BLS Handbook of Methods - Chapter 8 - NCS

Chapter 8
National Compensation Measures

T

he U.S. Bureau of Labor Statistics (BLS, the Bureau)
produces a diverse set of data from the National
Compensation Survey (NCS) program and the Work
Stoppages program.

IN THIS CHAPTER
Background ................................................................  1
Description .................................................................  1
Sample Design and Sampling Procedures .................  2
Data Collection ..........................................................  4
Occupational Selection and Classification .................  5
Weighting, Nonresponse Adjustment,
Imputation, and Benchmarking ..................................  7
Calculation and Reliability of the Estimates ..............  8
Uses and Limitations of the Data ............................... 18
Technical References ................................................. 19
Appendix: Major Work Stoppages Program .............. 22

The links to each of the NCS data products are as follows:
•

Employment cost levels and trends, http://www.bls
gov/ncs/ect/

•

Incidence of employer-provided benefits, http://
www.bls.gov/ncs/ebs/

•

Benefit plan provisions, http://www.bls.gov/ncs/ebs/

The link to Work Stoppages data products is http://www.
bls.gov/wsp

Background

the NCS will provide data by detailed worker characteristics,
such as work levels, union status, and part-time or full-time
work schedules.1

BLS collection and publication of wage data can be traced
back to the late 19th century, employee benefits data to the
mid-20th century. The NCS, introduced in 1996, collects a
broad range of compensation data that formerly had been
collected by three separate BLS programs. The Occupational
Compensation Survey program published national and local
area wage data for survey years 1991 through 1996. The
Employment Cost Index has been published since 1975.
The Employee Benefits Survey program started in 1979 and
collected and published data on employee benefits for survey
years 1980 through 1998. Until 2011, the NCS collected
data on employee compensation from a large sample
of establishments providing data on about 800 detailed
occupations in more than 150 local areas.
The NCS data are used to set pay levels of some federal
workers. With the enactment of the 2011 federal budget, the
Locality Pay Survey (LPS) portion of the NCS was eliminated.
Occupational data by locality will still be available through
the Occupational Employment Statistics (OES) program. To
continue to meet the requirement of the Federal Employees
Pay Comparability Act of 1990, data from the OES and NCS
programs will be used collaboratively. The OES program will
provide wage data by occupation for all localities nationwide;

Description
The NCS is an establishment-based survey that provides
comprehensive measures of (1) employer costs for employee
compensation, (2) compensation trends, and (3) the incidence
of employer-provided benefits among workers. The NCS also
collects data on provisions of selected employer-provided
benefit plans. The Employment Cost Index (ECI)—a
Principal Federal Economic Indicator—is estimated from
data collected by the NCS.
The NCS includes establishments with one or more
workers in private industry and in state and local
government, in the 50 States and the District of Columbia.
All workers covered by the survey are referred to by the
NCS as civilian workers. Major exclusions from the
  1 For the history of how BLS occupational wage surveys were used for federal pay comparability, see John E. Buckley, “Fifty Years of BLS surveys on
Federal employees’ pay,” Monthly Labor Review, September 2009, pp. 36–
46, http://www.bls.gov/opub/mlr/2009/09/art3full.pdf. For information
on historical NCS publications, link to Occupational Employment Statistics,
Employment Cost Index, Employment Cost Trends, and NCS Benefits.

1

In the NCS sample, certainty areas are any individual
areas with employment greater than 80 percent of the
sampling interval. (The sampling interval is the total
employment across all areas, divided by the number of areas
to be selected.) Certainty areas represent only themselves in
NCS local area estimates; smaller areas represent themselves
and other areas that are not part of the sample. The larger the
area, the greater is its chance of being selected.
In stage 1, certainty areas are identified, noncertainty areas
are stratified, and the remaining sample areas are selected
in accordance with the NCS Probability-Proportional-toSize (PPS) technique, in which local areas are allocated
approximately proportionally to total employment across
county clusters. These clusters consist of Metropolitan Areas,
Micropolitan Areas, and Outside Core Based Statistical Areas
(CBSAs) for each of the nine census divisions and for the
United States as a whole. Furthermore, the NCS combines
most contiguous Outside CBSA counties within the same
Census (area) division, to create clusters with employment
of at least 10,000 and with heterogeneous wage levels. The
result is a list of 436 county clusters formed from an original
list of 1,359 Outside CBSA counties. Stage 1 resulted in a
sampling of 152 areas, broken out as follows: 57 certainty
areas, 60 noncertainty Metropolitan Areas, 22 noncertainty
Micropolitan Areas, and 13 noncertainty Outside CBSA
county clusters.

survey are workers in federal and quasi-federal agencies,
military personnel, agricultural workers, workers in private
households, the self-employed, volunteers, unpaid workers,
individuals receiving long-term disability compensation,
and those working overseas. The NCS also excludes
individuals who set their own pay (e.g., proprietors, owners,
major stockholders, and partners in unincorporated firms)
and family members paid token wages.

Sample Design and Sampling Procedures
NCS data are collected from probability samples selected in
three stages: (1) a probability sample of geographic areas, (2)
a probability sample of establishments within sampled areas,
and (3) a probability sample of occupations within sampled
establishments. Probability samples are subject to sampling
error and nonsampling error, which are discussed under the
section “Calculation and Reliability of the Estimates.”
In February 2011, the Bureau began implementing a
change to the LPS component of the NCS; the component
is used to produce annual occupational earnings data for
the nation, each Census division, and selected geographic
areas. When this change is fully implemented, a modeling
technique that combines the national data from the NCS
with the locality data from the OES survey will be used to
produce these occupational earnings estimates. All other
data estimates computed with the use of NCS data, including
the ECI, Employer Costs for Employee Compensation
(ECEC), and various measures of access and participation
in employer-provided benefits, will continue to be produced.
With the elimination of a need to produce locality estimates
directly from the NCS sample, a more efficient nationally
based sample design is being introduced. This new sample
design, which began with collection in June 2012 and is
expected to be completely phased in with the December 2016
estimates, eliminates the current stage 1 process of selecting
geographic areas. (For more information on the new sample
design, see Gwyn R. Ferguson, Joan L. Coleman, and
Chester H. Ponikowski, “Update on the evaluation of sample
design issues in the National Compensation Survey,” paper
presented at the Joint Statistical Meetings of the American
Statistical Association, Alexandria, VA, August 31, 2011,
www.bls.gov/osmr/pdf/st110230.pdf.)

Area definitions
Geographic areas of the United States are defined by the
Office of Management and Budget (OMB) once every 10
years and are based on the results of the most recent decennial
census. In December 2003, OMB defined 361 Metropolitan
Statistical Areas and 573 Micropolitan Statistical Areas in the
nation. OMB also defined a list of Combined Statistical Areas
(CSAs) such that adjacent Metropolitan and Micropolitan
Statistical Areas, in various combinations, form a CSA if
certain specified criteria are met. There are 1,359 counties in
the nation that are not included in either type of statistical area.
Any county not included in a Metropolitan or Micropolitan
Statistical Area is referred to as an Outside CBSA.
To keep NCS products representative of the areas
surveyed, the NCS must phase in the new OMB-defined
sampled areas and phase out the old. The previous NCS area
sample was based on the December 1994 OMB definitions.
The 1994 sample consisted of 152 areas; however, because
of a difference between the 1994 and 2003 OMB area
definitions, and because of the NCS clustering of Outside
CBSAs, only a portion of the 2003 area sample’s 152 areas
include exactly the same municipal and county areas as those
of the 1994 sample.
The NCS converted the state and local government sample
of index establishments to the December 2003 OMB area
definitions in December 2007. The NCS began to convert the
private industry sample of index establishments to 2003 OMB
area definitions in December 2008, with the replacement of
one-fifth of the private industry sample under the new area
definition each year through 2012. (Definitions of index

Selecting geographic areas (Stage 1)
In stage 1, the NCS selects a sample of geographic areas in
the nation. The sampling frame is the list of establishments
from which the survey sample is selected. Selection of the
geographic areas has resulted in the methodology currently
used by the NCS.2 The paragraphs that follow describe, in
general terms, the methods by which the NCS selects area
samples.
  2 See the topic “Area sample selection” in the section titled “Technical References” at the end of this chapter for references regarding the research and
the decisionmaking process by which the current NCS sample frame was
created.

