Collective Bargaining Rights: Information on the Number of	 
Workers with and without Bargaining Rights (13-SEP-02,		 
GAO-02-835).							 
                                                                 
In 1935, the federal National Labor Relations Act provided U.S.  
workers the right to bargain over wages, hours, and other terms  
and conditions of employment with their employers, forming the	 
framework for collective bargaining in the United States. The Act
allowed workers to join together to form unions and required that
employers recognize certified employee unions and bargain "in	 
good faith." Although the Act applied broadly to "employees," it 
and subsequent amendments excluded certain groups of workers from
its coverage. Three-quarters of the civilian workforce--or 103	 
million of the 135 million people in the labor force as of	 
February 2001--had some form of collective bargaining rights from
federal, state, or local statutes. In contrast, 32 million	 
civilian workers were without collective bargaining rights under 
any law, either federal and state. The portion of the total labor
force with collective bargaining rights has likely increased	 
since 1959. Since 1959, no major group of workers has lost	 
bargaining rights under the Act. However, other federal, state,  
and local laws have extended rights to some workers in the groups
excluded from the Act, providing bargaining rights to 14.5	 
million workers, primarily nonprofit health care workers;	 
federal, state, and local government workers; and agricultural	 
workers. Under two recent Supreme Court cases affecting National 
Labor Relations Board decisions, some workers currently with	 
bargaining rights may either lose bargaining rights or have their
rights diminished. In the Kentucky River decision, the Supreme	 
Court ruled that the Board should revise its test for determining
whether a worker is a supervisor, and excluded group under the	 
Act, finding that the Board's test served to categorically	 
include certain employees as covered under the act. Because any  
future tests used by the Board to determine whether or not	 
employees are supervisors should be less categorical and more	 
case-specific, the Kentucky River decision could increase the	 
number of employees considered supervisory and thus excluded from
coverage under the act. In the case of Hoffman Plastic, the Court
reversed the Board's decision to award back pay to an		 
undocumented alien worker who was fired for union activity.	 
Although the Court did not exclude undocumented alien workers	 
from protection under the Act, it prohibited the Board from	 
awarding back pay to these documented alien workers whose rights 
had been violated, stating that this remedy would conflict with  
federal immigration law.					 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-02-835 					        
    ACCNO:   A05057						        
  TITLE:     Collective Bargaining Rights: Information on the Number  
of Workers with and without Bargaining Rights			 
     DATE:   09/13/2002 
  SUBJECT:   Collective bargaining				 
	     Labor force					 
	     Labor unions					 
	     Workers compensation				 
	     Census Bureau Current Population Survey		 

******************************************************************
** 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-02-835

Report to Congressional Requesters U. S. Senate

United States General Accounting Office

GAO

September 2002 COLLECTIVE BARGAINING RIGHTS

Information on the Number of Workers with and without Bargaining Rights

GAO- 02- 835

Page i GAO- 02- 835 Collective Bargaining Letter 1

Results in Brief 2 Background 4 Three- Quarters of U. S. Workers Have
Collective Bargaining Rights 5 Over 32 Million Workers Do Not Have
Collective Bargaining Rights 10 Percentage of Total Labor Force with
Bargaining Rights Likely

Increased Since 1959 14 Two Recent Supreme Court Cases Could Affect Gains
in Coverage

over the Last 40 Years 18

Appendix I Methodology 23 Overall Approach 23 Legal Review 23 Definition
of Bargaining Rights 23 Workforce Estimations and Analysis 24

Tables

Table 1: Number of Workers Receiving Collective Bargaining Rights from
1959 through February 2001 16 Table 2: Net Effects of Changes in Excluded
Groups on Percent of

the Labor Force That Has Bargaining Rights, 1959 and 2001 17 Table 3:
Minimum Annual Dollar Sales Volumes for Coverage under

NLRA 27 Table 4: Definitions of Supervisory Leveling Factors, 1997 NCS 30
Table 5: Comparison of NLRA Definitions of Private Sector

Excluded Groups and Those Used in GAO*s Quantitative Estimates 35 Table 6:
Census Level Estimates of the Foreign- Born Population by

Migrant Status in 2000 37 Table 7: Estimates of Residual Foreign- Born by
Age 37

Figures

Figure 1: Private Sector Collective Bargaining Coverage and Key Groups
with No Rights, February 2001 6 Figure 2: Comparison of Number of Workers
with Collective

Bargaining Rights with Total Number of Workers, by Industry, February 2001
7 Contents

Page ii GAO- 02- 835 Collective Bargaining

Figure 3: Percent of Work Force That Has Collective Bargaining Rights and
Percent Who Are Independent Contractors by Private Sector Industry,
February 2001 8 Figure 4: Percentage of Workers Affected by the Board*s
Dollar

Sales Volume Standards, by Industry, February 2001 28

Abbreviations

AFSCME American Federation of State, County, and Municipal Employees BLS
Bureau of Labor Statistics CES Current Employment Statistics Survey CPS
Current Population Survey INS Immigration and Naturalization Service IRCA
Immigration Reform and Control Act of 1986 NCS National Compensation
Survey NLRA National Labor Relations Act SBA Small Business Administration
SIC Standard Industry Classification SOC Standard Occupational
Classification USDA U. S. Department of Agriculture

Page 1 GAO- 02- 835 Collective Bargaining

September 13, 2002 The Honorable Edward M. Kennedy Chairman, Committee on
Health, Education,

Labor, and Pensions United States Senate

The Honorable Paul Wellstone Chairman, Subcommittee on Employment,

Safety and Training, Committee on Health, Education,

Labor, and Pensions United States Senate

In 1935, the federal National Labor Relations Act (NLRA), 1 also known as
the Wagner Act, provided many U. S. workers the right to bargain over
wages, hours, and other terms and conditions of employment with their
employers, forming the framework for collective bargaining in the United
States. The NLRA not only allowed workers to join together to form unions,
but also required that employers recognize certified employee unions and
bargain *in good faith.* While the NLRA applied broadly to *employees,* it
and subsequent amendments excluded certain groups of workers from its
coverage. Since then, other federal, state, and local statutes have
provided rights to some persons in these excluded groups.

As Congress reviews the extent that American workers have bargaining
rights, you asked us to determine and assess (1) how many workers have
statutory collective bargaining rights 2 in the current civilian U. S.
workforce, (2) the types and numbers of workers without such rights, (3)
how the extent of collective bargaining rights among the workforce may
have changed during the past 40 years and (4) the potential impact of

1 29 U. S. C. 151 et seq.

2 For this report, we consider statutes as providing *collective
bargaining rights* if they not only permit individuals to join together
and form unions, but also require employers to recognize employee
organizations and to *bargain in good faith.* Note that having collective
bargaining rights does not necessarily imply the exercise of those rights
through union membership or other forms of collective action. Although not
statutes, in several instances we included workers who received rights
through gubernatorial executive orders.

United States General Accounting Office Washington, DC 20548

Page 2 GAO- 02- 835 Collective Bargaining

two recent Supreme Court decisions* the Kentucky River and Hoffman Plastic
cases 3 *on the types and numbers of workers without collective bargaining
rights.

To provide you with this information, we reviewed the NLRA of 1935 and
subsequent amendments and Supreme Court decisions and National Labor
Relations Board (Board) cases with regard to the scope of the act*s
coverage. We also identified and reviewed those other federal, state, or
local statutes that provide collective bargaining rights to employees. We
met with staff, Board members, and the General Counsel of the Board and
outside organizations and experts to assist in our review. We then
developed a methodology, using data from the February 2001 Current
Population Survey (CPS) Supplement collected by the Bureau of the Census
and from other data sets, to construct a quantitative estimate of the
percentage of the labor force that currently has statutory collective
bargaining rights. Our work was conducted between November 2001 and June
2002 in accordance with generally accepted government auditing standards.

We estimate that about three- quarters of the civilian workforce* or about
103 million of the approximate 135 million people in the labor force as of
February 2001* had some form of collective bargaining rights from federal,
state, or local statutes. Among the 115 million private sector workers,
about 78 percent had bargaining rights, mostly from coverage under the
NLRA. Coverage varied among industries, being the highest (90 percent) for
the 20.4 million workers in the manufacturing sector. In general, coverage
in the private sector was higher than that in the public sector, where
about 66 percent of 20 million government workers had some form of
collective bargaining rights, most often under state or local statutes.

In contrast, about 32 million civilian workers were without collective
bargaining rights under any law, either federal or state. The largest
groups without rights were about 8.5 million independent contractors; 5.5
million employees of certain small businesses; 10.2 million supervisory/
managerial employees (including 8.6 million first- line supervisors); 6.9
million federal,

3 National Labor Relations Board v. Kentucky River Community Care, Inc.,
532 U. S. 706 (2001) and Hoffman Plastic Compounds, Inc. v. National Labor
Relations Board, 122 S. Ct. 1275 (2002). Results in Brief

Page 3 GAO- 02- 835 Collective Bargaining

state and local government workers; approximately 532,000 domestic
workers; and 357,000 agricultural workers.

Our analysis of available data suggests that the proportion of the total
labor force with collective bargaining rights has likely increased since
1959. Since 1959, no major group of workers has lost bargaining rights
under the NLRA. However, other federal, state, and local laws have
extended rights to some workers in the groups excluded from the NLRA,
providing bargaining rights to about 14.5 million workers, primarily
nonprofit health care workers; federal, state, and local government
workers; and agricultural workers. In addition, because of inflation,
there has likely been a decline in the proportion of the labor force
employed in small businesses with annual dollar sales volumes too small to
be under the Board*s jurisdiction. While other excluded groups like
managers and supervisors may have increased in size, it is unlikely that
their growth was sufficiently large to offset the gains in the number of
workers obtaining rights. 4

Under two recent Supreme Court cases affecting Board decisions, some
workers currently with bargaining rights may either lose bargaining rights
or have their rights diminished. In the Kentucky River decision, the
Supreme Court ruled that the Board should revise its test for determining
whether a worker is a supervisor, an excluded group under the NLRA,
finding that the Board*s test served to categorically include certain
employees as covered under the act. Because any future tests used by the
Board to determine whether or not employees are supervisors should be less
categorical and more case- specific, the Kentucky River decision could
increase the number of employees considered supervisory and thus excluded
from coverage under the act. The Board has not yet devised alternative
tests in response to the decision, and given the case- by- case
determination required by the Court, we are unable to estimate the number
of employees that could potentially be deemed supervisors as a result of
this decision. In the case of Hoffman Plastic, the Court reversed the
Board*s decision to award back pay to an undocumented alien worker who was
fired for union activity. While the Court did not exclude undocumented
alien workers from protection under the NLRA, per se, it prohibited the
Board from awarding back pay to these undocumented alien workers whose
rights had been violated, stating that this remedy would conflict with
federal immigration law. Since back pay is one of the

4 For specifics on our methodology, see app. I.

Page 4 GAO- 02- 835 Collective Bargaining

major remedies available to workers for a violation of their rights, the
Court*s decision effectively diminishes the bargaining rights of such
workers under the NLRA. Undocumented alien workers potentially affected by
the Hoffman decision are estimated to number about 5.5 million.

