Worker Protection: OSHA Inspections at Establishments Experiencing Labor
Unrest (Letter Report, 08/31/2000, GAO/HEHS-00-144).

Pursuant to a congressional request, GAO reviewed worker protection,
focusing on: (1) the extent to which employers experiencing labor unrest
are more likely to be inspected than employers not experiencing labor
unrest; and (2) whether the Occupational Safety and Health
Administration (OSHA) has policies for performing inspections during
labor unrest and whether these policies are followed.

GAO noted that: (1) establishments experiencing labor unrest are about
6.5 times more likely to be inspected by OSHA than establishments not
experiencing labor unrest (8.6 percent inspected compared with about 1.3
percent) during fiscal years (FY) 1994 through 1998; (2) the statutory
requirement that OSHA investigate valid complaints and OSHA's policy to
investigate a fatality or catastrophe may be related to the higher rate
of inspection for establishments experiencing labor unrest; (3) about 68
percent of the approximately 1,900 OSHA inspections conducted each year
at establishments experiencing labor unrest resulted from complaints,
fatalities, or catastrophes; (4) in contrast, only about 27 percent of
the approximately 100,000 total inspections OSHA conducted each year
resulted from complaints, fatalities, or catastrophes; (5) about 76
percent of the establishments with labor unrest that were inspected by
OSHA from FY 1994 through FY 1998 were unionized, as compared with about
24 percent of all establishments inspected by OSHA over this period; (6)
while it did not appear that unionized establishments were in general
more likely to receive a complaint-based inspection than nonunionized
establishments, GAO's analysis did find that, among establishments
experiencing labor unrest, there were a higher proportion of
complaint-based inspections at unionized establishments than at
nonunionized establishments; (7) OSHA's policy concerning inspections
during labor unrest provides discretion for programmed inspections; (8)
OSHA may delay programmed inspections during periods of labor unrest,
such as a strike, which would prevent OSHA inspectors from witnessing
actual work operations; (9) however, OSHA officials said that it has
rarely delayed these inspections; and (10) statutory requirements or
OSHA's long-standing policy dictate that inspections resulting from
valid complaints, fatalities, or catastrophes--which accounted for the
majority of OSHA's inspections at establishments with labor unrest
during FY 1994 through FY 1998--must be performed, regardless of whether
labor unrest exists.

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

 REPORTNUM:  HEHS-00-144
     TITLE:  Worker Protection: OSHA Inspections at Establishments
	     Experiencing Labor Unrest
      DATE:  08/31/2000
   SUBJECT:  Working conditions
	     Occupational safety
	     Labor law
	     Safety standards
	     Health hazards
	     Inspection
	     Labor unions
	     Industrial relations
	     Strikes (labor actions)
IDENTIFIER:  California
	     Michigan
	     Oregon
	     Washington

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Testimony.                                               **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
******************************************************************
GAO/HEHS-00-144

Appendix I: Objectives, Scope, and Methodology

16

Appendix II: Comments From the Department of Labor

22

Appendix III: GAO Contacts and Staff Acknowledgments

24

Table 1: Number of Establishments Experiencing Labor Unrest,
by Source, Fiscal Years 1994-98 17

Table 2: Estimated Number of Establishments With Labor Unrest Inspected by
OSHA, Fiscal Years 1994-98 18

Table 3: Number of Establishments With and Without Labor Unrest,
Fiscal Years 1994-98 19

Table 4: Number of OSHA Inspections at Establishments With and
Without Labor Unrest, Fiscal Years 1994-98 19

Table 5: Percentage of Establishments Without Labor Unrest
Inspected by OSHA, Fiscal Years 1994-98 20

Figure 1: Inspection Rates for Establishments With and Without
Labor Unrest, Fiscal Years 1994-98 8

Figure 2: Reasons for Inspections at Establishments With Labor
Unrest and for All OSHA Inspections, Fiscal Years 1994-98 10

Figure 3: Inspections of Unionized and Nonunionized Establishments, With
Labor Unrest and for All OSHA Inspections,
Fiscal Years 1994-98 12