2

establishments are given in detail in the section titled “Data
Collection.”) The transition has resulted in the collection of
private industry data from 227 areas. Thus, NCS publications
with reference periods of December 2007 through December
2012 may include data from as many as 227 local areas.3

new establishment samples, the NCS uses the most recent
version of NAICS as one of the stratification variables. The
NCS sample was stratified beginning in August 2007 on the
basis of the 2007 version of NAICS. NCS publications with
collection periods that include August 2007 contain data from
both the 2002 and 2007 NAICS industry codes. The 2007
NAICS includes revisions across several sectors. The most
significant revisions are in the information sector, particularly
within the telecommunications area; overall, however, the
change from the 2002 NAICS to the 2007 NAICS had little
effect on the resulting NCS estimates.4

Selecting establishments (Stage 2)
In stage 2, the NCS uses the PPS technique to select a
sample of private industry and state and local government
establishments within each of the sampled areas. That is, the
larger the establishment, the greater is its chance of being
selected.
With some minor exceptions, an establishment is a single
economic unit that engages in one, or predominantly one,
type of economic activity. For private industries in the survey,
the establishment usually is at a single physical location,
such as a mine, a factory, an office, or a store, that produces
goods or provides services. If a sampled establishment
is owned by a larger entity with many locations, only the
employment and characteristics of the establishment selected
for the sample are considered for the survey. For state and
local governments, an establishment can include more than
one physical location, such as a school district, a college, a
university, a hospital, a nursing home, an administrative body,
a court, a police department, a fire department, a health or
social service operation, a highway maintenance operation,
an urban transit operation, or some other governmental
unit within a defined geographic area or jurisdiction. Each
establishment is assigned a six-digit code from the North
American Industry Classification System (NAICS). When
a single physical location encompasses two or more distinct
economic activities, it is treated as two or more separate
establishments if separate payroll records are available and
certain other criteria are met.
The sampling frame, or universe, is the list of establishments from which the survey sample is selected. It is developed from state unemployment insurance reports. The most
recent month of reference available at the time the sample is
selected is used to develop sampling frames.

Panel structure
The NCS uses a panel structure to rotate establishments in
and out of the survey. A panel is a subset of all establishments
sampled for the survey that begin their participation in the NCS
at the same time. Each panel is composed of a representative
cross section of industries and geographic areas throughout
the United States that are within the scope of the survey.
Approximately one-fifth of the private industry establishment sample is reselected each year. The private industry sample is divided into five panels that enter and exit the
survey on a rotational basis. A panel of establishments is
introduced into the survey once each year. Private industry
panels stay in the survey for 5 years. Establishments that go
out of business or refuse to participate in the survey are not
replaced within the panel; rather, the NCS adjusts for establishments’ and occupations’ refusals as the panel proceeds.
With one panel of new establishments entering the survey,
and one exiting, each year, the sample is fully replaced over
an approximately 5-year period. This practice helps to reduce
respondent burden and keep the sample current. When a new
replacement panel is introduced into the survey, field economists conduct the initial interviews of establishments in the
new panel while updating the establishment records of the
other four panels.
The state and local government establishment sample
includes only one panel, replaced approximately once
every 10 years. This arrangement differs from the private
industry 5-year rotation because state and local government
establishments are, generally, more stable in terms of
establishment births and deaths as well as in the number of
employees. The NCS replaced its state and local government
index–establishment sample in its entirety in December 2007,
using the 2007 NAICS to stratify the sampling frame for the
selection of new establishments.
In areas newly surveyed in 2007, data were collected only
in state and local government establishments. From 2008
through 2012, the NCS used the 2007 NAICS to stratify the
sampling frame in order to introduce panels of new private
industry establishments from newly selected areas based on
the December 2003 OMB area definitions. Currently, the
NCS is transitioning to the 2012 NAICS.

Industry classification of establishments
All federal statistical agencies currently use NAICS to define
industries and classify survey establishments. The NCS,
which originally used Standard Industrial Classification
(SIC) codes to stratify establishments for selection, began a
transition from SIC to the 2002 version of NAICS in 2004;
the transition was completed in July 2007. NAICS revises
its industry classifications every 5 years to stay current
with industrial organization in North America. In selecting
  3 For more information, see Jason Tahonica, Lawrence R. Ernst, and Chester
H. Ponikowski, “Phase-in of the Redesigned National Compensation Survey
Area Sample,” paper presented at the American Statistical Association
Section on Survey Research Methods, Alexandria, VA, August 2005, https://
www.amstat.org/sections/srms/Proceedings/y2005/Files/JSM2005000156.pdf; for a list of current and historical OMB area definitions, see
“Metropolitan and Micropolitan: Metropolitan and Micropolitan Statistical
Areas Main” (U.S. Census Bureau, May 2012), https://www.census.gov/
population/www/metroareas/metrodef.html.

  4 For more information about the differences between the 2002 NAICS and
the 2007 NAICS, see “North American Industry Classification: Introduction
to NAICS” (U.S. Census Bureau), http://www.census.gov/epcd/naics07.

3

• What are the employees in the sampled, matched
occupation paid? The field economist collects data on
hours and wages from the payroll records covering the
12th of the month.

Probability sampling of occupations within sampled
establishments (Stage 3)
In stage 3, field economists use a technique to randomly select
the jobs to be sampled during the initial contact with the
sampled establishment. (See the section titled “Occupational
Selection and Classification” for details.)

• What are the duties and responsibilities of the job? The
field economist collects the information requested and
uses it to determine the number of points ascribed to each
of the pay factors of the job. From the sum of the points of
all of the pay factors, the work level of each job surveyed
is determined.6

Occupational classification and the transition to
new occupational definitions
Before 2004, the NCS used the Occupational Classification
System (OCS), which was based on the 1990 Census of
Population and Housing, to classify jobs in the survey’s
selection and publication of occupational data. The NCS
phased in the use of Standard Occupational Classification
(SOC) codes over several years. The survey first published
ECEC estimates using SOC codes in March 2004, the ECI
in March 2006, and NCS benefits publications in March of
2007. The survey first published local area earnings estimates
using the SOC in September 2006 and followed with the
NCS national and Census division publications in September
2007.
Starting in May 2010, the NCS began to use SOC 2010
definitions for initial data collection at an establishment.
Updates (followup data collection at an establishment) began
in September 2010. The 2010 SOC system contains 840
detailed occupations, aggregated into 461 broad occupations.5
The first NCS publication to use SOC 2010 was the March
2011 ECI release.