The NLRA, enacted in 1935, is the cornerstone of labor relations and
collective bargaining in the United States, providing the basic framework
governing private sector labor- management relations. It provides
employees the right to form unions and bargain collectively and requires
employers to recognize employee unions that demonstrate support from a
majority of employees and to bargain in good faith. The act includes the
payment of back pay and the reinstatement of employment as remedies for
certain violations of the act. 5 It also created an agency called the
National Labor Relations Board (Board) to administer and enforce the act.
Subsequently, in 1947, the Taft- Hartley Act 6 and the Landrum- Griffin
Act in 1959 7 amended the NLRA, among other things, to clarify coverage
under the NLRA, prohibit certain union activities, and set requirements on
union governance. Together, these statutes established the basic policies
and procedures under which most private sector collective bargaining still
operates today.

Although the NLRA applies in general to all *employees,* certain groups of
workers are excluded from its provisions, either by express statutory
language in the original act and its amendments or by Board or judicial
interpretation. Among the groups of workers excluded from the act are (1)
supervisors and managers; (2) independent contractors; (3) employees of
certain small businesses; (4) domestic workers; (5) agricultural workers;
and (6) federal, state, and local government employees. 8 These

5 Back pay is monetary compensation, including interest, for the wages
lost because of the violation. 6 Labor- Management Relations Act, ch. 120,
61 Stat. 136 (1947).

7 Labor- Management Reporting and Disclosure Act of 1959, P. L. 86- 257
(1959). 8 Other groups of excluded employees include workers who are
employees of a parent or a spouse, U. S. employees of international
organizations like the World Bank, confidential assistants of labor
relations managers, and lay teachers at religious schools. Because we were
unable to accurately estimate the numbers of these workers in the total
workforce using a national database, we have not included these groups in
our estimates. However, we believe that the numbers of workers in these
groups to be small, particularly in relation to the total workforce, and
thus would not have a major impact on our overall estimates. Background

Page 5 GAO- 02- 835 Collective Bargaining

categories of excluded workers have been defined through various tests,
which the Board can apply to determine coverage of the act. For example,
the act sets forth a three- part test for determining supervisory status.
Employees are statutory supervisors if (1) they hold the authority to
engage in any 1 of 12 listed supervisory functions* hire, transfer,
suspend, lay off, recall, promote, discharge, assign, reward, or
discipline other employees or responsibility to direct them, to adjust
their grievances, or effectively to recommend such action; (2) their
exercise of such authority is not of a merely routine or clerical nature,
but requires the use of independent judgment; and (3) their authority is
held in the interest of the employer.

While the NLRA underpins much collective bargaining activity in the United
States, other federal, state, and local statutes also provide bargaining
rights to many individuals excluded from the NLRA, particularly government
workers and agricultural workers. The bargaining rights and procedures
under these statutes generally differ from those under the NLRA, with
variation in important areas such as what may be bargained for, and how
disputes are handled. For example, many state laws covering public
employees prohibit the right to strike, while the NLRA does not.

We estimate that about 103 million workers had collective bargaining
rights in their primary job in February 2001. 9 These workers constituted
about 77 percent of the 135 million people in the civilian workforce. 10
The percentage of workers with bargaining rights varied among industries
and between the private and public sector. As certain workers are not
covered under the NLRA but rather under various other federal, state, and
local statutes, not all workers* particularly agricultural workers and
public sector workers* have rights equivalent to those under the NLRA.

9 This estimate is for the entire workforce and includes workers who were
both the employed and unemployed. For those workers who were unemployed,
we determined whether they had rights or not on the basis of their most
recent primary job.

10 In February 2001, the total workforce of full- time and part- time
workers included about 141 million people aged 16 and above. From this
total, we subtracted about 5.8 million people we classified as *self
employeds** those self- employed who were not independent contractors*
resulting in a workforce of about 135 million people. (See app. I.) Three-
Quarters of

U. S. Workers Have Collective Bargaining Rights

Page 6 GAO- 02- 835 Collective Bargaining

Nearly an estimated 90 million private sector workers had collective
bargaining rights, about 78 percent of all persons who worked for private
employers. 11 (See fig. 1.) The percentage of covered private sector
workers varied by industry, with certain industries having coverage below
70 percent* agriculture/ forestry, construction, and finance/ real estate.
Coverage also varied between the private and public sectors, with overall
coverage among government workers (about 66 percent of 20 million workers)
markedly lower than that among all private sector workers. Figure 2
compares the number of workers with bargaining rights with the total
number of workers, by industry.

Figure 1: Private Sector Collective Bargaining Coverage and Key Groups
with No Rights, February 2001

Source: GAO*s analysis of February 2001 CPS Supplement, the 1997 Economic
Census, and the 1997 Bureau of Labor Statistics National Compensation
Survey.

11 An alternative methodology to counting the number of workers with
rights would be to count the number of jobs covered by these laws. When we
compared our CPS estimates with those from the job- based Current
Employment Statistics survey, we found little difference in the percentage
of the nonagricultural labor force with collective bargaining rights. (See
app. I.) Private Sector Coverage

Varies by Industry while Public Sector Coverage Varies by State

Excluded groups

22% Bargaining

coverage 78%

Independent contractors 7%

Small business employees 5%

Supervisors 9%

Domestic workers 0.3% Agricultural workers 0.5%

Page 7 GAO- 02- 835 Collective Bargaining

Figure 2: Comparison of Number of Workers with Collective Bargaining
Rights with Total Number of Workers, by Industry, February 2001

Source: GAO*s analysis of February 2001 CPS Supplement.

Coverage varied largely with the proportion of excluded workers across
each industry. However, in general, the industries with the lowest
coverage rates were also those with a greater proportion of independent
contractors. (See fig. 3.)

Total labor force in industry (in millions) 0 5

10 15

20 25

30 35

40 45

Agriculture/ Fish forestry Mining

Construction Manufacturing

Trans/ Utility Wholesale trade

Retail trade Real estate

Services Governmentemployees Finance/

Number of workers without collective bargaining rights Number of workers
with collective bargaining rights

Page 8 GAO- 02- 835 Collective Bargaining

Figure 3: Percent of Work Force That Has Collective Bargaining Rights and
Percent Who Are Independent Contractors by Private Sector Industry,
February 2001

Source: GAO*s analysis of February 2001 CPS Supplement and the 1997
National Compensation Survey.

Among public sector workers, uneven coverage among states led to the
relatively low percentage of workers with bargaining rights. While 26
states and the District of Columbia have laws that provide collective
bargaining rights to essentially all public employees, 12 12 states
essentially do not have any laws for collective bargaining among state and
local

12 The 26 states are Alaska, California, Connecticut, Delaware, Florida,
Hawaii, Illinois, Iowa, Maine, Massachusetts, Michigan, Minnesota,
Montana, Nebraska, New Hampshire, New Jersey, New York, Ohio, Oregon,
Pennsylvania, Rhode Island, South Dakota, Vermont, Washington, and
Wisconsin. Delaware has an employer *opt- in* provision for local
government employees in cities and towns with fewer than 100 employees. As
with the NLRA, the state laws that provide collective bargaining rights to
public employees often exclude various groups of employees (e. g., many
states expressly exclude management officials) from coverage.

Percent of industry 0 20

40 60

80 100

Work force that has collective bargaining rights Industry workforce that
are independent contractors

Agriculture/ Fish forestry Mining

Construction Manufacturing

Trans/ Utility Wholesale trade

Retail trade Real estate

Services Finance/

Page 9 GAO- 02- 835 Collective Bargaining

employees. 13 The remaining 12 states have laws that provide bargaining
rights to specific groups of workers (e. g. state workers, teachers, or
firefighters) but not to all state and local government workers. 14

Among workers with bargaining rights, most are covered by the federal
NLRA. However, other federal, state, or local statutes cover some private
sector workers as well as those public workers who currently have rights.
While some of these statutes provide broader rights than those available
under the NLRA, other statutes offer substantially fewer rights than those
provided by the NLRA. Although we did not attempt to either catalog or
assess the other bargaining statutes, we did outline some of their key
differences.

Within the private sector, airline and railroad employees have somewhat
different collective bargaining rights than other workers. Airline and
railroad employees are covered by the federal Railway Labor Act, 15 a law
that predates the NLRA by 9 years. Among other differences, the Railway
Labor Act applies to *all subordinate officials,* thereby permitting many
supervisory employees to join in collective bargaining.

While agricultural workers are excluded from the NLRA, nine states provide
these workers with bargaining rights. However, these statutes vary from
the NLRA*s provisions. For example, in Arizona the state agricultural
statute has different standards for bargaining unit determination than the
NLRA. California also has different standards for bargaining unit

13 Twelve states do not have collective bargaining laws for public
employees. They are Alabama, Arizona, Arkansas, Colorado, Louisiana,
Mississippi, New Mexico, North Carolina, South Carolina, Texas, Virginia,
and West Virginia. In addition, Texas prohibits collective bargaining for
most groups of public employees. However, firefighters and police may
bargain in jurisdictions with approval from a majority of voters.

14 These states are Georgia, Indiana, Idaho, Kansas, Kentucky, Maryland,
Missouri, Nevada, North Dakota, Oklahoma, Tennessee, and Wyoming. Three of
these states, Indiana, Kentucky and Missouri, extend collective bargaining
rights to certain public employees through an executive order from the
governor. Many public employees may be covered by local laws, for example,
in Maryland they do not have a comprehensive law covering all public
employees. All state employees are covered under state labor laws, but
state statutes cover local employees only in certain counties. Local
governments in Maryland may have their own ordinances giving local public
employees collective bargaining rights, but these ordinances do not exist
in every county.

15 45 U. S. C. 151- 188. Employees covered by the Railway Labor Act are
excluded from the provisions of the NLRA. 29 U. S. C. 152( 3). Bargaining
Rights

Conferred under Other Laws Are Not Identical to Those under the NLRA

Page 10 GAO- 02- 835 Collective Bargaining

determination than the NLRA and allows agricultural workers to use
secondary boycotts, 16 actions prohibited under the NLRA.

Within the public sector, where federal, state, and local government
workers rely on an array of federal, state, and local statutes for
bargaining rights, most workers are prohibited from striking. Although
some states allow certain workers the right to strike for at least some
employees, 17 many states 18 provide compulsory binding interest
arbitration (a procedure unavailable under the NLRA that requires an
arbitrator to resolve differences in the event of an unsettled dispute
between the union representing the employees and the employer).