Figure 4: Percentage of Inspections by Industry for All
OSHA Inspections and Inspections at Establishments
With Labor Unrest, Fiscal Years 1994-98 21

BLS Bureau of Labor Statistics

FMCS Federal Mediation and Conciliation Service

IMIS Integrated Management Information System

NLRB National Labor Relations Board

OSHA Occupational Safety and Health Administration

Health, Education, and
Human Services Division

B−284100

August 31, 2000

The Honorable Peter Hoekstra
Chairman, Subcommittee on Oversight and Investigations
Committee on Education and the Workforce
House of Representatives

The Honorable Cass Ballenger
Chairman, Subcommittee on Workforce Protections
Committee on Education and the Workforce
House of Representatives

The Department of Labor's Occupational Safety and Health Administration
(OSHA) is responsible for ensuring healthful and safe working conditions in
nearly all of the approximately 7 million workplaces nationwide. To carry
out this responsibility, OSHA and its state partners conduct about 100,000
onsite inspections of employers each year, on a programmed or routine basis
and in response to fatalities or catastrophes1 or worker complaints alleging
violations of safety and health laws. Employees have the right to notify
OSHA of any potential safety and health violations, and OSHA2 is required by
law to investigate all valid3 complaints about serious hazards from
employees or their representatives. OSHA's long-standing policy also
requires the investigation of all fatalities and catastrophes.

In recent years, however, employers and members of Congress have raised
concerns that employees may file complaints with OSHA during periods of
labor unrest solely to secure an inspection by OSHA as a way to pressure
employers to meet employee demands. Others maintain that the conditions that
lead to labor unrest may often be related to those that lead to potential
health and safety violations. According to Labor officials, labor unrest and
resulting breakdowns in labor-management relations can disrupt the internal
resolution of health and safety problems and can cause workers and/or unions
to turn to outside compliance authorities such as OSHA. Given the different
viewpoints about OSHA's role and the timing of its inspections, you asked
that we determine (1) the extent to which employers experiencing labor
unrest are more likely to be inspected than employers not experiencing labor
unrest, and (2) whether OSHA has policies for performing inspections during
labor unrest and whether these policies are followed.

To address your request, we identified those establishments that experience
labor unrest. There is no specific definition for labor unrest, although it
includes situations such as work stoppages (strikes), allegations of unfair
labor practices by employers or unions, campaigns to organize employees into
a union, or efforts to remove an existing union from an establishment. We
found only two databases that track specific events that may be related to
labor unrest--the Federal Mediation and Conciliation Service (FMCS), which
tracks work stoppages, and the National Labor Relations Board (NLRB), which
investigates and resolves complaints about unfair labor practices and
processes petitions requesting employee elections to determine union
representation. Using these databases, we identified about 22,000
establishments each year from fiscal years 1994 through 1998 that
experienced conditions involving labor unrest. We estimated the OSHA
inspection rate for a random sample of these establishments and compared it
with an OSHA inspection rate we developed for establishments not
experiencing labor unrest.4 When possible, we provided information on the
characteristics of these inspections5 (see app. I for a more detailed
description of our scope and methodology). We performed our work between
June 1999 and June 2000 in accordance with generally accepted government
auditing standards.

Establishments experiencing labor unrest are about 6.5 times more likely to
be inspected by OSHA than establishments not experiencing labor unrest (8.6
percent inspected compared with about 1.3 percent) during fiscal years 1994
through 1998. The statutory requirement that OSHA investigate valid
complaints and OSHA's policy to investigate a fatality or catastrophe may be
related to the higher rate of inspection for establishments experiencing
labor unrest. About 68 percent of the approximately 1,900 OSHA inspections
conducted each year at establishments experiencing labor unrest resulted
from complaints, fatalities, or catastrophes. In contrast, only about 27
percent of the approximately 100,000 total inspections OSHA conducted each
year resulted from complaints, fatalities, or catastrophes. We found that
about 76 percent of the establishments with labor unrest that were inspected
by OSHA from fiscal year 1994 through fiscal year 1998 were unionized, as
compared with about 24 percent of all establishments inspected by OSHA over
this period. While it did not appear that unionized establishments were in
general more likely to receive a complaint-based inspection than
nonunionized establishments, our analysis did find that, among
establishments experiencing labor unrest, there were a higher proportion of
complaint-based inspections at unionized establishments than at nonunionized
establishments.