• How many hours does the employee work? The field
economist collects data on the usual work schedule of
each sampled, matched occupation. This information
helps to determine the employee’s hourly, weekly,
and annual earnings, as well as the employer’s cost of
benefits.
• What types of benefits do the employees receive? The field
economist gathers data on the availability to the worker
and cost to the employer of 18 selected benefits that
the employer may offer to the worker in each sampled,
matched job. The field economist collects summary plan
descriptions of the health and retirement plans offered
by the employer. Data on the availability of a number of
other benefits also are collected. Then the descriptions are
analyzed to generate data on benefit plan provisions (the
terms of coverage of the plan).
Collection period
A BLS field economist contacts the sampled establishment
for the initial collection of data. From each establishment,
the NCS collects data on wages, the cost of benefits, and
the incidence and provisions of benefits. These data are
used to produce ECI, ECEC, and NCS estimates of benefits.
Although initial data collection occurs at any time of the
year, ECI and ECEC updates are conducted over a 6-week
timeframe for the pay period that includes the 12th day of
the month for the months of March, June, September, and
December. For example, the news release “Employment
Cost Index—March 2012” contains data collected for
the pay period that included March 12, 2012, from each
employer scheduled for an ECI update. The publication
date of the news release was April, 27, 2012.7 The news
release “Employer Costs for Employee Compensation—
March 2012” contains data collected for the same pay
period—in this example, the pay period that included
March 12, 2012; however, its publication date was June 7,
2012.8 The ECEC publication dates are 3 months after the
reference month, the ECI publication dates 1 month after
the reference month.

Data Collection
BLS field economists employ a variety of methods, including
personal visits, mail, telephone, and email, to obtain data
from NCS survey respondents. At the initial and subsequent
contacts, field economists ask the following questions:
• What is the primary business activity of the establishment?
The field economist determines the correct industry code
for the establishment.
• What types of occupations does the establishment employ?
The field economist determines the correct SOC code and
work level for each sampled job.
• How many employees are there in each sampled job
that is matched to an occupational description? The
field economist determines how many employees in the
establishment can be defined by the occupational code for
the sampled job.
• Do workers in the matched, sampled occupation work
full or part time? Are they union or nonunion workers?
Paid by time or incentive? The field economist determines
these three work attributes of the employee in the matched
occupation. The field economist proceeds to collect wage
and benefits data on all of the workers with the same work
attributes in the matched occupation.

  6 For more information on pay factors and work levels, see “National Compensation Survey: Guide for Evaluating Your Firm’s Jobs and Pay” (U.S.
Bureau of Labor Statistics, October 2003), http://www.bls.gov/ncs/ocs/sp/
ncbr0004.pdf.
  7Visit http://www.bls.gov/news.release/pdf/eci.pdf for the most recent
ECI news release.
  8Visit http://www.bls.gov/news.release/pdf/ecec.pdf for the most recent
ECEC news release.

  5 For more information on SOC 2010, see “Standard Occupational Classification” (U.S. Bureau of Labor Statistics), http://www.bls.gov/SOC.

4

Each March of the collection cycle, the NCS collects
benefits incidence and key provisions data from establishments
contacted for survey updates that quarter. The March data are
published in the summer or fall of the same year.9
NCS detailed benefits provisions data typically are
collected from establishments at initiation visits over
approximately 18 months and are published the next year.10

tion, is conducted by field economists during the initial contact with the sampled establishment. There are four main
steps in this stage:
1. Selection of establishment jobs by the NCS Probability
Selection of Occupations (PSO) technique. With this
technique, the probability of selecting a given job is
proportional to the number of workers in the job in the
establishment

Occupational Selection and Classification

2. Classification of jobs into occupations based on the SOC
system

The NCS collects data on workers who are employed by the
sampled establishment. Persons working onsite at a surveyed
establishment, but paid by a contractor, are not included in
data collection from the establishment, unless the contractor
is part of the sample, in which case the NCS collects data
on employees of the contractor who are working offsite at
other establishments, as well as those working onsite. To
be included in the NCS, employees in sampled occupations
must receive cash payments (cash, check, or direct deposit
payments) from the establishment for services performed
and the establishment must pay the employer’s portion of
Medicare taxes on those individuals’ wages.
The number of workers in an establishment includes
workers on paid vacation or other types of leave; salaried officers, executives, and staff members of incorporated firms;
employees temporarily assigned to other units; and noncontract employees for whom the reporting unit is the permanent duty station, regardless of whether that unit issues their
paychecks.
In sampling jobs at an establishment, BLS field economists
use a method that ensures a random sample. Field economists
then match employees working in the sampled jobs with an
occupation (as defined in the SOC structure). Workers are
classified into occupations on the basis of the work performed
and the skills required in each occupation, not on the basis
of their education. For example, an employee trained as an
engineer, but working as a drafter, is reported as a drafter.
An employee who performs the duties of two or more
occupations is reported as working in the occupation that
requires the highest level of skill or in the occupation in which
the employee spends the most time if there is no measurable
difference in skill requirements. A quote is a sampled job that
has been matched with an SOC occupation; it includes all
workers in the job who have the same occupational attributes:
full-time or part-time status, union or nonunion status, and
whether they are paid on a time or an incentive basis.
Stages 1 and 2 of the NCS sampling procedures are described in the section titled “Sample Design and Sampling
Procedures.” Stage 3, occupational selection and classifica-

3. Identification of attributes of the worker in the occupation, such as full-time or part-time status, union or nonunion status, and whether the worker is paid on a time or
incentive basis
4. Identification of the work level of each occupation
Selection of occupations. In step 1, the field economist
receives the establishment’s complete list of employees and
their job titles. The field economist then uses the NCS PSO
technique to randomly select the jobs to be sampled. The
number of jobs selected for which data are collected is based
on the establishment’s employment size, according to the
following criteria:
Number of employees
Number of jobs selected

1–49

50–249

250 or more

Up to 4

6

8

Exceptions include state and local government units,
for which up to 20 jobs may be selected, and the aircraftmanufacturing industry units—those matching NAICS code
336411—for which up to 32 jobs may be selected.
Classification of jobs. In step 2, the field economist classifies
the sampled jobs into occupational categories based on the
workers’ actual job duties and responsibilities, not on their
job titles.
Each occupation is designated by a six-digit code that
is part of a hierarchical structure: detailed occupations are
grouped under broad occupations, broad occupations are
part of a minor group, and minor groups are part of a major
group. The SOC designates 23 major groups. Major group
codes end with 0000, minor group codes end with 000, and
broad occupation codes end with 0. For example, the detailed
occupation Orthodontists (code 29-1023) is under the broad
occupation Dentists (code 29-1020), which is under the
minor group Health Diagnosing and Treating Practitioners
(code 29-1000), which is under the major group Health Care
Practitioner and Technical Occupations (code 29-0000).11 In
the NCS, occupations can fall into any of 22 major groups;
only the major group 23 (SOC code 55-0000), military
specific occupations, is excluded.

  9 See, for example, National Compensation Survey: Employee Benefits
in the United States, March 2011, Bulletin 2771 (U.S. Bureau of Labor
Statistics, September 2011), http://www.bls.gov/ncs/ebs/benefits/2011/
ebbl0048.pdf.
  10 See, for example, National Compensation Survey: Health Plan Provisions in State and Local Government in the United States, 2011, Bulletin
2772 (U.S. Bureau of Labor Statistics, March 2012), http://www.bls.gov/
ncs/ebs/detailedprovisions/2011/ebbl0049.pdf.

  11See the entire list of SOC occupational categories in “Standard Occupational Classification” (U.S. Bureau of Labor Statistics, November 18, 2010),
http://www.bls.gov/soc/home.htm.