We estimate that about 32 million workers currently do not have any
collective bargaining rights. These workers include over 25 million
private sector workers* 8.5 million independent contractors, 5.5 million
employees of certain small businesses, 10.2 million managers and
supervisors (including 8.6 million first- line supervisors), 532,000
domestic workers, 19 and 357,000 agricultural workers. Those groups
without rights also include over 6.9 million federal, state, and local
government employees. In general, these workers do not have bargaining
rights under any federal or state statute. 20

16 A secondary boycott is an organized refusal to purchase the products
of, do business with, or perform services for (such as deliver goods) a
company that is doing business with another company where the employees
are on strike or in a labor dispute.

17 According to the American Federation of State, County, and Municipal
Employees, a union that represents 1.3 million public employees, 11 states
provide the right to strike to some public employees.

18 Twenty- one states and the District of Columbia have such provisions.
In many cases, mandatory binding arbitration serves as a *quid pro quo*
for the prohibition of strikes by public employees.

19 These 532,000 domestic workers do not include an additional 126, 000
domestic workers who classify themselves as independent contractors. 20 It
is likely that some of these workers, particularly the employees of
certain small businesses, have bargaining rights under state labor
relations statutes. Eighteen states have state labor relations acts or
*little Wagner Acts,* modeled after the federal NLRA, and although many
exclude the same groups as under the NLRA, some do not expressly exclude
agricultural workers, supervisors, or independent contractors and only one
excludes employees of small businesses. (See app. I.) Except for
agricultural workers, we did not have sufficient information to determine
the number of workers affected by these statutes. Over 32 Million

Workers Do Not Have Collective Bargaining Rights

Page 11 GAO- 02- 835 Collective Bargaining

About 7 percent (8.5 million) of all private sector workers classified
themselves as independent contractors 21 in February 2001. The percentage
of independent contractors in private sector industries varied widely. For
example, for the construction sector this figure was 19 percent and for
the manufacturing sector it was 1.5 percent. (See fig. 3.) Although the
NLRA specified that its provisions applied to employees, 22 it was not
until the 1947 Taft- Hartley amendments that *independent contractors*
were expressly excluded from the act. Traditionally, the Board and the
Supreme Court have looked at the factual circumstances underlying the
relationship between the worker and the employer to determine whether the
worker is an independent contractor or an employee. 23

Although the NLRA provides the Board with broad jurisdiction, the Board*s
jurisdiction does not extend to retail employers with annual dollar sales
volumes less than $500,000 and many non- retail employers with sales
volumes less than $50,000. Since the standards were modified in 1958, the
annual dollar sales volumes in the standards have remained unchanged. 24

21 Using the February 2001CPS Supplement, we first subtracted the sole
proprietors or selfemployeds* those people identifying themselves as self-
employed but not an independent contractor* from the labor force. For our
group of independent contractors, we counted all persons who identified
themselves as self- employed and as independent contractors or who were
wage/ salary workers and independent contractors. Independent contractors
are defined as workers who typically have control and judgment over how
contracted services are performed and, therefore, they decide the time and
place the work is to be done, supply their own necessary tools and
instruments, and have a risk of financial profit or loss. They also
typically perform work that requires a particular skill not ordinarily
used in the course of business.

22 However, according to the act, this shall not include any individual
employed as an agricultural laborer or in the domestic service of any
family or person in his or her home or in any individual employed by his
or her parent or spouse.

23 To do this, the Board has considered factors such as a worker*s
autonomy to perform work and where the work is being performed. See
National Labor Relations Board v. United Insurance Co. of America, 390 U.
S. 254 (1968).

24 Since 1950, the Board has held that its jurisdiction does not extend to
labor disputes involving certain small businesses. Requiring a *pronounced
impact* on interstate commerce before it asserts federal jurisdiction, the
Board set minimum standards for specific types of employers, generally
stated in terms of a yearly sales volume. In 1959, however, the Landrum-
Griffin amendments to the NLRA upheld the Board*s interpretation of its
jurisdictional thresholds, but limited such exclusions to the standards in
effect as of August 1, 1959. Our analysis extended only to those
industries with a specified dollar sales volume. (See app. I.) About 7
Percent of Private

Workforce Were Independent Contractors, Not Employees

About 5 Percent of Private Sector Workers Employed by Firms Too Small for
Board Jurisdiction

Page 12 GAO- 02- 835 Collective Bargaining

We estimate in February 2001 that about 5 percent (5.5 million) of private
sector workers were employed in firms with dollar sales volumes below the
Board*s jurisdictional reach. For most industries, the percentage of
employees excluded under these standards is very small* less than 1
percent for mining, construction, and manufacturing. However for retail
trade, the $500,000 sales limit excludes about 12 percent of all workers
in that industry (about 2.6 million workers). The standards also exclude
11 percent of the workers in the finance/ real estate sector (mainly real
estate brokers) and 4 percent of service workers. (See app. I.)

We estimate that about 10.2 million 25 (including 8.6 million first- line
supervisors) of 115 million private sector workers were either first- line
supervisors or higher- level managers in February 2001. 26 This group
comprised about 9 percent of the total private sector labor force.
Although the original NLRA, as enacted in 1935, was silent on the
exclusion of supervisors and management officials, early Board rulings
interpreted the NLRA as excluding managers from its protections. However,
whether supervisors* particularly the first- line supervisors (or the so-
called *foremen*)* should be excluded as managerial officials was much
debated in the years following the enactment of the NLRA in 1935. 27 In
1947, the Taft- Hartley amendments to the NLRA specifically excluded
supervisors from coverage of the NLRA. 28 The Board and the courts
continue to interpret the act as excluding managers.

25 The 10.2 million estimate is based on 1997 NCS calculations on full-
time employees. To the extent that part- time employees are also
supervisors, the total number of employees in this group is
underestimated. In addition, some managers and supervisors had already
been excluded when we corrected for employees in those firms deemed too
small for Board jurisdiction. We do not include them here to avoid double
counting. (See app. I.)

26 In keeping with the 1997 NCS definitions, we define first- line
supervisor as one who directs staff through face- to- face meetings and
where performing the same work as subordinates is not the principal duty.
Consistent with the NCS, we define second line supervisors as one who
directs staff through intermediate supervisors. See James Smith,

Supervisory Duties and the National Compensation Survey, in Compensation
and Working Conditions, Spring 2000, and appendix I.

27 The Supreme Court*s holding in Packard Motor Car Co. v. NLRB, 330 U. S.
485 (1947) that general foremen (supervisors and managers) were entitled
as a class to bargaining rights under the NLRA served to spur the
subsequent enactment of Taft- Hartley, which specifically excluded
supervisors from coverage.

28 29 U. S. C. 152( 11). Supervisors and Managers

Are the Largest Group of Employees without Collective Bargaining Rights

Page 13 GAO- 02- 835 Collective Bargaining

Domestic workers** domestic servants* employed by any homeowner or
resident of the home in which the services are rendered* have been
expressly excluded from NLRA since 1935. Moreover, most if not all states
that enacted labor relations acts similar to the NLRA expressly excluded
domestic workers. 29 In February 2001, there were about 658,000 domestic
workers in private households. However, about 20 percent of these workers
were classified as wage employees working as independent contractors.
Therefore, in our analysis, these workers were counted as independent
contractors, while the remaining 532,000 were considered domestic workers.

In 1935, the NLRA excluded any individual employed as an *agricultural
laborer.* 30 Due to the federal exclusion of agricultural laborers, nine
states have either enacted a separate agricultural labor relations act
(Arizona, California, Oregon, and Maine) or did not expressly exclude
agricultural workers in a state labor relations act (Hawaii,
Massachusetts, New Jersey, and Wisconsin), with one state (Kansas) in both
categories. Thus, agricultural workers have obtained some collective
bargaining rights under these state statutes. In February 2001, we
estimate about 680,000 agricultural workers whose primary job included
*any task associated with running or manufacturing a farm.* 31
Approximately 357,000 of these workers resided in a state that had some
bargaining rights for agricultural workers. Thus, about one- half of all
agricultural workers were without federal or state bargaining rights.

29 However, in California, the state labor code provides workers the right
to organize into unions, and it has been held to apply to employees in a
private household. See Annenberg v. Southern California District Council
of Laborers and its Affiliated Local 1184, 38 Cal. App. 3d 637 (1974). We
did not incorporate this or other state exceptions into our domestic
worker estimate.

30 Subsequently, the definition of agricultural laborer was expanded to
include not only crop workers but also workers who work on a farm in
practices incidental to farming, such as packing produce for shipment from
the farm. Because the number of workers matching the secondary definition
is not available using national survey data, we limited our analysis of
laborers to agricultural workers who work with crops and livestock. Thus,
we may be overestimating the percentage of agricultural workers who have
collective bargaining rights.

31 Our data are based on February 2001 employment levels. For most
agricultural commodities, February represents the annual employment
*trough.* Employment in agriculture fluctuates significantly over the year
and, in 2001, had an annual monthly average of about 745, 000 individuals
employed as agricultural laborers. Over One- Half Million

Employees Excluded as Domestic Workers

About 357,000 Agricultural Workers Excluded from Bargaining Rights

Page 14 GAO- 02- 835 Collective Bargaining

We estimate that about 6.9 million federal, state, and local government
workers out of about 20 million government workers did not have collective
bargaining rights as of February 2001. These include about 900,000 federal
employees who are not eligible to collectively bargain either because they
are considered management employees or because they work for a noncovered
agency. 32 They also include state and local public employees in the 12
states without bargaining rights, and those in the 12 states where only
some of the employees have rights. The other 13 million workers have some
collective bargaining rights.

Federal, state, and local government workers were excluded from the NLRA
in 1935. For many federal employees, the Federal Service Labor Management
Relations Statute 33 provides collective bargaining rights, specifically
for employees of most executive branch agencies as well as employees of
the Library of Congress and the Government Printing Office. In addition,
all but 12 states have some collective bargaining rights for at least one
group of government worker, and 26 states as well as the District of
Columbia have collective bargaining procedures that cover most public
employees.

Our analysis of available data suggests that the proportion of the labor
force with statutory collective bargaining rights has likely increased
since 1959. Over this period, no major groups of workers have lost
collective bargaining rights, and about 14.5 million workers have gained
bargaining rights through the enactment of federal, state, and local laws.
In addition, the proportion of the labor force employed in small
businesses that are excluded from NLRA*s jurisdiction has likely declined,
because the standard for determining NLRA*s jurisdiction dollar sales
volume has not been corrected for inflation. While some excluded groups
like independent contractors, managers, and supervisors may have increased
in size, it is

32 Among the major agencies not covered include: the Federal Bureau of
Investigation, the Central Intelligence Agency, the National Security
Agency, the Tennessee Valley Authority, and the U. S. Secret Service. 5 U.
S. C. 7103( a)( 3).

