Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 26
2
Alternative Sources of Wage Data
The charge to this panel included a request to “evaluate currently
available and potential data sources” for measuring and collecting pay
information from U.S. employers for the purpose of administering Section
709 of the Civil Rights Act of 1964. We begin our response to this part of
the charge with a discussion of the collection of earnings data from public-
sector employers on the EEO [equal employment opportunity] form 4, or
EEO-4. Indeed, the Equal Employment Opportunity Commission (EEOC)
has some experience from which to draw when considering the collection
of earnings data because the agency now collects wage band information
on the EEO-4 form.
We also discuss other possible sources of wage information and the
experiences of other agencies in collecting such information.1 We first con-
sider the capacity of existing federal administrative data series that include
earnings information from employers to meet a requirement for wage
information by gender, race, and national origin. If these administrative
data, mostly from tax collections, could suffice to provide the necessary
wage data for use in antidiscrimination enforcement, a new data collection
process could be avoided. Unfortunately, as discussed in this chapter, the
use of administrative data is not a promising path because of data incom-
pleteness and uncertain quality.
1 his report does not assess another data source that has appeared recently in which
T
individual employees self-report pay by employer, occupation, and location on a variety of
websites; these self-postings sometimes include pay stubs. These self-reports are not a random
sample, offer little or no demographic information, have variable or in many cases no coverage
of occupations, and are difficult to verify.
26
OCR for page 27
ALTERNATIVE SOURCES OF WAGE DATA 27
We then consider the experience of the Office of Federal Contract Com-
pliance Programs (OFCCP) of the U.S. Department of Labor (DOL) with
collection of earnings information on a trial basis a decade ago. The lessons
learned in that experiment should be considered by EEOC as it considers
collecting earnings information.
We also discuss the data collection programs of the states of New
Mexico and Minnesota and the Canadian province of Ontario. These ju-
risdictions now gather earnings information from employers for pay equity
purposes. We assess the potential of these collections to inform an EEOC
decision on whether and how to collect earnings information.
Finally, we consider survey-based wage information and discuss three
Bureau of Labor Statistics (BLS) surveys—the Current Employment Statis-
tics (CES) Survey, the National Compensation Survey (NCS), and the Oc-
cupational Employment Statistics (OES) Survey. These surveys can inform
the collection of wage data and provide a source of potential validation
information for data series that could be collected by EEOC, but we do
not judge them to be suitable sources for the wage data for EEO enforce-
ment purposes. They do not collect data by gender, race, or national origin;
they are covered by strict confidentiality provisions, which limit their use
for enforcement; and they do not cover all establishments covered by EEO
laws and executive orders.
DATA FROM EEO-4 REPORTS
As noted in Chapter 1, EEO-4 reports are collected in odd-numbered
years from state and local governments: in 2009 approximately 6,000
jurisdictions filed EEO-4 reports that covered 3,238,769 employees. The
report collects employment data by job group and salary ranges for race/
ethnicity and gender, with separate reports by function (e.g., streets and
highways, health, corrections). Data are also collected separately for part-
time employees and new hires.
The EEO-4 report is the only one that collects any wage-related data.
It collects annual salaries by job category for eight pay bands:
1. $1,000 to $15,999
2. $16,000 to $19,999
3. $20,000 to $24,999
4. $25,000 to $32,999
5. $33,000 to $42,999
6. $43,000 to $54,999
7. $55,000 to $69,999
8. $70,000 and over
OCR for page 28
28 COLLECTING COMPENSATION DATA FROM EMPLOYERS
The pay band data are collected for eight job categories:
1. officials and administrators
2. professionals
3. technicians
4. protective service workers
5. paraprofessionals
6. administrative support
7. skilled craft workers
8. service and maintenance workers
The wage data collected on this report have some limitations, accord-
ing to EEOC commissioner Stuart Ishimiru, who addressed the panel on
May 24, 2011. The form requests wage data by race, ethnic origin, and
gender, but the wages are reported in broad intervals that do not allow
for precise comparisons. Similarly, according to the commissioner, the job
categories for which wages are reported are so broad that they are rarely if
ever used to conduct wage disparity analyses. Despite these limitations, the
reports are used extensively by the U.S. Department of Justice (DOJ) for
administrative and enforcement purposes. Academic institutions use these
reports for self-assessment purposes.
ADMINISTRATIVE DATA
The federal government and state agencies now collect a massive
amount of wage data from employers and maintain them in the form of
administrative records of three tax systems. Two of these systems are ad-
ministered by federal agencies—the Internal Revenue Service (IRS) and the
Social Security Administration (SSA)—and one by state unemployment in-
surance agencies under the auspices of the DOL’s Employment and Training
Administration (for details, see Greenia, Appendix B of this volume). The
three administrative data systems are used primarily to collect taxes and
determine benefits for the purposes of administering and funding the federal
income tax system (by the IRS), the Social Security and Medicare programs
(by SSA), and the joint state-federal unemployment insurance (UI) system.