OSHA's policy concerning inspections during labor unrest provides discretion
for programmed inspections. OSHA may delay programmed inspections during
periods of labor unrest, such as a strike, which would prevent OSHA
inspectors from witnessing actual work operations. However, OSHA officials
said that it has rarely delayed these inspections. Statutory requirements or
OSHA's long-standing policy dictate that inspections resulting from valid
complaints, fatalities, or catastrophes? which accounted for the majority of
OSHA's inspections at establishments with labor unrest during fiscal years
1994 through 1998?must be performed, regardless of whether labor unrest
exists.

OSHA was created in 1970 to carry out the Occupational Safety and Health
Act, which declared a national policy of ensuring safe and healthful working
conditions for every man and woman. To carry out its mission, OSHA develops
and enforces workplace safety and health standards, educates employers and
employees about workplace hazards, and along with its state partners,6
conducts inspections of employers to assess compliance with applicable
safety and health standards. Of the approximately 100,000 inspections
conducted each year from fiscal year 1994 through fiscal year 1998, about
65,000 were conducted by the 25 states and territories with approved OSHA
programs. Four of these states?California, Michigan, Oregon, and
Washington?conducted about half of the state inspections.

OSHA and its state partners conduct inspections for a variety of reasons.
Programmed inspections are inspections that are planned at the beginning of
each year and are generally targeted at industries and particular employers
with high injury or illness rates. During fiscal years 1994 through 1998,
more than half of OSHA's inspections were programmed. Sixty percent of the
remaining inspections resulted from valid complaints (which OSHA is required
by law to investigate) and fatalities and catastrophes that occurred in the
workplace (which OSHA is required by policy to investigate).

Although OSHA has the authority to inspect most types of
establishments--from large manufacturing plants to small retail stores--
some limitations govern when an inspection can be conducted. For example,
OSHA does not have the authority to conduct any type of inspection at a farm
that employs 10 or fewer employees and that does not have a temporary labor
camp. For an establishment employing 10 or fewer employees in low hazard
industries, OSHA does not have the authority to conduct a programmed safety
inspection but can conduct other types of inspections.7 In addition,
although OSHA generally can conduct inspections in any industry, it tends to
focus its inspection resources on those industries that present the greatest
workplace hazards to workers. For example, in fiscal years 1994 through
1998, OSHA conducted about 44 percent of its inspections in the construction
industry, which has the highest death rate of any industry. OSHA conducted
about 26 percent of its inspections in manufacturing, 11 percent in the
service industry, 8 percent in wholesale/retail trade, and the remaining 11
percent in other industries.

The term "labor unrest" is commonly used to indicate some type of
dissatisfaction among workers, but there is no consensus about the ways in
which labor unrest develops or the forms it takes. However, it is clear that
there is some relationship between labor unrest and employees'
dissatisfaction with wages and working conditions. For example, a strike,
which we are using as an example of labor unrest, may result, at least in
part, from unsafe working conditions. On the other hand, a strike also might
lead to unsafe working conditions if an employer calls in replacement
workers who are less trained and thus may be more susceptible to injury.
Allegations of unfair labor practices, another example of labor unrest, can
be filed by employees if the employer subjects them to unsafe working
conditions, or by an employer if workers conduct a work slowdown, which also
could be in response to working conditions. As a result, conditions that
lead to labor unrest and those that lead to safety and health violations may
be related.

The OSHA inspection rate at establishments that experienced labor unrest
during fiscal years 1994 through 1998 was 6.5 times higher (8.6 percent)
than at establishments that did not experience such unrest (1.3 percent).
This may be related to OSHA's statutory requirement to investigate valid
complaints and union employees' awareness of their right to file complaints
with OSHA, in addition to its policy to investigate fatalities and
catastrophes.8

During fiscal years 1994 through 1998, OSHA inspected, on average, about 8.6
percent of the approximately 22,000 establishments each year that
experienced labor unrest (about 1,900 inspections a year). In contrast, OSHA
inspected, on average, about 1.3 percent of the approximately 7 million
establishments9 each year that did not experience labor unrest (about 93,000
inspections each year) (see fig. 1).