5

Identification of occupational attributes of the worker. In
step 3, the field economist records specific attributes of the
worker in the sampled job, for each selected occupation. Each
such occupation must include only workers with the same
attributes; for example, the occupation cannot include both
full-time and part-time workers. The occupational attributes
of workers, as determined by the NCS, are as follows:

accounting occupation is based on a different set of descriptions by which levels and points are assigned than the knowledge factor of an engineering occupation. The broad occupational categories for which unique descriptions are given are
as follows:
•
•
•
•
•
•
•
•

• Full-time or part-time status. The field economist
identifies the worker as holding either a full-time job or
a part-time job. For the NCS, full-time or part-time status
is not determined by the number of hours worked, but is
based instead on the establishment’s definition of those
terms.
• Time-based or incentive-based pay. The field economist
identifies the worker as having time-based or incentivebased pay, depending on whether any part of the pay was
based directly on the actual production of the worker, rather
than solely on the number of hours worked. Time workers
are those whose wages are based solely on an hourly rate
or salary. Incentive workers are those whose wages are
based at least partially on productivity payments, such as
piece rates, commissions, and production bonuses.

•
•
•
•
•
•
•
•

• Union or nonunion workers. The field economist records
whether the occupation is filled by union or nonunion
workers. The NCS defines a union worker as any employee in a union occupation who satisfies all of the following
conditions: a labor organization is recognized as the bargaining agent for all workers in the occupation; wage and
salary rates are determined through collective bargaining
or negotiations; and settlement terms, which must include
earnings provisions and may include benefit provisions,
are embodied in a signed, mutually binding collective bargaining agreement. A nonunion worker is an employee in
an occupation not meeting all of the NCS-defined conditions for union coverage.

•
•
•
•
•
•
•

Business Administration
Professional Accounting and Auditing
Information Technology
Professional Mathematics and Statistics
Professional Engineering and Architecture
Engineering and Scientific Technician
Professional Biological and Physical Science
Professional Economics, Sociology, Geography,
Psychology, and Similar Jobs
Social Welfare and Health Administration
Professional Legal
Administrative Legal
Professional Education
Professional Librarian, Museum Curator, and Archivist
Communications and Arts
Professional Medical
Medical, Hospital, Dental, Public Health, and Veterinary
Technician
Protective Service
Investigation, Inspection, and Compliance
Service
Sales
Office and Administrative Support
Miscellaneous Technician
Pilots and Air Transportation

The job is assigned points for the highest level at which all
requirements are met. The entire process is known as pointfactor leveling. If a specific work level cannot be determined
for a selected occupation, the data for that occupation are
recorded as not able to be leveled.

Determining the work level of the job. In step 4, field economists evaluate the job, using a “point-factor” system of points
ascribed to pay factors to determine the work level of a selected occupation. The NCS system uses four distinct factors:

Determining supervisory responsibilities. According to the
current version of the SOC,
Supervisors of professional and technical
workers usually have a background similar
to the workers they supervise, and are
therefore classified with the workers they
supervise. Likewise, team leaders, lead
workers and supervisors of production,
sales, and service workers who spend at
least 20 percent of their time performing
work similar to that of the workers they
supervise are classified with the workers
they supervise.12

• Knowledge
• Job controls and complexity
• Contacts
• Physical environment
Each factor consists of several degrees, each with an associated description and number of points. Generally, the greater
the impact, complexity, or difficulty of the factor, the higher
is the number of points assigned.
Except for the knowledge factor, the descriptions apply
to all occupational categories. Knowledge has a separate set
of descriptions for each of several broad occupational categories. For example, the knowledge factor of a professional

  12 “Standard Occupational Classification: Standard Occupational Classification (SOC) User Guide” (U.S. Bureau of Labor Statistics, March 25, 2011),
Classification Principles, #3, http://www.bls.gov/soc/socguide.htm.

6

Typically, supervisors have the authority to hire, transfer,
lay off, promote, reward, and discipline other employees.
For the NCS, field economists record whether the occupation
includes supervisory responsibilities and, if so, the level of
responsibility. By NCS definition, first-line supervisors direct
their staff through face-to-face meetings and are responsible
for conducting the employees’ performance appraisals whereas
second-line supervisors typically direct the actions of their
subordinates through first-line supervisors. The NCS also
evaluates most supervisory jobs on work levels based on the
four point factors previously described. A modified approach is
used for professional and administrative supervisors who direct
professional workers and are paid primarily for their supervisory
and managerial skills; the levels of such supervisory jobs are
determined on the basis of the duties and responsibilities of the
highest reporting position.13

time schedule, union or nonunion status, and time or
incentive type of pay), work schedule, and wage data.
Wages account for roughly 70 percent of compensation;
therefore, if wage data are not available, other data from
the establishment cannot be used in calculating estimates.
An establishment is considered nonresponding if it
provided neither earnings, occupational classification,
worker attributes, nor work schedule data for any occupation. Establishment nonresponse during the initial
interview is treated with adjustments that redistribute the
weights of nonrespondents to similar respondents on the
basis of characteristics such as the industry and size class
of the establishment. For example, if the nonresponding
establishment was in the manufacturing industry and had
an employment of 350 workers, the NCS would adjust
the weights of responding manufacturing establishments
with 250–499 workers by a nonresponse factor calculated
by dividing the sum of the product of establishment
employment and sample weight for responding and
nonresponding establishments by the sum of the product
of establishment employment and sample weight for
responding establishments.
At subsequent interviews of an establishment that had
provided data at initiation, establishment nonresponse is
treated by imputation, in which missing values for an
initially responding establishment are replaced by values
from the original interview, adjusted by the rate of change
among responding establishments. Establishments no
longer in operation or out of the scope of the survey, and
establishments with no workers within the scope of the
survey, are excluded from the survey estimates.

Weighting, Nonresponse Adjustment,
Imputation, and Benchmarking
Participation in the survey is voluntary; therefore, a company
official may refuse to participate in the initial survey or may be
unwilling or unable to update previously collected data for one
or more occupations during a subsequent contact. In addition,
some establishments selected from the sampling frame may be
out of the scope of the survey or may have gone out of business.
To address the problems of nonresponse and missing data, the
NCS adjusts the weights of the remaining establishments and
imputes missing values, to ensure that published compensation
estimates ultimately are representative of compensation in
the civilian, private industry, and state and local government
sectors. Beginning in 2006, the NCS implemented a number
of significant changes in the survey, including imputing for
temporary nonresponse situations and benchmarking estimated
employment. This section describes the current imputation
and benchmarking methods.14
Weight adjustments and imputation are made in accordance
with the following steps:

2. Item nonresponse occurs when an establishment
responds to the survey but is unable or unwilling
to provide some of the benefits data, occupational
classifications, or worker attributes for a given sampled
occupation. Item nonresponse is treated by item
imputation in certain situations. In item imputation,
missing values for an item are replaced by values
derived from establishments with similar characteristics
that completed the item. For NCS benefits estimates,
items can be imputed for nonresponse at initial and
subsequent data collection. For example, suppose that,
during the initial contact, an establishment reports
earnings data for a sampled occupation but refuses to
report whether those in the occupation receive paid
vacation leave; the NCS then imputes the incidence of
vacation leave for the occupation on the basis of the
incidence of vacation leave among similar occupations
in similar establishments. For NCS estimates other
than benefits estimates, earnings data are not imputed
for item nonresponse during the establishment’s
initial data collection but are imputed at subsequent
data collections. For example, if a manufacturing
establishment gave no information on the earnings per
hour of its full-time, nonunion assembly workers at the
initial collection of data, the NCS would not use any of

1. An establishment is considered responding if it provided
information on at least one usable occupation. An
occupation is classified as usable if the following data
are present: occupational attributes (full-time or part  13 For a complete description of point-factor leveling and the determination
of supervisory levels, see National Compensation Survey: Guide for Evaluating Your Firm’s Jobs and Pay (U.S. Bureau of Labor Statistics, October
2003), http://www.bls.gov/ncs/ocs/sp/ncbr0004.pdf.
  14 For more information on these and other changes in NCS methodology with respect to NCS wage products, see James E. Smith and Robert
W. Van Giezen, “Compensation and Working Conditions: Change Comes
to the National Compensation Survey Locality Wage Bulletins” (U.S. Bureau of Labor Statistics, January 24, 2007), http://www.bls.gov/opub/cwc/
cm20070122ar01p1.htm. For more information on these changes with respect to the ECI, see “Employment Cost Trends: Change Has Come to the
ECI” (U.S. Bureau of Labor Statistics, May 12, 2006), http://www.bls.gov/
ncs/ect/ecsm0001.htm, and with respect to BLS publications on benefits,
see “Employee Benefits Survey: Change to the NCS Benefits Products”
(U.S. Bureau of Labor Statistics, April 14, 2008), http://www.bls.gov/ncs/
ebs/ebsm0005.htm.