The data are used by the programs that collect them for purposes of
enforcement of their own laws and regulations. In select circumstances,
federal legislation has also authorized use of these data for enforcement
purposes in other programs. For example, a new hires database derived
from UI filings is used by the Administration for Children and Families in
the U.S. Department of Health and Human Services to facilitate finding
employed parents who are not making required child support payments
OCR for page 29
ALTERNATIVE SOURCES OF WAGE DATA 29
under the Personal Responsibility and Work Opportunity Reconciliation
Act of 1996.2
National compilations of statistics are produced from the three sets of
data by the pertinent statistical offices of IRS and SSA, as well as the Bureau
of Labor Statistics (BLS).3 In addition, the data are used for policy analysis
by the Joint Committee on Taxation of Congress, the Congressional Budget
Office, and the Office of Tax Analysis in the U.S. Department of the Trea-
sury. The data are also used for analysis by academic researchers, through
the Intergovernmental Personnel Act, as well as through the U.S. Census
Bureau’s Research Data Centers. Table 2-1 summarizes the availability of
items from each of these administrative records sources.
According to Greenia (Appendix B of this volume), the three sets of
data are interrelated. For example, the three tax-based systems depend on
the social security numbers (SSNs) assigned by SSA, the employer identifi-
cation numbers (EINs) assigned by IRS, the reporting of employment and
payroll at both the firm and individual worker level for federal and state
purposes, and other information from the administrative systems, such as
changes in name and address, to update the records.
The IRS has the duty to determine which workers are employees and
which are contractors. “The IRS decision is obtained by the filing of a
Form SS-8 for a firm or worker seeking to have IRS establish officially the
employee or independent contractor status of a particular worker. This
transaction then has ramifications for the other employee data collection
systems that are mandated by such legislation as the State Unemployment
Tax Act (SUTA) and the Federal Unemployment Tax Act (FUTA)” (Greenia,
Appendix B of this volume).
Thus, although only the SSA system has data on earnings by gender,
race, and national origin (items needed for enforcement purposes), it is pos-
sible, by virtue of their coverage and interrelationships, to link data from
the three tax systems so that each of them could produce some data on
employee earnings by gender, race and ethnicity, nativity, and age, by em-
ployer. These data could be used to inform EEOC’s enforcement programs,
although they most likely could not be used directly in enforcement actions.
2 or
F details, see http://www.acf.hhs.gov/programs/cse/newhire/library/ndnh/background_
guide.htm [July 2012].
3 RS data are primarily published by the Statistics of Income Division of IRS: see http://
I
www.irs.gov/taxstats/productsandpubs/article/0,,id=125133,00.html [July 2012]. SSA data are
published by the Office of Retirement and Disability Policy: see http://www.ssa.gov/policy/
docs/statcomps [July 2012]. And BLS data are published in the Quarterly Census of Employ-
ment and Wages series: see http://www.bls.gov/cew/cewbultn10.htm [July 2012]. The Census
Bureau also uses these data sets as input to several of its statistical programs.
OCR for page 30
30
TABLE 2-1 Available Items in Administrative Records Relevant to EEO
Earnings at Identity of Employee Employee Employee
Source Employee Level Employer Gender Race/Ethnicity Nativity
State Unemployment Insurance YES YES NO NO NO
State Employment Security Agency NO YES NO NO NO
Internal Revenue Service YES YES NO NO YESa
Social Security Administration YES YES YES YES YES
aOnly from individual taxpayer identification number (ITIN) applications.
SOURCE: Adapted from Greenia (Appendix B of this volume).
OCR for page 31
ALTERNATIVE SOURCES OF WAGE DATA 31
State Unemployment Insurance Data
In addition to complying with the Federal Unemployment Tax Act,
employers must also comply with the State Unemployment Tax Act by with-
holding and depositing tax or insurance payments from each employee’s
wages with state unemployment offices. These state unemployment taxes
fund unemployment benefits in each state or territory (including the District
of Columbia, Puerto Rico, and the Virgin Islands; see Greenia, Appendix
B of this volume).
This section presents a brief summary of the UI wage records and the
Quarterly Census of Employment and Wages (QCEW) Program that draws
on them. It discusses how the UI data are reported, collected, and shared
with the federal government, and assesses the potential usefulness of these
data for EEO enforcement purposes.
UI tax rates and coverage vary by state, as do the content and format
of the records a particular state collects. In general, all workers are covered
by the UI system with the exception of federal employees, independent con-
tractors, the self-employed, and some agricultural workers. A state collects
detailed employment and compensation data in quarterly reports from each
employer. The data include the SSN, name, and quarterly compensation
for each individual employee, as well as the employer name and EIN.4 The
products of this collection are known as UI wage records.
State employment security agencies also collect aggregate monthly em-
ployment (for the pay period containing the 12th of the month) for each
quarter and aggregate quarterly employee compensation from each em-
ployer in the state covered by state UI laws and for federal workers cov-
ered by the Unemployment Compensation for Federal Employees (UCFE)
Program. This data collection program, the QCEW, is administered and
partially funded by BLS.
Although states request data from employers at the establishment level
for multiple worksites or multi-establishment employers, there is no disin-
centive for an employer that does not comply with the request as long as
total employment is reported accurately and the appropriate amount of UI
tax is paid to the states (Greenia, Appendix B of this volume).