33 5 U. S. C. 7101. About 6.9 Million Public

Employees Do Not Have Bargaining Rights

Percentage of Total Labor Force with Bargaining Rights Likely Increased
Since 1959

Page 15 GAO- 02- 835 Collective Bargaining

unlikely that their growth was sufficient to offset the gains in the
number of workers obtaining rights. 34

The major categories of workers excluded from the NLRA* managers and
supervisors, independent contractors, employees of small businesses
affected by the Board*s jurisdiction standards, agricultural workers,
domestic workers, government workers, and workers covered by the Railway
Labor Act* were all established by 1959. The original 1935 act excluded
agricultural workers, domestic workers, government workers, and workers
covered by the Railway Labor Act. Subsequently, supervisors and
independent contractors were excluded in 1947, and between 1950 and 1959,
the Board set size standards for determining which small businesses would
be under its jurisdiction based on annual dollar sales volume.

Since 1959, federal, state, and local laws have extended rights to some of
the groups excluded under the NLRA*s coverage. Federal laws allowed many
federal workers bargaining rights, and added nonprofit health care workers
to those covered by the NLRA. State and local statutes provided bargaining
rights to many state and local government workers, as well as to many
agricultural workers. We estimate that about 14.5 million workers in
today*s workforce have bargaining rights as a result of the federal,
state, and local laws enacted since 1959. (See table 1.)

34 We chose 1959 because it was the last year when major comprehensive
amendments were made to the NLRA and when the Board jurisdiction standards
were last revised. However, it may be that this analysis is sensitive to
the comparison dates used. As one expert has pointed out, while the
percentage of workers with bargaining rights has likely increased between
1959 and 2001, this might not be true, for example, for the period 1981 to
2001. By 1981, nonprofit health care workers and some agricultural workers
had already gained rights, and major increases in public sector rights had
also occurred. This would leave less growth to balance out any increase in
excluded groups like independent contractors or supervisors. The early
1980s are considered an important period by some industrial experts
because it heralded the economic restructuring of a number of key
industries and a shift in the industrial relations policies of many
employers. However, data were not available to address this concern. About
14.5 Million Workers

Gained Bargaining Rights Since Last Major Exclusion in 1959

Page 16 GAO- 02- 835 Collective Bargaining

Table 1: Number of Workers Receiving Collective Bargaining Rights from
1959 through February 2001

a A few states did confer collective bargaining rights to certain groups
of public employees prior to 1959. b Postal Reorganization Act, 39 U. S.
C. 1001, provided collective bargaining rights to postal workers.

Source: GAO*s analysis, analysis of the February 2001 CPS Supplement and
Draft Comprehensive Transformation Plan, U. S. Postal Service, October 1,
2001.

Our analysis of available data on the key groups of employees excluded
under the NLRA suggests that the extent of collective bargaining rights
among the total civilian U. S. workforce has increased since 1959. The
number and proportion of public employees and supervisors with bargaining
rights increased over this period, and the effects of inflation on the
Board*s small business jurisdictional exclusions have likely reduced the
number of small business employees excluded from the act. 35 The
proportion of the labor force with bargaining rights has increased even
with the conservative assumptions that the percentage of the labor force
in businesses excluded from the Board*s jurisdiction has not changed since
1959, not correcting for the extension of bargaining rights to nonprofit
health care workers. It also assumes that the proportion of the labor
force excluded as independent contractors, managers, and supervisors had
more than doubled over this period. (See table 2.)

35 Regarding trends in private sector bargaining rights alone, increases
in workers with rights in agriculture and health care and the effects of
inflation on the Board*s small business jurisdiction standards must be
counterbalanced by increases in the percentage of the workforce that are
independent contractors, supervisors, or managers. Data limitations,
however, preclude a definitive answer.

Years bargaining rights received Worker group affected Number of

workers

1959- 2001 State and local government workers About 10,000,000 a 1978
Federal workers 1, 500,000 1970 b Postal workers 700,000 1972- 1997
Agricultural workers 360,000 1974 Nonprofit health care workers 1, 900,000

Total About 14,500,000

Net Increase Likely since 1959, Even with Growth in Some Major Excluded
Groups

Page 17 GAO- 02- 835 Collective Bargaining

Table 2: Net Effects of Changes in Excluded Groups on Percent of the Labor
Force That Has Bargaining Rights, 1959 and 2001

Excluded group Trend Approximate

percentage point change in labor force with bargaining rights

since 1959

Public employees About 12.2 million public employees have gained
bargaining rights since 1959. a +9 Employees of small businesses excluded
under Board jurisdiction standards

The number of workers affected by the Board's small business jurisdiction
exclusion was about 4 percent of the total civilian labor force in 2001.
The annual sales volumes required by the minimum standards for the Board*s
jurisdiction over a business have not been adjusted for inflation since
1959. For example, the current Board standard for retail employers of
$500,000 annual sales would be about $3 million in 2001, if adjusted for
inflation. In 2001, about one- third of all retail employees (over 7
million) worked for firms with annual sales under $3 million, while only
about 12 percent (about 2.6 million) work for firms with annual sales less
than $500,000. b We assume no change in the percentage of the labor force
excluded under these standards between 1959 and 2001, a conservative
assumption.

0 Independent contractors Comparable data on independent contractors for
1959 and 2001 are not

available. c We assume, conservatively, that the 6.5 percent of the 2001
total civilian labor force who were independent contractors had more than
doubled since 1959, increasing 3.5 percentage points.

-3.5 Managers and supervisors Although some evidence suggests that private
sector managers and

supervisors may have grown in number and as a share of the total civilian
labor force over this period, little data are available on these groups
for 1959 and 2001. d We assume that the 7.5 percent of the 2001 total
civilian labor force who were managers and supervisors had more than
doubled since 1959, increasing 4 percentage points.

-4 Agricultural workers and domestic workers Domestic workers were a small
proportion of the labor force in 1959 and 2001.

Virtually no agricultural workers had statutory bargaining rights in 1959
while about half of those workers had such rights in 2001. This would
increase the percentage of the total labor force in 2001 that had
bargaining rights. We assume, conservatively, that there was no change in
the relative size of either group as a percent of the labor force since
1959.

0 Net effect since 1959 +1. 5

a A few states did confer bargaining rights to certain groups of public
employees prior to 1959. Data are unavailable to estimate the number of
public employees in these states with rights prior to 1959. The percentage
of the nonfarm civilian labor force that the public employees in these
states comprised in 2001 was 0.5 percent, and we apply this same
percentage to 1959. b It is possible that industry shifts in the economy
towards sectors where businesses are typically of

smaller size could have led to a relative increase in the number of
employees below the dollar limit and thus an increase in the number of
employees without rights. Although, on balance, the sixfold increase in
inflation since 1959 more than likely resulted in a net increase in the
number of employees in smaller businesses with bargaining rights; we
assumed no change in the relative size of this exclusion between 1959 and
2001. c However, data on independent contractors suggests that this group
exhibited no growth between

1995 and 2001. U. S. General Accounting Office, Contingent Workers: Income
and Benefits Lag Behind Those in Rest of Workforce, GAO- 00- 76, June 30,
2000, Washington, D. C.

Page 18 GAO- 02- 835 Collective Bargaining

d Available evidence suggests that while managers and supervisors have
increased significantly as a percentage of the labor force, they have not
increased at the levels necessary to reduce the percent of the labor force
that has collective bargaining rights. A recent Department of Labor study
found that the percentage of supervisors in the services sector increased
from 8.6 to 12.8 percent between 1969 and 1999, about a 67 percent
increase. In retail trade, the percentage of supervisors increased from
7.5 to 12.0 percent between 1959 and 1999, a 63 percent increase. See U.
S. Department of Labor,

The *New Economy* and Its Impact on Executive, Administrative and
Professional Exemptions to the Fair Labor Standards Act (January 2001).

Source: GAO*s analysis of Bureau of Labor Statistics National Employment,
Hours and Earnings data, 1959 and 2001.

Under two recent Supreme Court rulings, some workers currently with
bargaining rights may either lose bargaining rights or have their rights
diminished. The Kentucky River ruling affected the test that the Board
uses to determine supervisory status, a status that can determine coverage
of the act and, therefore, bargaining rights. The Court ruled that the
Board*s test was inconsistent with the NLRA in that it introduced a
categorical restriction on the term *independent judgment,* a key concept
in the statutory definition of a supervisor. Because any future tests used
by the Board to determine whether or not employees are supervisors should
be less categorical and more fact- specific, the Kentucky River decision
could have the effect of increasing the number of employees considered
supervisory and thus excluded from coverage under the act. Given that the
Board has not yet devised alternative tests in response to the opinion, we
are unable to estimate the number of employees who could potentially be
deemed supervisors as a result of this decision. In the Hoffman Plastic

case, the Supreme Court ruled that undocumented alien workers were not
eligible for back pay under the NLRA. This case did not exclude
undocumented alien workers from coverage of the NLRA; however, because
back pay is one of the major remedies available for a violation, this
decision diminished the legal bargaining rights available to these workers
under the act. We estimate that this group of workers potentially affected
by this decision numbers about 5.5 million.

In the Kentucky River case, the Court ruled that the Board should revise
its test for determining supervisory status, a status that can determine
coverage under the act and, therefore, bargaining rights, for certain
groups of employees, in this instance, charge nurses. Specifically, the
case focused on the portion of this test that examines whether the nurses*
exercise of supervisory authority is not merely routine, but requires the
use of *independent judgment.* A previous line of Board decisions stated
that employees did not use independent judgment when they exercised
*ordinary professional or technical judgment in directing less- skilled
employees to deliver services in accordance with employer- specified Two
Recent Supreme

Court Cases Could Affect Gains in Coverage over the Last 40 Years

Kentucky River Requires a Determination of Supervisory Status on a Case-
by- Case Basis

Page 19 GAO- 02- 835 Collective Bargaining

standards* and thus were not supervisors excluded by the act*s coverage.
The Board reasoned, in this line of cases, that judgment informed by
professional or technical training or experience is not *independent,* and
thus, the employee should not be considered to be a supervisor.

The Court rejected the Board*s test in determining supervisory status and
held that for employees involved, the question of supervisory status turns
on whether the employee is exercising independent judgment. This
determination, according to the Court, must be based solely on the degree
of judgment exercised by the employee, not the type of judgment, that is,
professional or technical. 36 For this group of employees, the Court would
not accept an automatic or categorical finding that they are not
supervisors; rather, on a case- by- case basis, the Board must focus
solely on the degree of independent judgment that the employee exercises.
37

Because any future tests used by the Board to determine whether or not
certain employees are supervisors should be less categorical and more
fact- dependent, the Kentucky River decision could have the effect of
increasing the number of employees considered supervisory and thus
excluded from coverage under the act. 38 However, given the uncertainty of
the Board*s response to the opinion and the fact- intensive nature of the
legal tests involved, we are unable to estimate the number of employees
who could potentially be deemed supervisors as a result of the Kentucky
River case.

36 While the Court clearly rejected the Board*s approach in these cases,
the Court pointed out that it falls within the Board*s discretion to
determine, within reason, what scope of discretion qualifies as
independent so as to deem the employee a supervisor.

