade contractors are so small
that they do not meet the FLSA coverage criteria* because this information
is not tracked in WHD*s investigations database and we found no data on
employer coverage under FLSA. We also found, in our discussions with WHD
officials, that establishing coverage for employers in the construction
industry was not a problem for some of the district offices we visited,
particularly those on the east coast.

22. CPS data on occupations show that many children are employed as truck
drivers. However, our draft report incorrectly indicated that 43 percent
of all children who are employed illegally are employed as truck drivers.
We clarified the report to show that the percentage of children who are
employed illegally as truck drivers represents 42 percent of children who
are employed in prohibited occupations.

23. We continue to believe that the injury and fatality rates for children
who work in the construction industry warrant WHD attention. We also
commend WHD*s efforts to fund research by NIOSH on the health and safety
of young workers in the industry.

Appendix IV: Comments from The Department of Labor

Page 85 GAO- 02- 880 Child Labor

24. As noted in our report, manufacturing has one of the highest injury
rates for children. WHD*s opinion that targeting additional resources on
this industry would not be a wise resource investment and would have no
impact on occupational injuries and fatalities for children working in
this industry does not factor in these high injury and fatality rates. In
fact, some of the most serious child labor violations WHD found occurred
in sawmills and companies that produce wood pallets.

25. We commend WHD*s recent decision to conduct a second compliance survey
of the grocery, full service restaurant, and fast- food restaurant
industries in fiscal year 2004. However, as stated in our report, WHD has
not developed methods of measuring the success of all of its child labor
compliance efforts, particularly its education and outreach and other
compliance assistance activities. In addition, WHD has not conducted child
labor compliance surveys of other industries.

26. In our conversations with WHD headquarters officials and in WHD
publications, the number of investigations, violations found, and civil
monetary penalties were frequently cited as indicators of the success of
the WHD*s child labor enforcement efforts. In fact, on page 3 of the
agency*s comments on our report, Labor cites the increase from fiscal
years 2000 to 2001 in the number of child labor violations and the total
amount of penalties assessed as evidence of *WHD*s continued strong
commitment of resources to child labor compliance.*

27. We disagree with Labor*s statement that WHD routinely uses BLS and
NIOSH data to plan the allocation of its child labor resources. We found
that WHD has no procedures for routinely obtaining data from BLS and NIOSH
to plan its child labor enforcement efforts and the district offices we
visited confirmed that they do not receive this information for use in
planning their local child labor efforts. Second, Labor*s statement that
WHD routinely uses state workers* compensation data to plan its
enforcement efforts must be qualified to indicate that many states refuse
to release this information to WHD because of privacy concerns. Third, WHD
has, in the past, used only limited data from its WHISARD investigations
database to plan its enforcement efforts. In some cases, reports we
requested from the database, such as reports that showed the source of
investigations completed in fiscal years 2000 and 2001, was the first time
that WHD had run such reports. Finally, we believe the anecdotal
information on injuries and fatalities maintained by WHD*s national office
does not provide enough information to be useful in making WHD*s resource
allocation decisions. The information on fatalities and serious injuries
to children tracked by WHD*s national office and provided to us in May

Appendix IV: Comments from The Department of Labor

Page 86 GAO- 02- 880 Child Labor

2002 contained data on only 7 fatalities and 36 injuries reviewed by WHD
in fiscal year 2001 and 7 fatalities and 32 injuries reviewed in fiscal
year 2002.

28. We agree with Labor*s statement that our recommendation to consider
enhancing the data collected on children who work by expanding CPS to
include 14- year- olds or beginning additional cohorts of NLSY cannot be
implemented without additional study and resources, which is why we
recommended that the Secretary *consider* these actions instead of
recommending that they be taken. We clarified our conclusions to
acknowledge that collecting these data may require additional resources.

29. We continue to believe that WHD*s regional and district offices cannot
target their resources in the most effective manner without having better
criteria from WHD*s national office on how to target their enforcement
efforts, including information from WHD*s investigations database on
previous child labor investigations and from BLS and NIOSH on the
industries in which children work and those in which they are most likely
to be injured or killed.

Appendix V: GAO Contacts and Staff Acknowledgments

Page 87 GAO- 02- 880 Child Labor

Lori Rectanus, (202) 512- 9847 Revae E. Moran, (202) 512- 3863

Other major contributors to this report are Wendy Ahmed, Amy E. Buck,
Beverly A. Crawford, Charla R. Gilbert, Julian P. Klazkin, Ellen L.
Soltow, and Corinna A. Nicolaou. Appendix V: GAO Contacts and Staff

Acknowledgments GAO Contacts Staff Acknowledgments

(130071)

The General Accounting Office, the investigative arm of Congress, exists
to support Congress in meeting its constitutional responsibilities and to
help improve the performance and accountability of the federal government
for the American people. GAO examines the use of public funds; evaluates
federal programs and policies; and provides analyses, recommendations, and
other assistance to help Congress make informed oversight, policy, and
funding decisions. GAO*s commitment to good government is reflected in its
core values of accountability, integrity, and reliability.

The fastest and easiest way to obtain copies of GAO documents at no cost
is through the Internet. GAO*s Web site (www. gao. gov) contains abstracts
and fulltext files of current reports and testimony and an expanding
archive of older products. The Web site features a search engine to help
you locate documents using key words and phrases. You can print these
documents in their entirety, including charts and other graphics.

Each day, GAO issues a list of newly released reports, testimony, and
correspondence. GAO posts this list, known as *Today*s Reports,* on its
Web site daily. The list contains links to the full- text document files.
To have GAO e- mail this list to you every afternoon, go to www. gao. gov
and select *Subscribe to daily E- mail alert for newly released products*
under the GAO Reports heading.

The first copy of each printed report is free. Additional copies are $2
each. A check or money order should be made out to the Superintendent of
Documents. GAO also accepts VISA and Mastercard. Orders for 100 or more
copies mailed to a single address are discounted 25 percent. Orders should
be sent to:

U. S. General Accounting Office 441 G Street NW, Room LM Washington, D. C.
20548

To order by Phone: Voice: (202) 512- 6000 TDD: (202) 512- 2537 Fax: (202)
512- 6061

Contact: Web site: www. gao. gov/ fraudnet/ fraudnet. htm E- mail:
fraudnet@ gao. gov Automated answering system: (800) 424- 5454 or (202)
512- 7470

Jeff Nelligan, managing director, NelliganJ@ gao. gov (202) 512- 4800 U.
S. General Accounting Office, 441 G Street NW, Room 7149 Washington, D. C.
20548 GAO*s Mission

Obtaining Copies of GAO Reports and Testimony

Order by Mail or Phone To Report Fraud, Waste, and Abuse in Federal
Programs

Public Affairs
*** End of document. ***