In considering wage data for purposes of EEO enforcement, the UI data
system provides the earnings data needed and at the employee level, but it
also has several shortfalls:
• It is difficult, if not impossible, to disaggregate the data from multi-
establishment employers to the worksite level to match with the
EEO-1 reports (see Chapter 1).
4 he
T coverage varies by state; see Stevens (2002) for a complete review.
OCR for page 32
32 COLLECTING COMPENSATION DATA FROM EMPLOYERS
• There are no gender, race and ethnicity, or nativity data collected
for UI wage records, though there have been instances in which
demographic data from other sources, such as driver’s licenses files,
have been associated with the wage records (Glover, 2011; Moore,
2011) to enable analysis of UI wage information by gender. As
discussed below, it would be possible to match these records to SSA
demographic data.
• In order to obtain either of the two data components provided
to the states by employers—especially the detailed employee
earnings—it would be necessary to obtain the data directly from
employers (who would submit a copy of their UI filings to EEOC)
or to enter into separate agreements with each state, and it is likely
that both of these actions would require a legal action.
Internal Revenue Service Data
Since 1976, when the current simplified Combined Annual Wage Re-
porting (CAWR) program was established by the Tax Reform Act, employ-
ers have reported individual earnings statements and the amount of taxes
withheld (including federal income tax, Social Security tax, and Medicare
tax) on a single form (Form W-2 Wage and Tax Statement) for both IRS and
SSA purposes. The earnings details available from the W-2 are rich: wages
and salaries, deferred compensation (part of total compensation, even if not
taxable currently), and certain fringe benefits are reported, in addition to
capped Social Security earnings and uncapped Medicare earnings. Together,
the W-2 earnings variables provide a unique and comprehensive window
on earnings data at the employee level.
These individual W-2 forms are transmitted with another form (Form
W-3, Transmittal of Income and Tax Statements), which cumulates the in-
formation from the W-2 forms for each reporting establishment. Because of
this arrangement, it would be possible to obtain detailed annual employee
compensation, quarterly and annual aggregate employee compensation, and
number of employees at both the employee and employer level with links
to Social Security information through an SSN and EIN crosswalk. The
industry codes available at SSA, in full North American Industry Classifica-
tion System (NAICS) levels, can provide a further source of rich classifier
information on employers’ business activities. In addition, other tax forms
can provide various components of aggregate and even detailed employee
compensation: for example, compensation to corporate officers. Finally,
EIN and individual taxpayer identification numbers (ITINs) assignment
and other transactions would enable the tracking of new business births,
foreign-born workers without SSNs, and even the employee or contractor
status of a worker.
OCR for page 33
ALTERNATIVE SOURCES OF WAGE DATA 33
An employer is required to file an annual FUTA tax return (Form 940)5
for purposes of reporting and paying the federal unemployment taxes re-
quired by FUTA. Filing is required—at the aggregate employment level—for
each nonagricultural employee earning at least $1,500 in any quarter of
the year or for each employee who was employed for part or all of a day
in any 20 different weeks of the year.6 Although Form 940 does report
annual total compensation, it does not report the number of employees.
However, the compensation information may be useful for benchmarking
compensation data reported on other federal tax forms, such as Form W-2
and Form 941, as well as the UI data.
In summary, IRS data include a wealth of earnings information for
individual employees and employers. However, a limitation is that the IRS
data include establishment data only when the establishment is also an en-
terprise (and has an EIN). Another limitation is that the tax data contain no
information by gender (except, sporadically, for the IRS Statistics of Income
Division individual Form 1040 tax sample), race and ethnicity, or nativity
(except for ITIN applications).
Social Security Administration Data7
The data of most interest for examining pay equity issues are the de-
mographic data that are available on the application for a Social Security
Number (Form SS-5),8 which can be linked to federal tax data shared by
IRS. The application for an SSN captures gender, race and ethnicity, and
nativity—often shortly after birth for most U.S. citizens. In addition, it
captures citizenship status, which might be used as a proxy for or to supple-
ment nativity information.
Although the Form SS-5 data are self-reported (by the individual or a
parent), SSA uses supporting documentation for verification, particularly
for changes, such as a marriage license (name), passport (citizenship), and
birth certificate (place of birth). The Form SS-5 data, including updates,
are maintained in SSA’s Numerical Identification System file, referred to as
the Numident file.
Despite the richness of the demographic detail, the Numident file data
have some limitations. They are not updated as often as tax information
for such changes as name and address due to marriage or divorce (the tax
5 he
T form is available at: http://www.irs.gov/pub/irs-pdf/f940.pdf [December 2011].
6 or
F 2009 and 2010, agricultural employers were required to file if they paid cash wages
of $20,000 or more to farm workers during any calendar quarter or if they employed 10 or
more farm workers during some part of the day (whether or not at the same time) during any
20 or more different weeks in either year.
7 nformation in this section is based largely on Greenia (Appendix B of this volume).
I
8 his form is available at: http://www.ssa.gov/online/ss-5.pdf [July 2012].