Fiscal Years 1994−98

The percentage of establishments with labor unrest that were inspected
during fiscal years 1994 through 1998 ranged from about 5 percent to about
10 percent, while the percentage of establishments without labor unrest that
were inspected ranged from about 1.2 percent to about 1.6 percent.10

Inspection Rate for Establishments With Labor Unrest

The statutory requirement that OSHA investigate valid complaints and its
long-standing policy to investigate fatalities and catastrophes may help to
explain the higher rate of inspection for establishments with labor unrest.
About 68 percent of the inspections at establishments with labor unrest were
required because a complaint was filed (60 percent) or a fatality or
catastrophe occurred (8 percent). In contrast, only 27 percent of all
inspections OSHA conducted during this period were the result of a complaint
(22 percent) or a fatality or catastrophe (5 percent) (see fig. 2). Thus,
OSHA inspections at establishments experiencing labor unrest were 2.5 times
more likely to result from a complaint, fatality, or catastrophe than OSHA
inspections in general.

All OSHA Inspections, Fiscal Years 1994−98

Because of OSHA's requirements to investigate valid complaints and
fatalities and catastrophes, an OSHA inspection occurring at an
establishment with labor unrest may be wholly separate from any ongoing
labor unrest or may in fact be related to the labor unrest. For example, of
the four establishments we visited in California that had labor unrest and
were inspected by OSHA during periods of labor unrest, two of the
inspections were unrelated to the labor unrest that was ongoing at both
sites. One inspection resulted because an employee broke his ankle11; the
other inspection resulted when a parent filed a complaint citing the
possible presence of asbestos during the renovation of an educational
facility. For the other two inspections, however, there were indications
that the filing of the complaint with OSHA may have been a result of, or at
least related to, the labor unrest ongoing at the time. In one case, union
officials acknowledged that the OSHA complaint the union filed was related
to labor unrest, and in the second case, the employer believed that a
complaint filed by a union employee was in response to efforts to remove the
union from the workplace.

Despite the greater representation of complaint-driven inspections for
establishments with labor unrest, inspections at establishments with labor
unrest were just as likely to result in at least one violation or a
violation classified as "serious" as were all inspections conducted by OSHA.
For example, for both types of inspections during fiscal years 1994 through
1998, about 62 percent of all inspections resulted in at least one violation
and about 90 percent of those had 10 or fewer violations. Similarly, when
there were violations cited for establishments with labor unrest, about 51
percent of the violations were classified as "serious," as opposed to about
53 percent for all OSHA inspections.

Establishments With Labor Unrest

Approximately 76 percent of the OSHA inspections at establishments
experiencing labor unrest during fiscal years 1994 through 1998 were
conducted at unionized establishments. In comparison, about 24 percent of
all OSHA inspections during this time were conducted at unionized
establishments (see fig. 3).12

Unrest and for All OSHA Inspections, Fiscal Years 1994−98

According to OSHA officials, the higher rate of inspections for unionized
establishments experiencing labor unrest may in large part result from
OSHA's targeting procedures. They said that OSHA concentrates its
inspections in high-risk and relatively more unionized industry sectors,
which include construction, manufacturing, and oil/gas extraction. According
to OSHA officials, the greater dangers in these sectors and the greater
presence of unionization may also be related to a higher degree of labor
unrest.

Regarding complaint-driven inspections at establishments with labor unrest,
we found that 63 percent of those inspections at unionized establishments
resulted from complaints compared with about 52 percent for nonunionized
establishments. However, in general, OSHA inspections resulting from
complaints did not vary significantly between unionized and nonunionized
establishments, as the percentage of complaint-driven inspections at
unionized establishments was only slightly higher than for nonunionized
establishments (26 versus 21 percent, respectively). This indicates that
during periods of labor unrest, complaint-based inspections are somewhat
more likely at unionized establishments than at nonunionized establishments.