37 In considering the potential reach of this decision, it is important to
note that the decision requires a change in how the Board will determine
supervisory status for that group of employees who exercise ** ordinary
professional or technical judgment in directing less- skilled employees to
deliver services in accordance with employer- specified standards.*

38 In a recent case, the Board applied a revised test, in accordance with
Kentucky River, to determine supervisory status of certain tugboat pilots.
Reversing an earlier decision, the Board found that these pilots exercised
independent judgment in carrying out supervisory functions and thus were
supervisors under the act. See American Commercial Barge Line Company et
al., 337 N. L. R. B. 168 (2002).

Page 20 GAO- 02- 835 Collective Bargaining

The Supreme Court ruling in the Hoffman Plastic case reversed a Board
decision awarding back pay, a key remedy under the NLRA, to undocumented
alien workers. 39 The case involved a Board decision concerning an
employer*s termination of an undocumented alien worker to rid itself of
known union supporters, a violation of the act. The Board had ruled that
the employer should award back pay and conditional reinstatement 40 and,
in doing so, applied the protections and remedies of the act to
undocumented alien workers in the same manner as it does to other workers.
The Supreme Court reversed this decision and ruled that back pay remedies
cannot be awarded to an undocumented alien because such awards are in
conflict with the Immigration Reform and Control Act of 1986 (IRCA), which
makes it unlawful for employees to use fraudulent documents to establish
employment.

To remedy a violation of the NLRA, the Board had traditionally found that
unlawfully discharged employees were entitled to unconditional
reinstatement with back pay to make them whole for any losses they may
have suffered because of the unlawful layoffs. The Board maintained that
the congressional objective of deterring unauthorized immigration embodied
in IRCA and the labor law protections of the NLRA are complementary.
Specifically, according to the Board, coverage of undocumented workers by
the NLRA furthers IRCA*s goals by helping to both ensure reasonable
working conditions by decreasing competition from illegal aliens willing
to accept substandard wages and employment conditions and to eliminate
economic advantages* and thus incentives* to employers for hiring
undocumented workers in preference to American citizens or alien employees
working lawfully. The Board, therefore, concluded that awarding back pay
effectuates the policies of both IRCA and NLRA.

While the Court acknowledged that the NLRA vests the Board with broad
discretion in choosing an appropriate remedy where it has substantiated an
unfair labor practice, the Court concluded that awarding back pay to

39 An undocumented alien worker is a non- U. S. citizen present in the
United States in violation of the U. S. immigration laws, e. g., a person
who entered the United States without being inspected by an Immigration
officer or a non- U. S. citizen who entered legally but overstays or
violates his or her immigration status.

40 The Board stated that the employer*s reinstatement of the discharged
workers is dependent upon the employer demonstrating appropriate
documentation of eligibility for U. S. employment as required under the
Immigration Reform and Control Act of 1986. See

A. P. R. A. Fuel Oil Buyers Group, Inc., 320 N. L. R. B. 408 (1995). About
5.5 Million

Undocumented Alien Workers Have Rights Diminished

Page 21 GAO- 02- 835 Collective Bargaining

undocumented alien workers lies outside the limits of the Board*s
discretion. Although the Court*s decision reinforces the goals of federal
immigration policy, it diminishes the legal protections available to
undocumented alien immigrant workers under the NLRA. Back pay is an
important remedy under the act because it provides an incentive to report
law- violating behavior and imposes financial penalties on an employer and
job protection for the workers.

Adjusting year 2000 U. S. Census data on the foreign- born population, we
estimate a total of 8.7 million undocumented immigrants, of which 7.1
million are between the ages of 18 and 65. 41 Applying an estimate that 77
percent of private sector workers have collective bargaining rights under
the NLRA, we calculated that approximately 5.5 million unauthorized
immigrants are in the workforce and have collective bargaining rights
under the NLRA. 42 We were unable, however, to disaggregate this number by
industry.

As agreed with office, unless you publicly announce the contents of this
report earlier, we plan no further distribution until 30 days from the
report date. At that time we will send copies of this report to the
chairman, members and general counsel of the National Labor Relations
Board, the Secretary of Labor, and other interested parties. In addition,
the report will be available at no charge on the GAO Web site at http://
www. gao. gov.

41 This estimate also includes persons in *quasi- legal* status, persons
who cannot be conclusively determined as being in the country legally or
illegally. (See app. I.) 42 We derived our 77- percent estimate for NLRA
coverage from our total private sector coverage estimate of 78 percent and
correcting for coverage under the Railway Labor Act. (See app. I for
further details.)

Page 22 GAO- 02- 835 Collective Bargaining

If you or your staff should have any questions, please call me at (202)
512- 9889. The major contributors to this report are Charles A Jeszeck,
Assistant Director, who can be reached at (202) 512- 7036; Nancy Peters,
Analyst in Charge; Kara Kramer, Analyst; and Tom Beall, Paula Bonin, Mark
Ramage, and Joan Vogel.

Robert E. Robertson Director, Education, Workforce

and Income Security Issues

Appendix I: Methodology Page 23 GAO- 02- 835 Collective Bargaining

Our methodology consists of a review of key statutes, National Labor
Relations Board and U. S. Supreme Court decisions, a statistical analysis
of several national representative data sets, and discussions with Board
staff and officials and organizations that are active and familiar with
bargaining rights outside of the National Labor Relations Act of 1935
(NLRA). The organizing concept for the review of key statutes and the
statistical analysis was the NLRA (or the Wagner Act). We reviewed the
original act and subsequent amendments, and Supreme Court and Board
decisions to identify those segments of the labor force that were excluded
from coverage under the act. We then conducted a review of other relevant
federal, state, and local laws to identify those excluded workers who
obtained bargaining rights from other statutes. We consulted with Board
staff and representatives of several outside organizations for their
technical advice. Once we completed this review, we used a variety of
nationally representative data sets to construct a quantitative estimate.
Using an estimate of the workforce in February 2001, we subtracted those
groups of workers identified as excluded from the NLRA and then added back
those groups of workers who obtained bargaining rights from other laws.

The NLRA is the central law governing private sector labor management
relations in the United States. We reviewed the NLRA and subsequent
amendments to identify the key groups of employees who were excluded from
coverage under the act. We also reviewed pertinent Supreme Court and Board
decisions affecting the act*s scope of coverage. In particular we analyzed
Kentucky River and Hoffman Plastic Supreme Court decisions. We identified
and reviewed those other federal state and local laws that provided
certain bargaining rights to employees excluded from coverage under the
NLRA. These include the Railway Labor Act and the Federal Service Labor-
Management Relations Statute.

An important issue in this analysis is the definition of bargaining
rights. There is variation in the rights provided under the NLRA, the
Railway Labor Act and the many state and local laws governing collective
bargaining. These differences span a host of issues, from the procedures
governing representation elections, the right to strike, and binding
arbitration, to the scope of bargaining and the remedies for violations.
Although the right to strike could be considered part of a *core
definition of bargaining rights,* we based our definition on the concepts
of union recognition* permitting individuals to join together and form
unions and the requirement that employers recognize employee
organizations* and Appendix I: Methodology

Overall Approach Legal Review Definition of Bargaining Rights

Appendix I: Methodology Page 24 GAO- 02- 835 Collective Bargaining

*good faith bargaining** bargaining with intent to reach an agreement. 1
These are key elements of the rights granted under the NLRA. 2

Different definitions of collective bargaining rights would likely lead to
different empirical estimates of the percentage of the labor force that
has bargaining rights. However, because of the predominance of the NLRA
and the Railway Labor Act in the private sector, most of the differences
among laws occur in the public sector. Thus, for many alternative
definitions, for example, one based on the right to strike, the change in
the percent of the labor force with rights would be largely limited to
changes in the number of public employees who had this right.

To develop our quantitative estimates of the number of workers with
collective bargaining rights, we started with an estimate of the total
labor force and then subtracted those groups of workers we identified as
excluded from coverage under the NLRA. We then added back those groups of
excluded workers who obtained bargaining rights from other statutes.

For our quantitative estimates, we relied on the February 2001 Current
Population Survey (CPS) Supplement collected by the U. S. Bureau of the
Census. We used the CPS to estimate the total percentage of workers in the
civilian labor force with collective bargaining rights on their primary
jobs because it is a nationally representative dataset that also has
information available for independent contractors, a key group excluded
from coverage under the NLRA. Using the CPS, we adjusted our estimates for
six key groups excluded from coverage under the NLRA: (1) supervisors and
managers; (2) independent contractors; (3) employees of certain small
businesses; (4); domestic workers; (5) agricultural workers; and (6)
federal, state, and local government employees. We did not attempt to
construct estimates for the other groups excluded under the NLRA, for
example, religious organization employees and student employees, because
reliable data were not readily available. Because these

1 This definition would exclude laws that only provide that parties *meet
and confer.* 2 This report does not assess the adequacy of the rights
granted under the NLRA, in terms of fostering collective bargaining nor to
how it compares with any international standards on collective bargaining
or workers rights, nor the degree to which those rights are protected.
Workforce

Estimations and Analysis

Appendix I: Methodology Page 25 GAO- 02- 835 Collective Bargaining

other groups are small, we believe that their exclusion from our estimates
would have only a small effect. 3

Using CPS data, we first subtracted out sole proprietors or the
*selfemployeds* 4 *those persons identifying themselves as self- employed
but

not independent contractors* from our labor force estimates. Independent
contractors are all those who were identified as independent contractors,
consultants, and freelance workers in the supplement, regardless of
whether they were identified as wage and salary workers or self- employed
in the responses to basic CPS labor force status questions. For our group
of independent contractors, we added together those persons identifying
themselves as being self- employed and independent contractors and those
who identified themselves as wage/ salary workers and independent
contractors and subtracted this total from the total civilian labor force.
The CPS defines independent contractors as workers who typically have
control and judgment over how contracted services are performed and,
therefore, they decide the time and place the work is done, supply their
own necessary tools and instruments, and have a risk of financial profit
or loss. They also typically perform work that requires a particular skill
not ordinarily used in the course of business. About 88 percent of
independent contractors were identified as self- employed in the main
questionnaire, while 12 percent were identified as wage and salary
workers. Conversely, about half of the self- employed were identified as
independent contractors. Workers identified as self- employed in the basic
CPS are those workers who obtain customers on their own to provide a
product or service.

3 The excluded groups we did not estimate include religious organization
employees, employees of international organizations, children or spouses
employed by parents or spouses, and student employees. Not estimating
these excluded groups will tend to overstate the percentage of the labor
force with bargaining rights. Employees covered by the Railway Labor Act
and employees of nonprofit healthcare providers and educational
institutions and are included in our estimates.