7

the data on those workers. However, if the establishment
reported earnings per hour for its full-time, nonunion
assembly workers during the initial collection, but not
in a subsequent collection, the NCS would use the most
recent reported hourly earnings of full-time, nonunion
assembly workers in the establishment, adjusted for the
rate of change in hourly earnings of workers in similar
manufacturing establishments, to impute the missing
data.

use of 2002 employment counts from the Occupational
Employment Statistics survey.15
Before December 2006, the ECEC used only QCEW
employment counts for benchmarking; starting with
the quarter including December 2006, the ECEC began
using employment data from two BLS programs—the
QCEW and the Current Employment Statistics (CES)
program—for benchmarking. The CES data are used
to update QCEW data that are about 6 months old.
Combined, data from the two programs provide the
appropriate industry coverage and timeliness needed for
the ECEC. Starting with the publication of the March
2007 estimates, the NCS benefits incidence and key
provisions (I&P) series began using employment data
from the same two BLS programs as well. The NCS
continues to use employment counts from the most
recent QCEW data to benchmark estimates of detailed
NCS benefits provisions.16

3. A third factor adjusts for any special situations that may
have occurred during data collection. For example, when
a sample unit is one of two establishments owned by a
given company and the respondent provides data for
both locations combined instead of data for the sampled
unit, the weight of the sampled unit is adjusted to reflect
the employment data actually collected.
4. Finally, poststratification, or benchmarking, is the
process of adjusting the weight of each establishment
in the survey to match the most current distribution
of employment by industry. The NCS establishment
sample is drawn from the Quarterly Census of
Employment and Wages (QCEW). Because the
sample of establishments used to collect NCS data
was chosen over the past several years, establishment
weights reflect employment at the time of sampling.
The benchmark process updates those weights on the
basis of current employment. Benchmarking ensures
that survey estimates reflect the most current industry–
government (hereafter, ownership) employment counts
in proportions consistent with the private industry, state
government, and local government sectors. For example,
assume that 40 private industry, 10 local government,
and 5 state government units in the service sector were
selected from the NCS sampling frame containing
200,000 private workers, 30,000 local government
workers, and 10,000 state government workers. Suppose
also that, by the time of survey processing, the private
service sector experienced an employment increase
of 10,000 workers, or 5 percent, and no increase in
employment was experienced in the service sectors of
state and local government. Then, if the NCS did not
benchmark, the sample would underrepresent current
employment in the private industry service sector. In
this example, the NCS would adjust the sample weights
of the 40 service sector firms in private industry to
ensure that the number of workers in the sampling
frame rises to 210,000. The ownership employment
counts for the private industry service sector would
then reflect the current proportions of 84 percent for
private industry, 12 percent for local government,
and 4 percent for state government employment.
  The benchmark calculation is essentially the same
for all NCS data products; however, the input to the
calculation differs by data product. The ECI uses fixed
employment weights; the benchmark adjustment for
the ECI is calculated each quarter, currently with the

In NCS publications on benefits data, an appendix table on
survey establishment response in the technical notes shows
the number of establishments by sector in the sampling frame
and in the sample, the number responding to the survey, the
number out of the scope of the survey or out of business, and
the number unable or unwilling to participate in the survey.
A second appendix table shows the estimated number of
workers represented by the survey; this estimate includes the
number of all workers in all establishments within the scope
of the NCS, by sector and by occupational group, represented
by the survey—not the number of workers actually surveyed.
The number of workers represented by the survey is not
intended as an accurate employment count; rather, it indicates
only the relative importance of the occupational groups
studied in the survey.

Calculation and Reliability of the Estimates
This section describes computations and reliability measures
that apply to all NCS data products. At the end of the section
are links to Web pages for each NCS data product; the pages
include a description of the computations and reliability
measures specific to the individual product.
NCS estimates are derived from a sample of occupations
selected from responding establishments. Two types of
errors are possible in an estimate based on a sample survey:
sampling errors and nonsampling errors. Sampling errors
occur because the sample makes up only a part of the
population of interest. The sample used for the survey is
one of a number of possible samples that could have been
selected under the sample design, each producing its own
estimate. A measure of the variation among sample estimates
is the standard error. Nonsampling errors are data errors that
  15 See “Occupational Employment Statistics” (U.S. Bureau of Labor Statistics),
http://www.bls.gov/oes.
  16 For more information on ECEC benchmarking, see “Changes in Calculations for the BLS Employer Costs for Employee Compensation Data,
March 2007” (Bureau of Labor Statistics), http://www.bls.gov/ncs/ect/sp/
ececcalc.pdf.

8

stem from any source other than sampling error, such as data
collection errors and data-processing errors.
Standard errors can be used to measure the precision with
which an estimate from a particular sample approximates
the expected result of all possible samples. The chances
are about 68 out of 100 that an estimate from the survey
differs from a complete population figure by less than the
standard error. The chances are about 90 out of 100 that this
difference is less than 1.6 times the standard error. Statements
of comparison appearing in NCS publications are significant
at a level of 1.6 standard errors or better. This means that,
for differences cited, the estimated difference is less than 1.6
times the standard error of the difference. To assist users in
ascertaining the reliability of NCS series, standard errors for
all NCS estimates are available online.17
The ECI, ECEC, and benefits publications all use
some variation of balanced repeated replication (BRR),
a methodology employed to estimate the standard error.
The procedure for BRR is first to partition the sample
into variance strata composed of single sampling strata or
clusters of sampling strata and then to split the sample units
in each variance stratum evenly into two variance primary
sampling units (PSUs). Next, half-samples are chosen that
each contain exactly one variance PSU from each variance
stratum. Choices are not random, but are designed to yield a
“balanced” collection of half-samples. For each half-sample,
a “replicate” estimate is computed with the same formula for
the regular, or “full-sample,” estimate, except that the final
weights are adjusted. If a unit is in the half-sample, its weight
is multiplied by (2 – k); if not, its weight is multiplied by k.
For all NCS publications, k = 0.5, so the multipliers are 1.5
and 0.5. (Some of the weighting adjustments done as part of
the calculation of final weights also are recalculated for each
replicate.)
The BRR estimate of standard error with R half samples is

computer edits of the data (validation), and systematic
professional review of the data. The programs also serve as a
training device to provide feedback to field economists, or data
collectors, on errors. They provide information as well on the
sources of errors that can be remedied by improved collection
instructions or computer-processing edits. Extensive training
of field economists is conducted to maintain high standards
in data collection.
Once estimates of compensation cost changes, of wage
and compensation cost levels, or of benefit provisions are
produced, the estimates are verified, or validated. The focus
of the verification at this stage is a comparison of the estimates
with expectations. Expectations are based on economic
conditions; recent trends in similar data; and values prevalent
in the recent past among industries, occupations, union
status, region of the country, types of compensation, and so
on. Computer checks are used to identify anomalies, such
as wage changes outside the historical range. Another set of
checks verifies that there are enough observations supporting
each estimate and that there will be no way in which data
from a respondent could be identified. Once estimates that
are anomalies or that differ substantially from expectations in
other ways are identified, the data underlying those estimates
are examined in detail to try to explain the results. When the
review staff is convinced that the data are accurate and that
they are based on enough observations, the staff designates
the data as “fit for use” and the data are published.
Not all calculated series meet the criteria for publication.
Before any series is published, it is reviewed to make sure
that it meets specified statistical reliability and confidentiality
requirements. The review prevents the publication of a series
that could reveal information about a specific establishment
or have a large sampling error.
The following are links to descriptions of the methods
used to compute estimates for specific NCS products and of
the methods for calculating the reliability of those estimates:

1
�𝑟𝑟 − 𝑌𝑌
� )2 ,
∑𝑅𝑅𝑟𝑟=1(𝑌𝑌
SE�𝑌𝑌�� = �
𝑅𝑅(1−𝑘𝑘)2

Employment Cost Index (ECI) series

where

Employer Costs for Employee Compensation
(ECEC) series

the summation is over all half-samples r = 1,...,R,

Incidence and provisions of benefits

𝑌𝑌�𝑟𝑟 is the rth replicate estimate, and

ECI series
The ECI is a measure of the change in the cost of labor,
independent of the influence of employment shifts among
occupations and industry categories. The total compensation
series includes changes in wages and salaries and in employer
costs for employee benefits. The ECI calculates indexes of
total compensation, wages and salaries, and benefits separately
for all civilian workers in the United States (as defined by the
NCS), for private industry workers, and for workers in state
and local government. For all of these categories, the ECI
calculates the same indexes by occupational group, worker
attribute, industry group, and establishment characteristic.
Seasonally adjusted series are calculated as well.
Employer costs for employee benefits are collected for paid
leave—vacations, holidays, sick leave, and personal leave;

𝑌𝑌�𝑟𝑟 is the full-sample estimate.

Percent relative standard error data are provided alongside
estimates in NCS ECEC publications, which display the
standard error as a percentage of the full-sample estimate.
The percent relative standard error is given by
̂ /Y.̂
%RSE = 100 × SE(Y)
Data collection and processing errors are mitigated
primarily through quality assurance programs that include
the use of data collection reinterviews, observed interviews,
  17 See “National Compensation Survey” (U.S. Bureau of Labor Statistics),
http://www.bls.gov/ncs.

9

supplemental pay—premium pay for work done in addition
to that performed during the regular work schedule (such as
overtime, weekends, and holidays) and for shift differentials,
and nonproduction bonuses (such as yearend, referral, and
attendance bonuses); insurance benefits—life, health, short-term
disability, and long-term disability insurance; retirement and
savings benefits—defined benefit and defined contribution plans;
and legally required benefits—Social Security, Medicare, federal
and state unemployment insurance, and workers’ compensation.
As mentioned earlier, the ECI is a Principal Federal Economic Indicator.18

where
It is the index at period t,
W0, i is the estimated base-period wage bill for the ith cell, and
M t, i is the multiplicatively accumulated average wage change
in the ith cell from time 0 (the base period) to time t (the
current quarter).
In the preceding equation, W0, i, the wage bill, is the
product of the average wage of workers in the cell at the base
period (0) and the number of workers represented by the cell.
For the ECI, the number of workers represented by the cell
is held fixed. Also, in essence, M t, i projects the base-period
average wage level for the cell forward to the current quarter.
Note that M t, i can be written as M t, i = M t - I, i × R t, i ,

Computation of ECI series
The ECI is a Laspeyres index. The basic computational
framework for the ECI is the standard formula for an
index number with fixed index weights, modified by the
special statistical conditions that apply to the ECI. The
text that follows describes ECI measures of wage changes,
but indexes of changes in benefits and of changes in total
compensation (the latter defined by the NCS as the sum of
changes in wages and benefits) are calculated in essentially
the same manner.
An index number for the ECI is simply a weighted average
of the cumulative average wage changes within each of the
ECI basic cells, with base-period “wage bills” serving as the
fixed weights. A basic cell for the ECI is composed of raw
wage data from a narrowly defined set of workers, sorted by
ownership sector, industry, and occupational groups in which
they work. The ECI cell structure sorts the industry codes
into 1 of 3 ownership sectors: private, state government, or
local government. Workers within private establishments
are sorted into 1 of 58 industry categories that are defined
primarily by three-digit NAICS codes. Workers in either state
or local government are sorted into 13 industry categories;
the government industry categories are as broad as “all
goods-producing industries” and as narrow as “hospitals.”
Each of these private and government industry groups is
arrayed across nine aggregate occupational groups, which are
ordered numerically by their SOC codes. Altogether, there
are 522 (58 × 9) private industry occupational cells and 234
(13 × 9 × 2) state and local government industry occupational
cells, for a total of 756 ECI basic cells.
For each basic cell, a base-period wage bill (W0) is
computed and the wage bill is updated each quarter by
observed rates of change from the ECI survey sample. The
simplified formula for a basic cell is

∑ (W

0, i

It =

∑ WW
∑

where
R t, i is the ratio of the current-quarter weighted average
wage in the cell to the previous-quarter weighted average
wage in the cell, both calculated in the current quarter from
matched-sampled quotes. Using only matched quotes in the
ratio eliminates the inclusion of wage changes that might
be caused by shifting workers within establishment jobs.
That is, the ECI sample tracks changes in wages within
establishment jobs and not for individual workers of the
establishment. The weights applied are the sample weights
of the survey.
All wage indexes are computed from the following data:
• Average straight-time hourly earnings for six-digit SOC
code occupations, or groups of occupations, in those
sample establishments for which data are available for
both the current and previous quarters. These earnings are
called matched quotes. In addition to being identified by
the six-digit SOC code, a quote within an establishment is
identified from quarter to quarter by its union status, fulltime or part-time status, method of pay (time or incentive),
and job level.
• Employment for each of the basic cells, enumerated with
the use of 2002 employment data from the BLS OES
survey.
• Sample weights that reflect both employment in each
establishment or occupation surveyed and the probability
of selection of that establishment or occupation.
Computation of the index for a calendar quarter involves
five principal steps:

M t, i )
× 100,

1. Sampled occupation (i.e., quote) weights are applied to
the average occupational hourly wage for every quote in
a sampled establishment that has reported both currentquarter and previous-quarter wage data. These data are
used to calculate a weighted average wage for each
basic cell (i.e., for each occupational group within each
industry) for the current and previous survey periods.

0, i
0, i

  18 Principal Federal Economic Indicators are the major statistical series that
describe the current condition of the economy. For more details, see OMB
Statistical Policy Directive No. 3, in the Federal Register, September 25,
1985, http://www.bea.gov/about/pdf/federalregister09251985.pdf.

10

2. The ratio of the current-quarter to the previous-quarter
weighted average wage is then calculated for each
cell i. This ratio (R t, i ) is used as an estimate of the
current-quarter (t) wage change for that basic cell and
is multiplied by the previous-quarter (t − 1) cumulative
average wage change for the cell (M t - I, i ). The product
M t, i is a measure of the cumulative percent wage change
in the cell since the base period.

workers indexes are allowed to vary over time, these indexes
are not strictly comparable to the aggregate, industry, and
occupation indexes.
Keeping the index current
For a fixed-weighted index to remain economically relevant
over a span of periods, it is necessary to make changes to the
computations of the indexes on occasion. Beginning with the
release of the March 2006 data, the following major changes
were made in the way the ECI is calculated:

3. The measure of cumulative percent wage change is
multiplied by the base-period wage bill (W0, i ) to generate
an estimate of the current-quarter wage bill for the cell.