Union, NLRB, and OSHA officials, as well as employer representatives, said
that unionized employees are in general more aware of their worker
protection rights and may be more likely to file complaints with regulatory
agencies, especially during periods of labor unrest. An OSHA official said
that unions are also more aware of the role of regulatory agencies, both in
terms of protecting workers' rights and as a bargaining tool during contract
negotiations. Furthermore, according to a researcher well known for work in
this area, establishments with labor unrest are probably predominately
unionized because these employees are more likely to speak out and file
complaints with agencies such as OSHA in times of unrest.13

Unrest

Federal and state OSHA policies allow programmed inspections to be delayed
if a condition such as a strike is significant enough to prohibit OSHA
inspectors from witnessing normal work operations. However, according to
federal and most state OSHA officials, this has rarely occurred. Two states
highlighted conditions that would affect the decision to conduct a
programmed inspection. Kentucky OSHA officials stated that programmed
inspections would not be conducted if a picket line existed. The official
said that several reasons existed for this policy, including concern for the
safety of state inspectors and a desire not to get involved in
employer/employee controversies. Vermont OSHA officials said that the state
would postpone a programmed inspection if the company was on strike before
the start of the inspection, and if a strike began while the OSHA inspectors
were on site, OSHA inspectors would leave the premises and continue the
inspection after the strike was finished.

However, both federal and state OSHA inspection policies follow the
statutory requirement that OSHA investigate valid complaints, and state
policies follow federal policy regarding the investigation of fatalities and
catastrophes. Under these policies, federal and state OSHA officials stated
that inspections resulting from complaints, fatalities, and catastrophes are
to be conducted even during periods of labor unrest. They also said that, as
discussed above, these types of inspections accounted for the majority of
those OSHA performed at establishments with labor unrest from fiscal years
1994 through 1998.

In commenting on a draft of this report, Labor said that it recognized the
limitations in the data available to address the objectives of the review.
It believed that, because of these limitations, we could not conclude that
labor unrest led to an increased likelihood of OSHA inspections. While the
report estimates the percentage of establishments with labor unrest that
were inspected by OSHA, we do not draw any conclusions about a specific
relationship between the existence of labor unrest and OSHA inspections. In
fact, the report discusses a variety other factors that affect an
establishment's likelihood of inspection by OSHA. Labor's comments are
reproduced in app. II. Other technical comments provided by Labor were
incorporated as appropriate.

We are sending copies of this report to the Honorable Alexis M. Herman,
Secretary of Labor; the Honorable Leonard R. Page, General Counsel of the
National Labor Relations Board; the Honorable C. Richard Barnes, Director of
the Federal Mediation and Conciliation Service; relevant congressional
committees; and others who are interested. Copies will be made available to
others on request.

If you or your staffs have any questions concerning this report, please call
me at (202) 512-7215. Other GAO contacts and staff acknowledgments are
listed in app. III.

Marnie S. Shaul
Associate Director
Education, Workforce, and
Income Security Issues

Objectives, Scope, and Methodology

We were asked to determine (1) the extent to which employers experiencing
labor unrest are more likely to be inspected than employers not experiencing
labor unrest, and (2) whether the Occupational Safety and Health
Administration (OSHA) has policies for performing inspections during labor
unrest and whether these policies are being followed. To determine whether
employers experiencing labor unrest are more likely to be inspected than
employers not experiencing labor unrest, we (1) identified the universe of
establishments that experienced labor unrest and determined how many of
these were inspected by OSHA and (2) determined how many of the
establishments that did not experience labor unrest were inspected by OSHA.
To identify policies for conducting inspections during labor unrest, we
obtained OSHA policies and interviewed federal OSHA officials. We also
obtained inspection policies from each of the 25 states and territories that
conduct their own safety and health programs.

We also interviewed officials from the National Labor Relations Board
(NLRB); the Federal Mediation and Conciliation Service (FMCS); the Bureau of
Labor Statistics (BLS); representatives of the National Association of
Manufacturers, the Labor Policy Association, the AFL-CIO, and a California
longshoremen union; and David Weil, an associate professor of economics from
Boston University known for his research on occupational safety and health.
In addition, we visited four California employers that had experienced
various forms of labor unrest and had been inspected by OSHA during a period
of labor unrest to obtain information about the inspections.