4 The sequencing of our analysis was an important consideration so that we
could avoid potential double counting and potential overestimation of
workers in particular excluded groups. We first subtracted the self-
employeds and independent contractors out of the total labor force. We
then calculated the number of employees excluded by the Board*s
jurisdictional standards on business size, then managers and supervisors
and other groups. Self- employed Persons and

Independent Contractors

Appendix I: Methodology Page 26 GAO- 02- 835 Collective Bargaining

Although the Board has broad jurisdiction under the NLRA, it typically
declines jurisdiction over certain small employers with annual dollar
sales volumes less than a specified level as well as employers who meet
certain other criteria. (See table 3.) However, the CPS does not identify
workers by firm size, where size is determined by annual dollar sales
volume of the firm. To identify those employees who would be excluded
under the Board*s jurisdictional standards, we applied these sales volume
standards to 1997 data we obtained from the Small Business Administration
(SBA). SBA periodically requests that the Census calculate the number of
employees employed by firms with specific dollar sales volumes, by
industry classification. We chose the SBA data because this was the only
dataset available that categorically separates each industry by volume of
sales receipt. We chose data for 1997 because that was the most recent
data available that still used the Standard Industrial Classification
(SIC) code system of industry classification and thus would be comparable
with the classification used in the CPS. The SBA data identifies the
number of workers by firms* annual dollar sales volumes so that we were
able to calculate the number of employees with and without bargaining
rights according to the Board*s jurisdictional standards. We estimated the
exclusionary effect of the Board*s small business jurisdictional standards
by calculating a percent of employees in each industry who would be
employed by a firm with annual dollar sales below the Board*s standard and
thus exempt under the Board*s guidelines. We then applied these industry
exclusion percentages to the February 2001 CPS Supplement estimates of the
total number of workers in each industry classification, yielding a number
of excluded workers for each industry classification. 5 (See fig. 4.) Once
we estimated these excluded industry totals, we subtracted them from the
CPS labor force total. Summing all of the industry totals provides an
estimate of the total number of employees effectively excluded from NLRA
coverage by the Board*s jurisdictional

5 We applied the percentages derived from the 1997 SBA data to our 2001
CPS industry workforce estimates. To the extent that the distribution of
small businesses across industries has changed demonstrably over the time
interval, we may not have an accurate estimate of the number of workers
excluded by these standards in 2001. However, SBA officials believe that
the industrial distribution of small businesses changes very slowly over
time so this should not be a major problem. Employees of Small

Businesses with Annual Dollar Sales Volumes Below the Board Jurisdictional
Limit Standards

Appendix I: Methodology Page 27 GAO- 02- 835 Collective Bargaining

standards. 6 We did not attempt to estimate those Board jurisdictional
standards that did not specify an annual dollar sales volume.

Table 3: Minimum Annual Dollar Sales Volumes for Coverage under NLRA
Industry/ economic activity Minimum annual sales volume for NLRA
jurisdiction

Nonretail industries $50,000 Retail $500,000 Office buildings and shopping
centers $100,000 Public utilities $250,000 Newspapers $200,000 Radio and
TV stations $100,000 Hotels, motels, apartments, and condominiums $500,000
National defense enterprises Substantial impact on national defense,
irrespective of whether

other standards are met. Employer association If any member meets other
jurisdictional standards or the

combined operations of all members meet any such standard. Single employer
engaged in multiple enterprises Overall operations meet any standard.
Instrumentalities links and channels of interstate commerce $50,000 Other
transit systems a $250,000 Restaurants and country clubs $500,000

a Local and suburban transit, interurban highway passenger transportation,
and transportation utilities. The Board also extends NLRA jurisdiction to
secondary employers involved in labor disputes. Source: Hardin, P. (1992).
Developing Labor Law: The Board, the Courts, and the National Labor
Relations Act. Bureau of National Affairs.

6 The total number of employees excluded under the Board*s jurisdictional
standards also includes supervisors and managers in those firms; these
employees would still be excluded from NLRA coverage even if there were no
Board jurisdictional limits. Assuming that the percent of supervisors in
these small businesses was the same as our estimate for the labor force as
a whole, about 9 percent or about 500,000 employees in these excluded
small businesses would be excluded under NLRA*s supervisor/ manager
exclusion.

Appendix I: Methodology Page 28 GAO- 02- 835 Collective Bargaining

Figure 4: Percentage of Workers Affected by the Board*s Dollar Sales
Volume Standards, by Industry, February 2001

Note: The Agriculture/ Fish/ Forestry, Mining, Manufacturing, and
Wholesale Trade industries all had less than 1 percent of their workers
excluded by the Board*s jurisdictional standards. The numbers in the chart
refer to the estimated number of workers affected by the Board*s
jurisdictional standards in February 2001.

Source: GAO*s analysis of February 2001 CPS Supplement, 2000 Economic
Census, and the 1997 NCS.

The NLRA also excludes managers and supervisors from coverage. To estimate
the percentage of managers and supervisors by industry, we used data from
the 1997 Bureau of Labor Statistics* National Compensation Supervisors/
Managers

Percent of industry work force 0 2

4 6

8 10

12 14

Construction Trans/ Utility

Retail trade Services

Finance/ Real estate Industry total

80,862 122,491 2,652,123

1,594,191 988,854

115,024,759

Appendix I: Methodology Page 29 GAO- 02- 835 Collective Bargaining

Survey (NCS). 7 The NCS provides recent nationally representative
employer- reported data on job duties at the establishment level,
permitting the estimation of the number of jobs in a particular industry
that are supervisory or managerial in nature. Among the information
collected in the job survey is a ranking or leveling factor measuring
supervisory duties. This factor is designed to account for the additional
responsibilities of supervisors and to indicate the hierarchical level of
the positioning of the organization. Using this generic leveling factor,
jobs are classified according to their supervisory duties. (See table 4.)

. 7 BLS collects data for the NCS from more than 16,000 establishments,
weighted to represent more than 335,000 establishments employing almost 67
million workers. When an establishment is first surveyed, specific jobs
are selected through a sampling procedure and then classified to an
appropriate occupation. The work level of the particular job is
established by assessing the duties and responsibilities of the job
according to the survey*s nine generic leveling factors: (1) knowledge,
(2) supervision received, (3) guidelines, (4) complexity, (5) scope and
effect, (6) personal contacts, (7) purpose of contacts, (8) physical
demands, and (9) work environment. Each factor contains a number of
levels, and every level has an associated description and point value. See
James Smith,

Supervisory Duties and the National Compensation Survey, Compensation and
Working Conditions, Spring 2000.

Appendix I: Methodology Page 30 GAO- 02- 835 Collective Bargaining

Table 4: Definitions of Supervisory Leveling Factors, 1997 NCS Level
Definition

Team leader, group leader, or lead worker a Employee sets pace of work for
the group and shows other

workers in the group how to perform assigned tasks. He or she performs the
same work as the group in addition to lead duties. First line supervisor
Directs staff through face- to- face meetings. Organizational

structure is not complex, and internal and administrative procedures are
simple. Performing the same work as subordinates is not the principal
duty. Second line supervisor Directs staff through intermediate
supervisors. Internal

procedures and administrative controls are formal. Organization structure
is complex and is divided into subordinate groups that may differ from
each other as to subject matter and function. Third line supervisor
Directs staff through two or more subordinate supervisory

levels with several subdivisions at each level. Programs are usually
interlocked on a direct and continuing basis with other organization
segments, requiring constant attention to extensive formal coordination,
clearances, and procedural controls.

Note: An additional leveling factor is no supervisory responsibilities.
This category would include line workers. a Team leaders, group leaders,
or lead workers are considered in the NCS as representing

nonsupervisory positions. Based on the NCS, there are approximately 3.4
million private sector fulltime employees who would fall into the *team
leader* category, including 1.7 million team leaders in professional and
technical occupations. Team leaders include employees in professional and
technical occupations who may exercise considerable independent judgment.
Professional and technical occupations have a much higher percentage of
team leaders than first line supervisors; examples of these occupations
would include registered and licensed practical nurses, computer systems
analysts, and computer scientists.

Source: James Smith, Supervisory Duties and the National Compensation
Survey, Compensation and Working Conditions, Spring 2000.

We obtained NCS- based estimates of the supervisory level percentages for
wage and salary employees with full- time status by industry. We then
applied these percentages to the CPS full- time workforce estimates by
industry to estimate the number of full- time managerial and supervisory
employees. We defined supervisors as first- line supervisors and managers
as second- line and third- line supervisors.

We assume the industry percentage of supervisors applies to all firms in
an industry regardless of size. To avoid double counting, we applied the
NCS industry level supervisor percentages to the CPS total labor force
estimate of full- time wage and salary workers by industry net of the
small business

Appendix I: Methodology Page 31 GAO- 02- 835 Collective Bargaining

exclusion. 8 To calculate the percentage of workers excluded as managers
and supervisors, we divided the total number of full- time supervisors by
the total number of private sector employees. To the extent that managers
and supervisors are part- time employees, we underestimated the total
number of employees in that category, although we believe that this number
is small.

The NCS, while it provides reasonable data to estimate the number of
supervisors excluded under the NLRA, has several limitations. The NCS data
used in this report were collected in 1996 and thus may not completely
reflect the hierarchical structure of establishments in 2001. In addition,
the NCS data are based on a description of jobs, and while reasonable
conceptually, it is not strictly comparable with our workerbased estimates
from the CPS. In addition, there were certain occupations that could not
be leveled (e. g., musicians), and there are some small groups of jobs
that were restricted to a certain level (e. g., personnel managers were
restricted to supervisory duties of the first tier or higher). However, we
believe that these limitations are probably not large enough to have a
significant effect on our estimate.

The NLRA excludes coverage of domestic workers. Using the February 2001
CPS, we estimated the number of domestic workers to be excluded by first
focusing on the personal service/ private household (SIC 761). Within that
SIC, we then defined domestic workers as employees in the following
Standard Occupational Classification (SOC) codes: laundry/ ironers: 403,
cooks 404, child- care providers 406, and cleaner/ servant/ housekeeper
407. We summed the total number of employees in these SOC codes to
estimate the number of those excluded in this category and then subtracted
them from the CPS total labor force. 9

Using the U. S. Department of Agriculture (USDA) definition for *hired
hands,* we estimated all excluded agricultural workers from the CPS

8 The Railway Labor Act has a different definition of supervisor than the
NLRA. For those industries covered by the Railway Labor Act, we defined
supervisors as second- level supervisors and managers as third- level
supervisors.

9 Because of sampling limitations, we did not correct our estimated number
of domestic workers by subtracting those domestic workers, for example,
those in California, who might have rights under other laws. Making such a
correction would slightly increase the percentage of the labor force that
has collective bargaining rights. Domestic Workers

Agricultural Workers

Appendix I: Methodology Page 32 GAO- 02- 835 Collective Bargaining

estimates for agricultural wage and salary workers. 10 USDA defines hired
hands as SOC 475 for managers, farms, excluding horticultural; 476 for
managers, horticultural specialty farms; 477 for supervisors, farm
workers; 479 for general farm workers; and 484 for general nursery
workers. We used these same SOC codes as a definition to exclude
agricultural workers from the workforce in the CPS. Because nine states
have separate laws that give agricultural laborers bargaining rights with
either (1) an agricultural labor act 11 or (2) private employer labor
relation acts without exclusion of agricultural workers, 12 we added the
estimates of agricultural workers from these states back into the total
estimate to account for state and local laws. The estimated number of
total agricultural workers in the CPS is likely low because the CPS tends
to undercount agricultural workers and February is often the low
employment month for many crops. It is unclear how these factors affect
our estimate of the number of agricultural workers with collective
bargaining rights.