• Indexes were rebased from June 1989 = 100 to December
2005 = 100.

4. The current-quarter and previous-quarter wage bills are
then summed over all cells within the scope of the index.
For example, for the manufacturing wage index, the wage
bills would be summed across all cells in manufacturing.

• New fixed employment weights were introduced, using
2002 employment counts from the OES survey.19
• The BLS classification of industries was changed from the
SIC system to the 2002 NAICS.

5. The summed current-quarter wage bill (∑W0, i M t, i ) is
divided by the summed base-period wage bill (∑W0, i ).
The result, multiplied by 100, is the current-quarter
index (It), which is then divided by the previous-quarter
index (It − 1) to provide a measure of quarter-to-quarter
change, referred to as an “index link relative.”

• The BLS occupational classification was changed from
the OCS to the 2000 SOC.
• Imputation methods were changed.20
In August 2007, the NCS introduced the 2007 NAICS,
which, overall, had little effect on NCS estimates.
In 2009, the NCS began publishing continuous-series
historical information to assist data users in locating ECI data
on occupational and industry series that ran before and after
the changes just noted.21

Computations for the occupational and industry group
indexes follow the same procedures as those for the overall
indexes, except for summation. For example, for an index
for a broad occupational group, the wage bills are summed
across all cells which are a subset of that occupational group,
with indexes for industry groups calculated analogously.
Some caution is warranted in calculating the indexes for
private industry nursing care facilities. Those indexes are
estimated with the use of fixed-employment weights derived
from staffing patterns estimated from the four-digit NAICS
industry group 6231, nursing care facilities, a subindustry of
the larger industry group 623, nursing and residential care
facilities. The basic-cell cost weights for nursing care facilities
were constructed after the basic-cell fixed weights for group 623
were computed and prepared for use in the index computation
system. Consequently, the fixed weights for the four-digit
industry were not directly constructed as linear disaggregates of
NAICS group 623. Because the nursing care facilities indexes
are not linearly associated with their higher level aggregates,
they are not strictly comparable to those aggregates.
Computation procedures for measures of change in the
regional, union and nonunion, and excluding-incentive
workers indexes differ from those of the national wage
indexes because the sample is not large enough to hold the
wage bills constant at the level of detail of the former indexes.
For these indexes, the prevailing distribution in the sample
(e.g., between union and nonunion attributes within each
ownership–industry–occupation cell) is used to apportion
the previous-quarter wage bill in that cell (e.g., between the
union and nonunion indexes) each quarter. The portion of the
wage bill assigned to the union index is then moved by the
percent change in the union wages in the cell, and similarly
for the nonunion index. Therefore, the relative employment
of the union index in each cell is not held constant over time.
Because the weights of the region, union, and time-paid

Reliability of the ECI estimates
To assist users in ascertaining the reliability of indexes,
standard errors for all ECI estimates (excluding seasonally
adjusted series) are available on the BLS website.22
Publication of the ECI series
The ECI publishes indexes of total compensation, wages and
salaries, and benefits separately for all civilian workers in
the United States (as defined by the NCS), private industry
workers, and state and local government workers, and,
within each of these sectors, by occupational group, worker
attributes, industry group, and establishment characteristics.
More than 400 unique indexes and their associated quarterly
and 12-month changes in employers’ costs are published
quarterly. Seasonally adjusted indexes are published as well.
In 2008, ECI estimates were published for 14 selected
metropolitan areas for the first time; they are now published
quarterly. A 15th metropolitan area was added in 2009.
Metropolitan area data are limited to estimates of total
  19 See Stephanie L. Costo, "Introducing 2002 weights for the Employment
Cost Index," April 2006, pp. 28–32, http://www.bls.gov/opub/mlr/2006/04/
art5full.pdf.
20
 See Song Yi, "Accounting for missing data in the Employment Cost Index," Monthly Labor Review, April 2006, pp. 22–27, http://www.bls.gov/
opub/mlr/2006/04/art4full.pdf.
  21 See the subsection “Publication of index series”; for more information
on all the changes, see “Employment Cost Trends: Change Has Come to the
ECI” (U.S. Bureau of Labor Statistics, July 14, 2006), http://www.bls.gov/
ncs/ect/sp/ecsm0001.htm.
  22 See “Employment Cost Trends” (U.S. Bureau of Labor Statistics, October
26, 2009), http://www.bls.gov/ect/ectvar.htm, published shortly after publication of the news release.

11

compensation and of wages and salaries, for 12-month
periods beginning with reference date December 2005 and
for subsequent 12-month periods ending in March, June,
September, and December.23
Historical current-dollar ECI series that use industry categories based on SIC and that classify jobs into occupational
classifications according to OCS are available beginning with
the first publication of each series through December 2005.24
ECI current-dollar series based on the 2002 and 2007 NAICS
and the 2000 SOC also are available beginning March 2001,
with December 2005 = 100 as the base period.25
Historical constant-dollar ECI series also are available.
The constant-dollar ECI measures trends in compensation,
adjusted for changes in consumer prices as measured by
the CPI. The CPI-U, U.S. City Average, All Items is used to
adjust all ECI series except for the regional ECI series, which
are adjusted by the corresponding CPI regional indexes. ECI
constant-dollar series based on the 2002 and 2007 NAICS and
the 2000 SOC are available beginning with March 2001 data,
with December 2005 = 100 as the base period.26 Historical
constant-dollar ECI series that use industry categories based
on the SIC and that classify occupations according to the
OCS also are available, dating from the first publication of
each series through December 2005.27 Seasonally adjusted
constant-dollar ECI series are not available.
A historical listing that uses December 2005 = 100 as the
base period is available for all continuous ECI occupational
and industry series that existed before the March 2006
revisions and continued afterwards.28
  23 For available data, see “Employment Cost Trends” (U.S. Bureau of Labor
Statistics), http://www.bls.gov/ect, as well as news releases, for each of the
15 local areas. (For additional information, see Albert E. Schwenk, “BLS
Introduces New Employment Cost Indexes for 14 Metropolitan Areas” (U.S.
Bureau of Labor Statistics, September 24, 2008), http://www.bls.gov/opub/
cwc/cm20080922ar01p1.htm. ECI estimates for the four Census regions
and nine Census divisions are included in tables 6 and 10 of the national
“Employment Cost Index” news release. The Census regions and divisions
are as follows:
Northeast region: New England division—Connecticut, Maine,
Massachusetts, New Hampshire, Rhode Island, Vermont; Middle Atlantic
division—New Jersey, New York, Pennsylvania. Midwest region: East North
Central division— Illinois, Indiana, Michigan, Ohio, Wisconsin; West North
Central division—Iowa, Kansas, Minnesota, Missouri, Nebraska, North
Dakota, South Dakota. South region: South Atlantic division—Delaware,
District of Columbia, Florida, Georgia, Maryland, North Carolina, South
Carolina, Virginia, West Virginia; East South Central division—Alabama,
Kentucky, Mississippi, Tennessee; West South Central division—Arkansas,
Louisiana, Oklahoma, Texas. West region: Mountain division—Arizona,
Colorado, Idaho, Montana, Nevada, New Mexico, Utah, Wyoming; Pacific
division—Alaska, California, Hawaii, Oregon, Washington.
  24 See Employment Cost Index: Historical Listing, Current-dollar, 1975–
2005 (U.S. Bureau of Labor Statistics, May 10, 2006), http://www.bls.gov/
web/echistry.pdf.
  25 See, for example, Employment Cost Index Historical Listing—Volume III
(U.S. Bureau of Labor Statistics, January 2013), http://www.bls.gov/web/
echistrynaics.pdf.
  26 See, for example, Employment Cost Index Historical Listing—Volume IV
(U.S. Bureau of Labor Statistics, January 2013), http://www.bls.gov/web/
ecconstnaics.pdf.
  27 See, for example, Employment Cost Index Historical Listing—Volume II
(U.S. Bureau of Labor Statistics, January 2013), http://www.bls.gov/web/
ecconst.pdf.
  28 See, for example, Employment Cost Index Historical Listing—Volume V
(U.S. Bureau of Labor Statistics, January 2013), http://www.bls.gov/web/
eci/ecicois.pdf.