Labor unrest is a vague term with many connotations and no precise
definition. To perform our analysis, we first identified the various
documented forms of unrest. Accordingly, we obtained information from the
only two sources that systematically collect and maintain information on
events that may be related to labor unrest--NLRB and FMCS. Using these
sources, we were able to identify establishments that experienced (1)
charges of unfair labor practices, (2) union campaigns to organize
employees, (3) efforts to decertify (remove) a union, or (4) work stoppages
(strikes). For purposes of this report, we define labor unrest as the
occurrence of one or more of these four conditions.

From NLRB we obtained a listing of all cases on file during fiscal years
1994 through 1998 that involved charges of unfair labor practices (either
against an employer or against a union), union efforts to organize
employees, and efforts to have a union decertified as a bargaining
representative of employees. This accounted for approximately 40,000 cases
each year. From FMCS we obtained a listing of all establishments that had a
work stoppage during fiscal years 1994 through 1998. This resulted in about
400 cases each year. These establishments may have had multiple cases on
file with NLRB during a fiscal year and may also have had a work stoppage
according to FMCS. As a result, we had to review manually each case in the
two agencies' databases to obtain an unduplicated count of establishments
for each fiscal year (see table 1).

              1994    1995    1996    1997   1998    Combined
 Cases listed
 NLRB         41,415  40,444  39,366  40,627 37,647  199,499
 FMCS         477     388     375     386    430     2,056
 Total        41,892  40,832  39,741  41,013 38,077  201,555
 Unduplicated count
 NLRB         21,811  21,927  20,724  21,978 20,499  106,939
 FMCS         275     211     208     224    222     1,140
 Total        22,086  22,138  20,932  22,202 20,721  108,079

Once we had an unduplicated list of establishments that experienced labor
unrest, we selected a random sample of 400 establishments from each of the 5
fiscal years. Our sampling procedure permitted us to generalize the sample
results to the entire universe of establishments experiencing labor unrest.

After we selected our sample establishments, we compared them with OSHA's
Integrated Management Information System (IMIS), which tracks OSHA
inspection activity, to determine whether an inspection was conducted at a
particular establishment. Because labor unrest can exist over a period of
time, we decided to include any OSHA inspections conducted at the sample
establishments 6 months before the earliest date on record with NLRB or FMCS
and 6 months after the latest date. For example, if NLRB recorded charges of
unfair labor practices filed against an employer on November 9, 1997, and
again on September 15, 1998, then we counted all OSHA inspections at this
establishment between May 9, 1997, and March 15, 1999, as a match.
Inspections before May 9, 1997, and after March 15, 1998, would be
considered outside the period of labor unrest. Using this methodology, we
determined that OSHA inspected 172 of the 2,000 sample establishments. We
then projected these results to the universe of establishments experiencing
labor unrest. As shown in table 2, we estimated that, of the approximately
108,000 establishments with labor unrest during this period, about 9,300
were inspected by OSHA (see table 2).

                            1994   1995    1996    1997    1998   Combined
 Sample size                400    400     400     400     400    2,000
 Number inspected           21     39      31      40      41     172
 Percentage inspecteda      5.3    9.8     7.8     10.0    10.3   8.6
 Number of establishments
 with labor unrest          22,086 22,138  20,932  22,202  20,721 108,079
 Estimated number
 inspecteda                 1,171  2,170   1,633   2,220   2,134  9,328

aBecause the yearly estimates of the percentage of establishments with labor
unrest that were inspected are based on samples, there is a sampling error
of roughly plus or minus 3 percentage points for each of those estimates at
the 95 percent confidence level. Thus, for example, there is a 95 percent
chance that in fiscal year 1998 the number of establishments with labor
unrest that were inspected was between 1,513 (7.3 percent) and 2,756 (13.3
percent). The difference between the rate of inspection among establishments
experiencing labor unrest shown in table 2 and the rate of inspections among
those without unrest shown in table 5 is statistically significant at the
.05 level for each fiscal year.