T
OCR for page 34
34 COLLECTING COMPENSATION DATA FROM EMPLOYERS
information at IRS may be updated before the Numident data). In addition,
although nativity data classified by country might be considered relatively
reliable, researchers have noted that some of the “foreign born” may be,
in fact, the progeny of U.S. citizens, say, for military and other Americans
stationed overseas, where birth occurs. In conjunction with citizenship
status, however, the data are probably useful for indicating native versus
foreign-born status.
EQUAL OPPORTUNITY SURVEY PILOT
In order to identify federal contractors with potential problems of pay
discrimination that could warrant further review or evaluation by OFCCP
or to support a contractor self-audit, OFCCP has long been interested in de-
veloping a screening tool to enable the agency to identify supply and service
contractors whose compensation data indicate that further investigation is
warranted. This interest led to initiation of a pilot survey to collect earnings
data with demographic and job group information from federal govern-
ment contractors. An employer survey was developed and undertaken by
the OFCCP. The OFCCP experience is instructive for EEOC as it considers
collecting wage information by gender, race, and national origin.
As discussed in Chapter 1, the authority for this collection rests in
Executive Order 11246, as amended, which requires that federal govern-
ment contractors and subcontractors “take affirmative action to ensure that
applicants are employed, and that employees are treated during employ-
ment, without regard to their race, color, religion, sex, or national origin.”
Affirmative action under the executive order requires that contractors take
affirmative steps to identify and eliminate impediments to equal employ-
ment opportunity. The affirmative steps include numerous record-keeping
obligations designed, first, to assist the contractor and then OFCCP in
monitoring the contractor’s employment practices.9
In the early 2000s, the OFCCP listed three objectives for the survey
(71 FR 3374):
1. to improve the deployment of scarce federal government resources
toward contractors most likely to be out of compliance;
2. to increase agency efficiency by building on the tiered-review pro-
cess already accomplished by OFCCP’s regulatory reform efforts,
thereby allowing better resource allocation; and
3. to increase compliance with equal opportunity requirements by im-
proving contractor self-awareness and encourage self-evaluations.
9 or
F full text of Executive Order 11246, as amended, see http://www.dol.gov/ofccp/regs/
statutes/eo11246.htm [July 2012].
OCR for page 35
ALTERNATIVE SOURCES OF WAGE DATA 35
Field testing for the survey of federal contractors to collect wage in-
formation, as well as other new data items, was conducted in 1999. In
2000, OFCCP issued a requirement that nonconstruction contractor es-
tablishments designated by OFCCP prepare and file the new Equal Op-
portunity Survey. On a pilot basis, in April 2000, the EO Survey was sent
to 7,000 contractors. One part of the survey (Part C) collected data on
monetary compensation (expressed as an annual amount) and on tenure
for four groups—minority females, nonminority females, minority males,
and nonminority males—by the EEO-1 report categories applicable at that
time: (1) officials and managers; (2) professionals; (3) technicians; (4) sales
workers; (5) office and clerical workers; (6) craft workers; (7) operatives;
(8) laborers; and (9) service workers. The questionnaire instructions de-
fined annual monetary compensation as “an employee’s base rate (wage or
salary), plus other earnings such as cost-of-living allowance, hazard pay,
or other increment paid to all employees regardless of tenure on the job,
extrapolated and expressed in terms of a full year.”10 The annual monetary
compensation measure was not to include the value of benefits, overtime,
or one-time payments, such as relocation expenses.
The survey did obtain annual monetary compensation information—98.3
percent of respondents provided a numerical response to the compensation
item. Reported median average annual compensation by gender and oc-
cupation appeared to be “broadly consistent” with other well-established
data sets, such as the decennial census, the Current Population Survey, and
other salary surveys (Bendick, 2000, p. 9).
After receipt of pilot survey responses, OFCCP commissioned a study
to determine whether the pilot survey results could be used to predict
whether a contractor would have findings of noncompliance. The study
concluded, based on the first wave of survey responses, that the survey
could contribute to improvements in procedures for selecting establish-
ments for compliance evaluations (Bendick, 2000, p. i).
The OFCCP proceeded with the EO Survey that was sent to contractors
beginning in December 2000 and continuing to December 2004. It included
information, in summary form, about personnel activities, compensation,
and tenure, as well as the contractor’s affirmative action program. A total
of 53,000 forms were sent.
To assess the quality and usefulness of these data, the OFCCP en-
gaged an outside contractor to evaluate the collection to that point. The
evaluation criteria were based on comparisons of survey results with the
results of OFCCP compliance evaluations of a sample of supply and service
contractor establishments that had completed the 2002 EO Survey. The
10 .S.
U Department of Labor form, available at: http://www.management-advantage.com/
media/eosurvey.pdf [July 2012].
OCR for page 36
36 COLLECTING COMPENSATION DATA FROM EMPLOYERS
comparison study focused on 1,888 establishments that had completed
compliance reviews and had reliable EO Survey data. Of these 1,888 cases,
OFCCP found systemic discrimination in 67 cases (3.5 percent). Results
of the compliance reviews and survey data were analyzed to determine
whether a model could be developed that would predict which contractors
in the sample were engaged in systemic discrimination based solely on the
EO Survey data submitted (Abt Associates, Inc., 2005, pp. 23–37).