Although the NLRA excludes public employees from coverage, many public
employees have rights under other laws. 13 To estimate the number of state
and local public employees with rights, we used data from the

10 USDA conducted its research using the CPS as a basis to define
agricultural workers and also uses the National Agricultural Statistical
Survey to estimate and describe this population. We chose USDA*s
definition because it is derived from a 12- month data filing method to
calculate the number of agricultural workers in the United States so that
we would be able to separate the number of agricultural workers in
February to correspond with our data. See J. Runyan, USDA, (2001), which
also includes annual demographic and geographic averages of these workers.

11 These states are Arizona, California, Oregon, Kansas, and Maine. 12
These states are Hawaii, Massachusetts, New Jersey, Wisconsin, and Kansas.
13 Prior to 1995, the Board declined jurisdiction over employers whose
employees worked under a contract with a government entity, i. e., an
entity not covered by the NLRA, where the employer/ contractor lacked
control over the essential terms and conditions of employment. See Res-
Care, Inc., 280 N. L. R. B. 670 (1986). The Board no longer limits its
jurisdiction in this way. Management Training Corporation, 317 N. L. R. B.
1355 (1995). In this report, such workers are treated as private sector
employees. Public Employees

Appendix I: Methodology Page 33 GAO- 02- 835 Collective Bargaining

2000 Census Annual Survey of Government Employment, 14 a data source that
provides detailed information on state and local government employees. We
applied all of the state and local laws that excluded certain occupations
from coverage* in many cases police officers, firefighters, and teachers*
to estimate a percentage of workers with collective bargaining rights
within those occupations. 15 We applied a similar methodology for federal
employees. Starting with the CPS estimates for the total number of federal
employees, we first subtracted the number of employees in the legislative
and judicial branches of government and in the U. S. Postal Service, using
their relative shares of federal employment as estimated in using the
Federal Civilian Workforce Statistics: Employment and Trends as of March
2001. 16 We then adjusted for those agencies, for example, the Central
Intelligence Agency, who have been exempted from bargaining rights. 17 We
then calculated the percentage of federal employees who were managers or
supervisors using the September 2001 Office or Personnel Management*s
Central Personnel Data File (CPDF) and reduced the remainder by the number
of exempted manager and supervisors in the executive branch workforce. 18
For postal workers,

14 The U. S. Census Annual Survey of Government Employment provides data
on full- time and part- time government employment. The survey reports on
hours worked and payroll statistics by type of government (state, county,
city, township, special district, school district) and by government
function (elementary and secondary education, higher education, police
protection, fire protection, financial administration, other government
administration, judicial and legal, highways, public welfare, solid waste
management, sewerage, parks and recreation, health, hospitals, water
supply, electric power, gas supply, transit, natural resources,
correction, libraries, air transportation, water transport and terminals,
other education, state liquor stores, social insurance administration, and
housing and community development). See The U. S. Census Annual Survey of
Government Employment 2001.

15 For purposes of enumeration, we counted as having collective bargaining
rights those public employees in three states who obtained rights meeting
our definitional criteria from a gubernatorial executive order, rather
than a statute. In those states with collective bargaining laws that
include an *opt- in* provision, we also counted as having rights those
employees who received rights from public employers who actually opted to
provide those rights.

16 Federal Civilian Workforce Statistics: Employment and Trends as of
March 2001,

U. S. Office of Personnel Management, OMSCE- OW1- 05- 01, May 2001. 17 We
assumed this estimate to be about 15 percent of the federal workforce.
However, we did not review all federal agencies to identify those workers
who are exempted. Furthermore, for some of these agencies, like the
National Security Agency, the employment levels are unavailable. See
Federal Civilian Workforce Statistics: Employment and Trends as of March
2001, U. S. Office of Personnel Management, OMSOE- OW1- 05- 01, May 2001,
note on p. 7.

18 This estimate from the CPDF was 12.6 percent.

Appendix I: Methodology Page 34 GAO- 02- 835 Collective Bargaining

we applied the percentage of the postal workforce estimated by the U. S.
Postal Service to be management employees and then added the remainder
back to the number of net non- postal federal workers to obtain a total
number of federal employees with bargaining rights. 19

While our datasets define many of the employee categories relevant to the
NLRA, few of these groups are definitionally identical to those specified
in the NLRA. (See table 5.) For example, although we were able to use the
NLRA*s definitions to estimate the number of employees excluded by the
Board*s small business jurisdictional standards, the definitions of
independent contractors, supervisors, and agricultural workers from the
datasets that we used are not identical to the NLRA*s definitions. To the
extent that the definitions of these groups we use in our datasets
operationally diverge from the NLRA, our estimates may not accurately
capture the size of the NLRA*s defined groups. However, we believe that
there is considerable overlap between the two sets of definitions and that
there is no systematic bias that we can identify in using the CPS and
other datasets to estimate the NLRA categories.

19 This estimate was 10.8 percent. See Draft Comprehensive Transformation
Plan,

U. S. Postal Service, October 1, 2001, p. 19. Definitional Differences

Between the CPS, NCS, and NLRA

Appendix I: Methodology Page 35 GAO- 02- 835 Collective Bargaining

Table 5: Comparison of NLRA Definitions of Private Sector Excluded Groups
and Those Used in GAO*s Quantitative Estimates

Excluded group NLRA definition Definition Used in quantitative estimates

Independent contractor An independent contractor relationship is
determined by applying the common law agency test and considering all
incidents of the individual*s relationship to the employing entity. See
NLRB v. United Insurance Co. of America, 390 U. S. 254 (1968); Roadway
Package System, Inc., 326 N. L. R. B. 842 (1998); and Dial- a- Mattress
Operating Corp., 326 N. R. L. B. 884 (1998).

Wage and salary workers or self- employed in the CPS who answered
affirmatively to the question, *Last week, were you working as an
independent contractor, an independent consultant, or a free- lance
worker? That is, someone who obtains customers on their own to provide a
product or service.*

Supervisor/ manager A supervisor is one who has the authority, in the
interest of the employer, to hire, transfer, suspend, lay off, recall,
promote, discharge, assign, reward, or discipline other employees or
responsibly to direct them, or to adjust their grievances, or effectively
to recommend such action, if . . . such authority is not merely of a
routine or clerical nature, but requires the use of independent judgment.
29 U. S. C. S: 152( 11). A managerial employee is one who formulates and
effectuates management policies by expressing and making operative the
decisions of their employer, who exercises discretion within, or even
independently of, established employer policy, and who is aligned with
management. NLRB v. Yeshiva University, 444 U. S. 672 (1980).

First, second and third- line supervisors as defined in table 4.

Domestic worker A domestic worker is one whose employment falls within the
commonly accepted meaning of the term *domestic servant.* Ankh Services,
Inc., 243 NLRB 478 (1979); 29 U. S. C. S: 152( 3).

Employees in SOC codes 403: laundry/ ironers, 404: cooks, 406: child- care
providers, and 407: cleaner/ servant/ housekeeper, employed in SIC code
761, personal service/ private household. Agricultural worker An
agricultural worker is one who engages in

farming in all its branches, including the cultivation and tillage of the
soil, dairying, the production, cultivation, growing, and harvesting of
any agricultural or horticultural commodities, the raising of livestock,
bees, fur- bearing animals, or poultry, and any practices (including any
forestry or lumbering operations) performed by a farmer or on a farm as an
incident to or in conjunction with such farming operations, including
preparation for market, delivery to storage or to market or to carriers
for transportation to market. National Labor Relations Board Appropriation
Act, 60 Stat. 698, ch. 672, Title IV (1947), referring to the Fair Labor
Standards Act, 27 U. S. C. S: 203( f).

*Hired hands* defined as SOC Codes 475: managers, farms, excluding
horticultural; 476: managers, horticultural specialty farms; 477:
supervisors, farm workers; 479: general farm workers; 484: general nursery
workers.

Employees of certain small businesses See table 3. Same annual sales
dollar volume.

Source: GAO legal analysis and definitions from selected datasets.

Appendix I: Methodology Page 36 GAO- 02- 835 Collective Bargaining

We also constructed an estimate of the undocumented alien immigrant
population that could have their rights diminished by the Hoffman Plastic

decision. To construct this estimate, we used the U. S. Census Bureau*s
estimated number of 8.7 million for the residual foreign- born population
(see table 6) and adjusted it for a variety of factors. 20 First, to
estimate the number of eligible workforce participants, we counted only
those above the age of 18 and below 65 in the 8.7 million number, yielding
a total of about 7.1 million persons. (See table 6.) We then accounted for
the percent that are likely to be covered by collective bargaining rights
in the workforce. We apply an estimate of the private sector labor force
covered by the NLRA of 77 percent to obtain an estimate of about 5.5
million persons (5,485,538). Although the bulk of this estimate contains
undocumented alien workers, it also includes a small group of so- called
quasi- legal residents; persons who are in a variety of transitional
immigrant categories but at the time do not have explicit legal or illegal
status. 21

20 This residual foreign- born population is not an estimate of the number
of unauthorized immigrants. This estimate also includes people who are
here legally but are not yet included in the official estimates of legal
migrants and refugees. Because the method was derived from a residual
methodology, any limitations in the methods or in the measurement of other
migration components are reflected in the residual number. For a
discussion of issues in estimating undocumented alien workers see Lindsay
Lowell and Robert Suro (2002). How Many Undocumented? The Numbers Behind
the U. S.- Mexico Migration Talks. The Pew Hispanic Center. Bean, F., R.
Corona, R. Tuiran, K. Woodrow- Lafield and J. Van Hook. (2001). Circular,
Invisible, and Ambiguous Migrants: Components of Difference in Estimates
of the Number of Unauthorized Mexican Migrants in the United States.
Demography 38: 411- 422. Passel, Jeffrey S., Estimates of Undocumented
Immigrants Living in the United States: 2000, The Urban Institute. August
2001.

21 The Census estimates the quasi- legal migrant population to be
approximately 1.7 million. This estimate includes refugees (who have not
yet adjusted status) and asylum applicants (awaiting claim adjudication)
(200, 000 and 400,000, respectively), migrants deported during the decade
(200, 000), and population legalizing (adjusting status) during the decade
(900,000). Estimation of the Number

of Undocumented Alien Workers

Appendix I: Methodology Page 37 GAO- 02- 835 Collective Bargaining

Table 6: Census Level Estimates of the Foreign- Born Population by Migrant
Status in 2000

Migrant status Number

Foreign- born population 31,098,945 Legal immigrants 21,612,023 Temporary
immigrants 781,507 Residual foreign- born 8, 705,419 a a The residual
foreign- born estimate also includes people in quasi- legal status who are
awaiting action on their legal migration requests. Source: J. Gregory
Robinson. ESCAP II: Demographic Analysis Results (2001). U. S. Census
Bureau.