According to BLS, approximately 7 million establishments existed nationwide
each year from 1994 through 1998. We had already determined, using the NLRB
and FMCS database, that there were approximately 22,000 establishments each
year that experienced labor unrest. We then subtracted this number from the
total establishments identified by BLS (see table 3) to calculate the total
number of establishments without labor unrest.

                1994      1995      1996      1997     1998      Combined
 Total
 establishments 6,620,651 6,761,169 7,008,812 7,057,7947,354,744 34,803,170
 nationwide
 Number of
 establishments
 with labor     22,086    22,138    20,932    22,202   20,721    108,079
 unrest
 Number of
 establishments
 without labor  6,598,565 6,739,031 6,987,880 7,035,5927,334,023 34,695,091
 unrest

We then determined the percentage of establishments without labor unrest
that were inspected by OSHA. We first converted the number of inspections
conducted on a yearly basis (about 100,000) to establishments inspected, as
OSHA may conduct numerous inspections at a single establishment. To do so,
we identified the number of different addresses at which OSHA conducted
inspections during fiscal years 1994 through 1998, the method OSHA officials
suggested to approximate unique establishments inspected. We had already
identified the number of inspections OSHA conducted at establishments with
labor unrest, so we subtracted this number from the estimated number of
establishments at which OSHA conducted inspections (see table 4).

                       1994     1995    1996    1997   1998    Combined
 With unique addresses 106,076  93,369  84,165  94,462 94,106  472,178
 With labor unresta    1,171    2,170   1,633   2,220  2,134   9,328
 Without labor unrest  104,905  91,199  82,532  92,242 91,972  462,850

aSee note under table 2.

We then divided the number of OSHA inspections at establishments without
labor unrest for each fiscal year by the number of establishments nationwide
that did not have any labor unrest to determine the proportion of those
establishments that were inspected by OSHA (see table 5).

                 1994     1995      1996      1997      1998     Combined
 Number of
 establishments
 without labor   6,598,5656,739,031 6,987,880 7,035,592 7,334,02334,695,091
 unrest
 OSHA
 inspections at
 establishments  104,905  91,199    82,532    92,242    91,972   462,850
 without labor
 unrest
 Percentage of
 establishments
 without labor
 unrest          1.6      1.4       1.2       1.3       1.3      1.3
 inspected by
 OSHAa

aSee note under table 2.

Because we used a sampling procedure to estimate the OSHA inspection rate at
establishments experiencing labor unrest, we cannot identify whether a
particular establishment in OSHA's database that was not in our sample
experienced labor unrest. Thus, we could not obtain inspection
characteristics for those OSHA inspections conducted at establishments
without labor unrest. As a result, we can compare inspection characteristics
for those establishments that appeared in our sample only against all OSHA
inspections. Inspections at establishments experiencing labor unrest
comprise only about 2 percent of all OSHA inspections. The characteristics
we compared include:

ï¿½ reason for inspection,

ï¿½ union status of the establishment inspected,

ï¿½ size of the establishment inspected, and

ï¿½ industry designation of the inspected establishment.

Information on the first two variables is contained in the report. We found
that IMIS data on establishment size were of questionable reliability, so we
did not present information on this variable. For the last variable,
industry designation, we found that differences existed between the
percentage of various industries inspected by OSHA in general and during
times of labor unrest. We found that although OSHA was in general more
likely to inspect construction establishments, for establishments with labor
unrest, OSHA was more likely to inspect manufacturing establishments (see
fig. 4).

A number of different factors could have contributed to this. For example,
NLRB's database, the primary database we used to identify establishments
experiencing labor unrest, might contain primarily manufacturing
establishments. However, NLRB officials would not confirm whether
manufacturing establishments would be more apt to be in its database than
any other industry. Because we were unable to obtain an explanation for the
differences in inspection rates, we present this data for informational
purposes only.

Comments From the Department of Labor

GAO Contacts and Staff Acknowledgments

Lori Rectanus, (202) 512-9847
Wayne Sylvia, (617) 565-7492

In addition to those named above, the following persons made important
contributions to the report: Linda W. Choy, Joseph G. Evans, Lyle H. Lanier,
Jr., Kevin F. Murphy, and James P. Wright.