Based on that evaluation, OFCCP concluded that the EO Survey did
not improve deployment of enforcement resources toward contractors most
likely to be out of compliance and did not lead to greater self-awareness or
encourage self-evaluations. OFCCP further concluded that the information
in the survey largely duplicated information gathered in compliance visits,
although that finding does not necessarily undercut the potential value of
the survey, given that the purpose of such a survey is to obtain similar data
as those gathered in compliance visits for the purpose of targeting enforce-
ment resources.
The evaluation also found that the EO Survey imposed a burden on
respondents. Each survey form was estimated to take each respondent 21
hours to complete. Based on an estimated 10,000 respondents per year, the
EO Survey was estimated to cost contractor establishments 210,000 hours
per year. Using data from the BLS 2004 National Compensation Survey,
the total annual cost imposed on the regulated community by the survey
requirements was close to $6 million. However, whether this level of burden
was large or small compared with other regulatory requirements was not
established, nor was the burden considered in relationship to the costs to
employees of instances of wage discrimination that the survey might help
uncover.
OFCCP’s bottom-line conclusion was that the EO Survey had failed
to provide the utility anticipated when the regulation was promulgated in
2000, and, consequently, it eliminated the survey. Reinstatement of the EO
Survey, or the establishment of a similar survey, would require regulation
or legislation.
Yet Bendick (2006) pointed out that the data on which both the endick
B
(2000) and Abt Associates, Inc. (2005) studies drew had limitations that
make it difficult to reach definitive conclusions about the value—or lack
of value—of the EEO Survey for targeting enforcement resources. For
example, the survey data were contaminated by the fact that compliance
evaluations were conducted for many of the employers in the sample before
the survey, so that employers had the opportunity to improve their practices
by the time the survey was fielded.
The current OFCCP initiative is summarized by the National Equal
Pay Enforcement Task Force (2010, p. 5):
OCR for page 37
ALTERNATIVE SOURCES OF WAGE DATA 37
Through publication of an Advance Notice of Proposed Rulemaking
( NPRM) [in 2010],11 OFCCP [has sought] the input of stakeholders
A
in evaluating whether the EO Survey should be redesigned to collect
d
ifferent data than previously sought, and whether there are any ways to
further limit the burden of data collection for employers. The implemen-
tation of a redesigned survey is expected to result in better identification
of those contractors who are likely to be out of compliance, particularly
with regard to compensation discrimination; a narrowing of the issues on
which the resulting review will focus; and identification of contactors for
corporation-wide and industry-focused reviews.12
U.S. STATE AND CANADIAN PROVINCIAL SURVEYS
On their own initiative, two U.S. states and a Canadian province have
designed and fielded data collections to provide information on earnings
by demographic characteristics for pay equity purposes. Though small and
localized and based on comparable worth comparisons and thus not fully
responsive to the needs of the EEOC, these initiatives provide experience
in evaluating the feasibility of collecting wage information for antidis-
crimination enforcement purposes. The lessons from these initiatives are
summarized in this section.
Earnings Data Collection in New Mexico
As reported by Martha Burk in a presentation to the panel, in 2003
the New Mexico state legislature created a Pay Equity Task Force to study
wage disparities between men and women and between minorities and non-
minorities in both the public and private sectors. The task force issued a
report with numerous recommendations, and, in January 2009 Governor
Bill Richardson issued an executive order declaring pay equity a priority
for the state.13
In September 2009, the Pay Equity Task Force issued a report ad-
dressing pay gaps and job segregation in the state workforce and in the
workforce of state contractors. With respect to the latter, the report pro-
vided a rationale and model for requiring entities receiving state contracts
to submit pay gap reports as a condition of contracting.14 In December
2009, Governor Richardson issued a second executive order directing that
11 he
T ANPRM was published as 76 FR 49399.
12 he
T report further notes on p. 5 that “the EEO Survey has been rescinded, and its
reinstatement, or the establishment of a similar survey, must be by regulation or legislation.”
13
For details, see New Mexico Pay Equity Initiative, available at: www.generalservices.state.
nm.us/spd/pay_e.html [July 2012], p. 2.
14
Available at: http://www.generalservices.state.nm.us/spd/report093009.pdf [July 2012].
OCR for page 38
38 COLLECTING COMPENSATION DATA FROM EMPLOYERS
a contractor reporting system be implemented and appointing a working
group to design and facilitate such reporting.15
The new requirements were phased in gradually and, beginning July 1,
2010, all recipients of state contracts were required to submit a gender pay
equity report after a contract was awarded. More recently, these reports
have been required to be submitted as part of the response to a solicitation
or request for proposal (RFP). The reporting requirement applies to all state
agencies that let contracts and all categories of purchasing.
In her testimony to the panel, Burk stated that the early experience with
collecting these data is noteworthy for the lack of resistance by employers
and the absence of reports of difficulties in complying. About 3,200 firms
are covered by the reporting requirement, ranging in size from Intel (New
Mexico’s largest employer, with more than 3,000 employees) to firms with
only 10 employees. Over the first 7 months of implementation, fewer than
50 contractors contacted the state for assistance in understanding the re-
quirement or preparing their reports.16 Contractors seem generally to have
accepted the requirement as a normal part of the contracting process and
have lodged no complaints about the requirement.