Table 7: Estimates of Residual Foreign- Born by Age Age of work eligible
Total number of persons in that age group

18- 29 3,483,802 30- 49 2,605,123 50+ 627,538 50- 64 407,613

Total 7,124,076

Note: The number of persons in each age group includes persons of all
races. Source: Costanzo, J., Davis, C., Caribert, I., Goodkind, D., and
Ramirez, R. Estimating the Residual Foreign Born Population in 1990 and
2000. U. S. Census, Population Division. May 2002.

There were several limitations to consider when determining the estimate
of undocumented alien immigrants due to the characteristics and
circumstances of this population. We include the quasi- legal population
in the residual foreign- born estimate because it includes people who are
waiting for action to be taken on their legal migration requests. We
included this group in our total estimate because they were assumed to be
illegal at the time they were counted in the Census. 22

To obtain an estimate of the workers affected by the Hoffman Plastic

decision, we assumed that the distribution of undocumented alien workers
across industries was identical to that of the total civilian labor force.
Our methodology for the entire workforce was conducted in February 2001
while the Census estimates are for the entire year of 2000. In addition,

22 We also assumed that this group continuously replaces itself by new
quasi- legal immigrants of similar characteristics who are awaiting legal
action on their migration requests so it would not affect our estimates of
the total work eligible population.

Appendix I: Methodology Page 38 GAO- 02- 835 Collective Bargaining

although undocumented alien workers are likely not to be distributed in
this manner, it is unclear what their distribution is. For example,
undocumented alien workers are probably less likely to be employed in
government employment, although they may be more likely to work in small
businesses that are exempt from Board jurisdiction and as domestic
workers. In addition, we based our assumption that the percentage of
private sector workers with collective bargaining rights under the NLRA
(77 percent), on our private sector estimate of 78 percent, and then
corrected for coverage under the Railway Labor Act. 23 However, there is a
high percentage of undocumented alien workers in agriculture, another
industry excluded from the NLRA, suggesting that the 77 percent may be an
overestimate. On the other hand, undocumented workers may be more likely
to be rank and file workers rather than supervisors and thus covered by
the NLRA. We also assumed 100 percent Census coverage of the foreign- born
population, although, historically, Hispanic populations have been
undercounted in the Census, and the majority of undocumented alien
immigrants are of Hispanic origin. Such an undercount would imply that our
estimate is an undercount as well. However, there is currently no
available data on the size of the undercounted population in the Census.
24 On balance, the 77- percent estimate appears to represent a reasonable
approach.

We chose to estimate the percentage of the labor force with bargaining
rights using data based on persons because bargaining rights is a concept
associated with persons and not jobs. However, because some workers are
multiple jobholders, it is possible that deciding whether workers have
rights on the basis of their primary job could result in discrepant
misclassifications. 25 An alternative methodology would be to identify
those jobs that would confer rights to individuals. Using the 2001 Current
Employment Statistics Survey (CES) that collects information through the

23 Using the CPS Supplement, there were approximately 1.3 million workers
in the airline and railway industries, or about 1 percent of the civilian
labor force, in 2001. 24 We also assumed a labor force participation rate
(the total of the employed and unemployed) for undocumented alien workers
of 100 percent. While data are not available to estimate the labor force
participation rate of undocumented workers, it is likely to be quite high.

25 For example, a worker who was primarily an independent contractor but
who also worked at a covered job secondarily would not be counted under
our methodology. Estimates of the

Percentage of the Labor Force with Collective Bargaining Rights Based on
*Jobs* Data

Appendix I: Methodology Page 39 GAO- 02- 835 Collective Bargaining

survey of employers, 26 we constructed estimates for the percentage of the
non- farm non- supervisory, private labor force that would have bargaining
rights and then compared the estimates with the CPS. While we found that
the CES survey had a total nonfarm private workforce estimate that was
somewhat higher than the CPS, our estimates for the percentage of the
nonfarm private labor force with bargaining rights were quite close. 27

Our estimates of workers either having or not having collective bargaining
rights are imprecise. For the most part, our final national population
estimates are derived from data obtained from different national survey
samples of the worker population taken at different times. These surveys
are subject to both nonsampling and sampling errors. Nonsampling errors
can stem from many sources, such as a failure to sample a segment of the
population, inability to obtain information from respondents in the
sample, differences in interpretation of a question, or errors made in the
collection or processing of the data obtained.

Sampling errors occur because the observations used to develop an estimate
are not based on the entire population but rather on one of a number of
samples of the same size that could have been selected using the sample
design. For any given survey, estimates derived from the different samples
would differ from each other. A measure of such variation in samples,
known as the standard error or sampling error, can be used to develop a
confidence interval around an estimate. The confidence interval
establishes an upper and lower bound on the estimate that would encompass
the true population value with an expressed degree of probability. 28

26 The CES survey is an establishment payroll survey administered to a
monthly sample of nearly 400,000 business establishments nationwide. The
primary statistics derived from the survey are monthly estimates of
employment, hours, and earnings for the nation, states, and major
metropolitan areas.

27 While the number of nonfarm private sector workers was about 90 percent
of comparable CES figures, the percentage of the labor force with
bargaining rights in the CPS was about 1 percent point higher (83 percent
of nonfarm workers in the private sector had rights in the CPS compared
with 82 percent in the CES).

28 For extensive discussion of survey errors and their evaluation, see The
Office of Management and Budget*s Statistical Policy Working Paper 31:
Measuring and Reporting Sources of Error in Surveys, June 2001.
Reliability of Workforce

Estimates

Appendix I: Methodology Page 40 GAO- 02- 835 Collective Bargaining

In an effort to partially reflect the imprecision associated with our
estimates, we also developed an associated estimate of an upper and lower
bound to convey a range in which the true population value was most likely
to fall. The estimates that we developed are based on estimates of
sampling error only. We did not attempt to assess or account for the
various other sources of error in the data sources we used to develop our
workforce estimates of eligibility for collective bargaining. For example,
we do not know nor can we assess the errors associated with using data
referring to different years, based on somewhat different industry
classification systems, or being collected from individuals versus
establishments.

For our estimates of the following labor segments* independent
contractors, domestic workers, agricultural workers, employees of certain
small businesses, federal, state and local government employees, and
federal, state and local government employees eligible for collective
bargaining rights-- we produced confidence intervals at the 95- percent
level based on standard error estimates provided in the technical
documentation for the February 2001 CPS.

Our confidence interval estimates for independent contractors, employees
of certain small businesses, and state and local government employees were
at or below + 3 percent of the figures reported. Federal employees were at
or below + 6 percent of the figures reported. For domestic workers and
agricultural workers, our confidence interval estimates were at or below +
15 percent of the figures reported.

We also used February 2001 CPS standard error estimates to calculate the
confidence intervals for our estimates of certain labor segments by
private sector industries. For all industries, our confidence intervals
for estimates of the proportion that are employees of certain small
businesses is + 1 percentage point or less. The independent contractor
confidence intervals for all industries is + 3 percentage points or less.

For the NCS, which we used to obtain estimates of managers and supervisors
in the private sector and the state and local government workforce, there
were no available estimates of sampling error for employee estimates. In
preparing the NCS data on occupational wages and benefits, BLS only
produces measures of sampling error on earnings and not estimates of the
number of employees whose job characteristics include supervisory or
managerial work factors. Instead, we developed approximate estimates of
sampling error using information about the

Appendix I: Methodology Page 41 GAO- 02- 835 Collective Bargaining

allocation of the NCS sample 29 and applying a conservative assumption of
the survey*s design effect.

Our confidence interval estimate for excluded private sector managers and
supervisors overall is + 18 percent of the figure reported. For first-
line supervisors, our estimated interval is + 21 percent.

The multiple steps and data sources used in our estimation procedures
precluded direct computation of confidence intervals for our overall
estimates of workers eligible for collective bargaining either nationally
or by industry. In an effort to provide some quantified bounds reflecting
the imprecision of our national and overall industry estimates, we used
mean and variance estimates (by industry) for various components of the
overall estimates to approximate the sampling distributions for each of
those components. We then used these distributions to simulate how our
overall estimate would vary due to sampling. 30

Our overall bounds estimate for the total civilian workforce, total
government workforce, and total private sector workforce eligible for
collective bargaining rights is + 3 percent or less. For all industries
except mining and construction, our bounds estimates of the proportion
that are eligible for bargaining rights is + 6 percentage points or less.
For construction, the bound is + 9 percentage points, and for mining it is
+ 19 percentage points.

Although we made an effort to identify and count those workers who might
have rights from laws other than the NLRA, there are some for whom we did
not do so. In particular, one group of these employees would be those
covered by state private sector labor relations laws (Little Wagner Acts).
There are 18 states that have laws that provide collective bargaining
rights to some employees who would otherwise be excluded

29 Technical Note in Occupational Compensation Survey: Occupational Wages
in the United States, 1997, Bureau of Labor Statistics, U. S. Department
of Labor, Revised September 1999.

30 While there are no widely accepted methodologies for combining sampling
errors associated with estimates derived from multiple data sources, we
believe that this nonparametric approach is a reasonable method to use for
approximating bounds around our national estimates. State Labor Relations
Laws

Appendix I: Methodology Page 42 GAO- 02- 835 Collective Bargaining

from the federal NLRA. 31 Of these 18 states, 17 states extend coverage to
employees of small businesses, 11 states to independent contractors, 8
states include supervisors, 5 to agricultural laborers, and 1 includes
domestic servants. 32 Except for agriculture, we did not incorporate the
coverage under these state laws into our quantitative estimates and on
balance this would lead to an underestimation of the percentage of the
labor force with collective bargaining rights.

31 The states with a state labor relations act are Colorado, Connecticut,
Hawaii, Kansas, Massachusetts, Michigan, Minnesota, New Jersey, New York,
North Dakota, Oregon, Pennsylvania, Rhode Island, South Dakota, Utah,
Vermont, West Virginia, and Wisconsin.

32 State labor relations laws covering agricultural laborers are Colorado,
Connecticut, Michigan, Minnesota, and Wisconsin. Wisconsin*s labor
relations law covers domestic servants. State labor relations laws
covering supervisors are Connecticut, Kansas, Massachusetts, New Jersey,
New York, Pennsylvania, Rhode Island, and Utah. State labor relations laws
covering independent contractors: are Connecticut, Kansas, Massachusetts,
Michigan, Minnesota, New Jersey, New York, Pennsylvania, Rhode Island,
South Dakota, and Utah. Those states covering employees of small
businesses are Colorado, Connecticut, Kansas, Massachusetts, Michigan,
Minnesota, New Jersey, New York, North Dakota, Oregon, Pennsylvania, Rhode
Island, South Dakota, Utah, Vermont, West Virginia, and Wisconsin.

(130097)

Appendix I: Methodology Page 43 GAO- 02- 835 Collective Bargaining

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. ***