(205398)

Table 1: Number of Establishments Experiencing Labor Unrest,
by Source, Fiscal Years 1994-98 17

Table 2: Estimated Number of Establishments With Labor Unrest Inspected by
OSHA, Fiscal Years 1994-98 18

Table 3: Number of Establishments With and Without Labor Unrest,
Fiscal Years 1994-98 19

Table 4: Number of OSHA Inspections at Establishments With and
Without Labor Unrest, Fiscal Years 1994-98 19

Table 5: Percentage of Establishments Without Labor Unrest
Inspected by OSHA, Fiscal Years 1994-98 20

Figure 1: Inspection Rates for Establishments With and Without
Labor Unrest, Fiscal Years 1994-98 8

Figure 2: Reasons for Inspections at Establishments With Labor
Unrest and for All OSHA Inspections, Fiscal Years 1994-98 10

Figure 3: Inspections of Unionized and Nonunionized Establishments, With
Labor Unrest and for All OSHA Inspections,
Fiscal Years 1994-98 12

Figure 4: Percentage of Inspections by Industry for All
OSHA Inspections and Inspections at Establishments
With Labor Unrest, Fiscal Years 1994-98 21
  

1. A catastrophe is defined as the hospitalization of three or more workers
due to an accident or illness caused by a workplace hazard.

2. Our discussion of OSHA throughout this report includes the activities of
OSHA and its state partners.

3. Under the law, there must be "reasonable grounds to believe" that the
violation or danger complained of exists. See 29 U.S.C. 657(f)(1).

4. The inspection rate we developed for establishments without labor unrest
was based on the universe of about 7 million establishments minus those
establishments experiencing labor unrest. Although OSHA would be able to
perform an inspection at any of these establishments if a valid complaint
was filed or a fatality or catastrophe occurred, in actuality, some
establishments within this 7 million are more likely than others to be
inspected.

5. Because our sample only included establishments with labor unrest, we do
not have information on the characteristics of establishments without labor
unrest. Our comparisons of inspection characteristics, such as reason for
inspection, are between OSHA inspections for all establishments and
inspections of establishments experiencing labor unrest, which comprise
about 2 percent of all OSHA inspections.

6. States are allowed to operate their own safety and health programs as
long as OSHA determines that they are at least as rigorous as federal OSHA
programs.

7. These limitations on OSHA's authority have appeared regularly in OSHA's
yearly appropriation.

8. The information we present concerning inspection characteristics
illustrates the association between labor unrest and OSHA inspections; it
does not demonstrate a causal relationship between the two. There are
numerous factors that may affect inspection rates and labor unrest that are
outside the scope of this review.

9. As pointed out by OSHA officials, OSHA's policy of targeting high-hazard
industries means that some establishments within this 7 million are more
likely than others to be inspected. We recognize that if we were to reduce
that figure to include only those establishments OSHA officials say are most
likely to be inspected, this would increase the inspection rate for
establishments not experiencing labor unrest. However, OSHA has the
authority inspect any of these establishments based on a complaint.

10. Because the yearly estimates of the percentage of establishments with
labor unrest that were inspected are based on samples, there is a sampling
error of roughly plus or minus 3 percentage points surrounding each of those
estimates at the 95 percent confidence level. The difference between the
rate of inspection among establishments experiencing labor unrest and the
rate of inspections among those without unrest is statistically significant
at the .05 level for each fiscal year.

11. California OSHA's policy is to respond to serious work-related accidents
in the same manner as federal OSHA responds to fatalities and catastrophes.
In the case of nonserious work-related accidents, its inspection policy is
discretionary.

12. Information is not available on the percentage of establishments
nationwide that are unionized. For comparison, however, about 10 percent of
all private workers were represented by a union in 1998, ranging from 27
percent in the transportation and public utilities industry to about 2
percent in the agriculture industry.

13. David Weil, "Building Safety: The Role of Construction Unions in the
Enforcement of OSHA," Journal of Labor Research, Vol. 13, No. 1 (1992), and
"Enforcing OSHA: The Role of Labor Union," Industrial Relations, Vol. 30,
No. 1 (1991).
*** End of document. ***