As of early 2011, cataloguing of contractor reports was in progress,
and systematic analysis had not begun. However, cursory examination
of selected reports indicated that a majority of reporting employers have
employees in only three or four of the nine EEO-1 form job categories
provided for on the reporting forms, and a few (e.g., janitorial compa-
nies) have employees in only one job category besides the owner/manager.
Moreover, job segregation by gender is not unusual, in which case gender
pay disparities cannot be computed because of lack of wage data for both
genders. When pay comparisons can be made, percentage gender pay gaps
tend to range as low as 2–3 percent, with most in the 10–25 percent range.
An unusual few were observed as high as 45 percent.
The report uses EEO-1 job categories for reporting because many
contractors and payroll processing firms are already familiar with these cat-
egories. Using them avoided the need for a new taxonomy and also avoided
the difficulties of analyzing data for job titles or groupings that were not
comparable across firms.
The data each contractor is required to report consist of the number of
employees by gender (including full- and part-time workers) in each EEO-1
15 xecutive Order 2009-049, available at: http://www.generalservices.state.nm.us/spd/pay_e.
E
html [July 2012].
16 n addition to offering “live” assistance by telephone, the state provides easy access on
I
a website (see http://www.generalservvices.state.nm.us/spd/pay_e.html [July 2012]) to docu-
ments, including worksheets with instructions and reporting forms with instructions.
OCR for page 39
ALTERNATIVE SOURCES OF WAGE DATA 39
job category and the gender pay gap (stated as a percentage) in each cat-
egory. Individual compensation is not reported.
Uniform reporting is enhanced by the fact that average hourly wages
by gender and job category, taking into account hours worked, are com-
puted following detailed instructions for producing these averages and
entered into a worksheet. (Average hourly wages in each job category are
computed by dividing the total compensation by gender by the total hours
worked by that gender.)
Employers generally enter the appropriate numbers in the worksheet
by use of an accounting/payroll system that is capable of classifying em-
ployees and aggregating compensation and hours worked by gender and
job category. If employers do not have such a system to classify employees
by job category, gender, time worked, and compensation levels, the state
has provided an alternative downloadable employee data entry spread-
sheet for performing the necessary calculations “from scratch.” Using this
spreadsheet, employers enter employee identification, job category, gender,
whether full or part time, total annual compensation, and total hours
worked for each employee.
Standard formulas for computing the gaps are embedded in required
spreadsheets, which are provided to employers online, and the results of the
computations are exported to a standard final report format. To maintain
confidentiality of the wage and gender information, contractors do not turn
in worksheets showing dollar amounts, but instead report only the ratios
of average earnings for women to those of men in the same EEO-1 job
category. Proprietary information is retained by contractors. However, they
are encouraged to use this information for internal tracking of potential
compensation disparities between women and men.
The Office of the State Auditor has oversight over state agencies’ imple-
mentation of the reporting requirement and over the reports submitted.
Procedures for auditing are still under development, and no audits have yet
taken place. To date, the requirement is simply to submit a report; however,
bids that fail to comply with the reporting requirement are disqualified.
Minnesota Pay Equity Survey
Since 1982 the state of Minnesota has had a pay equity law for state
employees based on the concept of comparable worth. The law was ex-
tended to all city, county, school, and other public jurisdictions by the
1984 Local Government Pay Equity Act. The 1984 law requires each lo-
cal government jurisdiction to use a job evaluation system to determine
comparable work value and to submit a report to the state government
at 3-year intervals with a comparable worth value estimate (the value of
work as measured by skill, effort, responsibility, and working conditions)
OCR for page 40
40 COLLECTING COMPENSATION DATA FROM EMPLOYERS
and the minimum and maximum monthly salary for the job class by gen-
der. The state uses the reports to assess how well the local jurisdictions are
complying with the state law. Pay equity laws in Minnesota address only
gender-based wage disparities.
Ontario Pay Equity Survey
A relatively recent pay equity survey in Ontario, Canada, collects wage
information from the province’s relatively large employers. The collection
was enabled by a 2009 amendment to the Ontario Pay Equity Act, which
provides that the Pay Equity Office may collect information for the purpose
of providing reports to the Minister of Labour. The program was launched
in January 2011. It involves canvassing all Ontario employers using a
simple form that employers populate to provide current compensation data:
see Table 2-2. The raw data are submitted to the Pay Equity Office, which
assesses the data to determine if a wage gap exists.
For the first phase, Ontario workplaces with more than 500 employees
were selected to enter the program. Using lists of Ontario employers de-
veloped by an external provider, employers who had not been visited by a
review officer in the past 10 years and were not unionized were requested
to submit current, basic wage data on the positions and incumbents in their
organizations.
The information is analyzed to determine whether a wage gap appears
to exist against a set of criteria developed by seasoned review officers. A
committee of review officers meets monthly to review the analyses and
findings and finalize the assessment. The employer of the establishment is
advised whether the review officer determined there is an apparent wage
gap and is provided with tools and information to allow the employer to
consider whether they are pay equity compliant.
TABLE 2-2 Ontario Pay Equity Form
Pay as of
Employee December
(may use 31, 2010— Salary Range
Job Title/ symbol rather Held by: hourly, weekly, of Position Years of
Position than name) Male/Female annually (if applicable) Service
OCR for page 41
ALTERNATIVE SOURCES OF WAGE DATA 41
The response rate for the first mailings was about 80 percent. This
relatively high response rate was attained by significant follow-up efforts;
in addition, response is encouraged because nonresponders are singled out
for investigation through a proactive monitoring program for compliance.
SURVEY-BASED WAGE INFORMATION
In this section we consider the experience of surveys that collect wage
information, specifically, three BLS employer surveys: the National Com-
pensation Survey (NCS), the Current Employment Statistics (CES) survey,
and the Occupational Employment Statistics (OES) survey.
We do not discuss Census Bureau data sources. The Census Bureau’s
business surveys have extensive establishment coverage but do not col-
lect wage or demographic information. The decennial population census
captures data on gender, race, and ethnicity, but, of course, only every
10 years and without socioeconomic establishment detail. The Current
Population Survey and the American Community Survey collect wage data
by gender, race/ethnicity, nativity (native/foreign born), and many other
characteristics, but not by establishment. The Census Bureau’s Longitudinal
Employer-Household Dynamics (LEHD) System links UI and QCEW data
on employers and employees (obtained through individual agreements with
states) with additional employer and employee data from censuses and sur-
veys. The data, which include wage information, are available only to quali-
fied researchers at one of the Census Bureau’s Research Data Centers.17
All information collected by the federal government for statistical pur-
poses, including the data in these three BLS surveys, is collected under a
pledge of confidentiality according to the provisions of the 2002 Confiden-
tial Information Protection and Statistical Efficiency Act (CIPSEA). This
means that the data cannot be shared for purposes of antidiscrimination
enforcement; however, the information may be used to assist in analysis
relevant to wage discrimination, and the ability of the survey to collect
wage information may be instructive for EEOC.
National Compensation Survey
The NCS is an establishment-based survey that annually provides esti-
mates of occupational earnings, employer costs for employee compensation,
compensation trends, wages in one geographic area relative to other geo-
graphic areas, the incidence of employer-provided benefits among workers,
and provisions of employer-provided benefit plans. The employment cost
17 or
F details, see http://lehd.did.census.gov/led/ [July 2012].
OCR for page 42
42 COLLECTING COMPENSATION DATA FROM EMPLOYERS
index (ECI)—a principal federal economic indicator—is estimated from
data collected by the NCS.18
The NCS samples private industry establishments with one or more
workers and state and local governments across the 50 states and the
District of Columbia. Each sampled establishment—over 35,000 establish-
ments in 2010—is asked to report on selected occupations. As stated in the
BLS Handbook of Methods, major exclusions from the survey are workers
in federal and quasi-federal agencies, military personnel, agricultural work-
ers, workers in private households, the self-employed, volunteers, unpaid
workers, individuals receiving long-term disability compensation, and in-
dividuals working overseas. Currently, the NCS also excludes individuals
who set their own pay (e.g., proprietors, owners, major stockholders, and
partners in unincorporated firms) and family members being paid token
wages; however, these exclusions are being reevaluated (Bureau of Labor
Statistics, no date).
Among the products of the survey are estimated average hourly wages
for over 800 occupations in approximately 80 metropolitan and selected
nonmetropolitan localities, weekly and annual earnings and hours for full-
time workers, and earnings by work level that permit wage comparisons
across occupational groups. The survey collects no demographic detail,
however, and it is therefore not directly useful for analysis that might fa-
cilitate antidiscrimination enforcement.
Current Employment Statistics Survey
The CES is an establishment payroll survey that is based on a monthly
survey of approximately 141,000 businesses and government agencies rep-
resenting approximately 486,000 worksites throughout the United States.19
The primary statistics derived from the survey are monthly estimates of
employment, hours, and earnings for the nation, states, and major metro-
politan areas. Preliminary national estimates for a given reference month
are typically released on the third Friday after the conclusion of the refer-
ence week, which is the week that includes the 12th of the month.
National estimates of average weekly hours and average hourly earn-
ings are made for the private sector for all employees and for production
and nonsupervisory employees. Detail is available for about 750 industries.
Average weekly overtime hours in manufacturing are also available.
Hours and earnings are derived from reports of gross payrolls and cor-
responding paid hours. However, hours for salaried workers who may have
18 or
F details, see http://www.bls.gov/eci/# [July 2012].
19 nformation
I in this section is largely reproduced from http://www.bls.gov/ces/cescope.
htm [July 2012].
OCR for page 43
ALTERNATIVE SOURCES OF WAGE DATA 43
set compensation but volatility in their hours are often reported as standard
weekly hours rather than hours actually worked and paid. The payroll for
employees covered by the CES is reported before deductions of any kind,
for example, for Social Security, federal and state withholding tax, union
dues, or retirement plans. Included in the payroll reports is pay for over-
time, vacations, holidays, and sick leave paid directly by the firm. Bonuses,
commissions, and other types of nonwage cash payments are excluded un-
less they are earned and paid regularly (at least once a month). Employee
benefits paid by the employer, as well as in-kind payments, are excluded.
Total hours during the pay period include all hours worked (including
overtime hours), and hours paid for holidays, vacations, and sick leave.
Total hours differ from the concept of scheduled hours worked. Average
weekly hours reflect effects of numerous factors, such as unpaid absentee-
ism, labor turnover, part-time work, strikes, and fluctuations in work sched-
ules for economic reasons. Overtime hours in manufacturing are collected
when overtime premiums were paid and the hours were in excess of the
number of straight-time hours in a workday or workweek. No information
is collected by gender, race/ethnicity, or nativity.
Occupational Employment Statistics Survey
The OES survey is a semiannual mail survey designed to measure oc-
cupational employment and wage rates among full- and part-time wage and
salary workers in nonfarm establishments in the United States.20 The survey
does not include the self-employed, owners and partners in unincorporated
firms, household workers, or unpaid family workers.
The OES survey is a cooperative program between BLS and state work-
force agencies (SWAs). BLS funds the survey and provides the procedures
and technical support, while the SWAs collect most of the data.21
The OES is a very large survey. Its estimates are constructed from a
sample of about 1.2 million establishments grouped into six semiannual
panels over a 3-year period. Each year, forms are mailed to two panels of
approximately 200,000 establishments, one panel in May and the other
in November. Thus, for example, the May 2010 estimates were based on
responses from six panels—May 2010, November 2009, May 2009, No-
vember 2008, May 2008, and November 2007.
The overall national response rate for six panels is about 78 percent
based on establishments and 74 percent based on employment. The survey
covers all employer size classes, and response rates are actually higher among
smaller employers. The survey’s coverage is extensive—approximately 63
20 nformation
I in this section is largely reproduced from http://www.bls.gov/oes/ [July 2012].
21 ata
D for 180 large firms are collected directly by BLS.
OCR for page 44
44 COLLECTING COMPENSATION DATA FROM EMPLOYERS
TABLE 2-3 Occupational Employment Statistics by Number of
Employees, May 2010
Number of Employees Viable Sample Units Respondents
1–9 437,389 380,215
10–19 195,755 155,320
20–49 202,642 148,143
50–99 107,175 71,562
100–249 84,492 55,090
250–499 35,225 22,780
500–999 13,620 8,778
More than 999 8,747 6,346
NOTE: The respondents were establishments.
SOURCE: Data from Bureau of Labor Statistics special tabulation for the panel.
percent of total national employment is represented by the unweighted
employment of sampled establishments across all six semiannual panels.
The OES survey draws its sample from state UI files. The survey sample
is stratified by metropolitan and nonmetropolitan area, industry, and size.
To provide the most occupational coverage, larger employers are more
likely to be selected than smaller employers (see Table 2-3).
The data available from the OES include cross-industry occupational
employment and wage estimates for over 500 areas, including the nation,
states, and the District of Columbia, metropolitan statistical areas (MSAs),
metropolitan divisions (the result of MSA subdivisions) non etropolitan
m
areas, and territories; national industry-specific estimates at the 2007
N
AICS 3-, 4-, and selected 5-digit industry levels; and national estimates
by ownership across all industries and for schools and hospitals (Bureau
of Labor Statistics, 2010a). No data are collected by gender, race/ethnicity,
or nativity.
The OES survey categorizes workers into nearly 800 detailed occupa-
tions based on the Office of Management and Budget’s Standard Occu-
pational Classification (SOC) system. The detailed occupations cover 22
of the 23 SOC major occupational groups. The May 2010 OES estimates
mark the first set of estimates based in part on data collected using the 2010
SOC system, and after May 2012, the OES data will reflect the full set of
detailed occupations in the 2010 SOC. Importantly, the 2010 SOC occupa-
tions will be capable of being cross-walked into the EEOC job categories
when EEOC completes an update of the crosswalk between the EEOC job
categories and the 2010 SOC.
OCR for page 45
ALTERNATIVE SOURCES OF WAGE DATA 45
SUMMARY
Several surveys have been developed specifically to measure pay dis-
crimination, and there are several survey-based and administrative records-
based sources of estimates of earnings. They vary widely in their approach
to measurement, their coverage of employers, and their content: for exam-
ple, only some of them collect demographic as well as earnings information.
Only two of the data sources for establishments contain information on
hours and whether the employee is on a temporary or permanent schedule,
and neither of those sources includes demographic information.
It is clear that there is no current source of earnings data that incorpo-
rates the demographic, occupation, work schedule, and employer informa-
tion necessary to support an antidiscrimination enforcement and analytical
program. A new reporting mechanism would have to be put in place to
produce earnings by gender, race, and ethnicity for establishments.
Nonetheless, the fact that earnings data are now generally reported to
the taxing authorities and to federal (and state) government statistical and
enforcement agencies suggests that it might be feasible to collect earnings
information by gender, race, and national origin in an EEOC data collec-
tion program. It also suggests that the EEOC may be able identify other
data collections that could serve as sources of benchmarks to assist in vali-
dating the information that might be collected as part of a new reporting
arrangement.