12

Data from the ECI that provide 12-month percent
changes in employer costs for health insurance in private
industry also are available, from March 1982 to the
present.29
Seasonal adjustment
Over the course of a year, rates of change in the cost of
wages and benefits, as measured in the ECI, reflect events
that follow a more or less regular pattern. These events
include expansions and contractions of economic activity
that occur in specific periods of the year, such as increased
work in the construction industry during warm weather.
As another example, ECI 3-month rates of change for
wage-and-benefit costs in state and local governments,
which include state and local education as a substantial
part, show larger rates of increase in September, reflecting
new contracts associated with the beginning of new school
sessions. Such regular patterns in an economic time
series typically are referred to as seasonal effects. The
process of estimating and removing these effects from an
economic series is called seasonal adjustment. Seasonal
adjustment makes it easier for analysts to observe changes
in data exclusive of seasonal effects. Economists and other
researchers are particularly interested in observing cyclical
and long-run movements of economic series so that they
can better understand the economic behavior of various
sectors of the economy.
In evaluating changes in a seasonally adjusted series,
it is important to note that seasonal adjustment is an
approximation based on past experience. Seasonally
adjusted estimates have a broader margin of error than
the original data on which they are based, because they
are subject to errors associated with seasonal factor
estimation in addition to sampling and nonsampling
errors.
Seasonal adjustment is performed with the X-12
ARIMA program developed by staff of the Statistical
Research Division of the U.S. Census Bureau. The X-12
ARIMA program includes enhancements to both the X-11
variant of the Census Method II seasonal adjustment
program and the X-11 ARIMA program developed by
Statistics Canada.
At the beginning of each calendar year, seasonal
adjustment factors are calculated for use during the coming
year. The seasonal factors for the coming year are published
on the BLS website. Revisions of seasonally adjusted indexes
and 3-month percent changes for the most recent 5 years also
are published on that site.
ECI series are seasonally adjusted by either a direct or
an indirect method. In the direct method, an original, or
unadjusted, index is divided by its seasonal factor. In the
indirect method (also called composite seasonal adjustment),
  29 See, for example, Employment Cost Index: Health Benefits (U.S. Bureau
of Labor Statistics, January 2013), http://www.bls.gov/ect/sp/echealth.pdf.
Note that official ECI series—those designated for use by agencies of the
federal government—are based on SIC and OCS through December 2005
and on NAICS and SOC from March 2006 forward.

the seasonally adjusted index is calculated as a weighted sum
of seasonally adjusted index components.30

had been linked to the annual percent increase in the
General Schedule (GS) federal civil service pay scale
under the Federal Employees Pay Comparability Act of
1990. In 1999, with a widening gap between military and
private industry pay, Congress enacted legislation that
tied annual military pay increases to the annual increase in
the ECI plus 0.5 percent. The legislation was in effect for
Fiscal Years 2001–06.

More information on ECI data
The ECI has been designated a Principal Federal Economic
Indicator by the Office of Management and Budget. It
provides a measure of labor costs that has continuous series
on wages and salaries and on total compensation, as well as
subseries by occupational and industry group. The ECI is
used by the Federal Reserve Board to monitor the effects of
fiscal and monetary policies and to formulate those policies.
It enables analysts and policymakers to assess the effects of
labor cost changes on the economy, both in the aggregate
and by sector. The ECI is particularly important in studies of
the relationships among prices, productivity, labor costs, and
employment. The index also is used to determine increases
in Medicare payments to hospitals and doctors and as a labor
cost escalator in long-term contracts.
In identifying data to be used in contract negotiations,
it is important to note that differences by bargaining status
may be due to factors other than union status, such as the
occupational and industry mix. An important consideration in
choosing a series for escalation is the sampling error.31

• U.S. economic policy decisions. The Federal Reserve
uses the ECI as a major economic indicator for monetary
policy decisionmaking.
• Escalator clauses in collective bargaining contracts.
Wage escalator clauses can allow for a pay increase that is
dependent upon the ECI.
• Escalator clauses in U.S. government contracts. Various
ECI series are used as labor cost escalators in U.S.
government contracts. For example, the Department of
Defense uses both the wages and salaries cost series and
the benefits costs series as escalation factors in numerous
defense contracts, including contracts for computer
research, and the Environmental Protection Agency uses
the series “total compensation for professional and related
workers” as the designated cost escalator in a number of
systems design services.

Examples of ECI data uses
• Federal pay adjustments. The ECI is used to determine
white-collar32 pay adjustments under the Federal Employees Pay Comparability Act.

• Adjustments to Medicare reimbursements for hospital,
physician, and related services. The U.S. Department
of Health and Human Services, Centers for Medicare
and Medicaid Services, uses data from the ECI to determine allowable increases in reimbursements to hospitals,
skilled-nursing facilities, home health care organizations,
and physicians under Medicare’s Prospective Payment
Systems (PPS). The PPS designates the level of reimbursement for Medicare-covered services, in accordance
with the diagnosis and geographic location of care. PPS
reimbursements are adjusted annually on the basis of a
number of factors, including changes in compensation for
medical and related personnel.34

• Active-duty military pay adjustments. In November 2003,
Congress passed a permanent law requiring that annual
basic pay increases for active-duty military personnel be
indexed to the annual increase in the ECI, beginning with
Fiscal Year 2007.33 However, each year since Fiscal Year
2004, Congress has enacted a special “National Defense
Authorization Act” that sets the annual military increase
at a level superseding that of 37 USC 1009. Prior to this
legislation, the wages of active-duty military personnel
  30 For more information about seasonal adjustment issues, see E. Raphael
Branch, James A. Buszuwski, Albert E. Schwenk, and Mark Gough, “Transitional Employment Cost Indexes for seasonal adjustment,” Monthly Labor
Review, April 2008, pp. 25–39, http://www.bls.gov/opub/mlr/2008/04/
art3full.pdf. For more information on the ECI conversion to new industry
and occupational classification systems, see E. Raphael Branch and Lowell
Mason, “Seasonal adjustment in the ECI and the conversion to NAICS and
SOC,” Monthly Labor Review, April 2006, pp. 12–21, http://www.bls.gov/
opub/mlr/2006/04/art3full.pdf. Also informative is E. Raphael Branch,
“Changes in the publication of seasonally adjusted Employment Cost Index
series,” Monthly Labor Review, March 2013, pp. 68–85, http://www.bls.
gov/opub/mlr/2013/03/art5full.pdf.
  31 For more information on choosing a series for escalation, see “Employment Cost Trends: How to Use the Employment Cost Index for Escalation”
(U.S. Bureau of Labor Statistics, June 17, 2008, http://www.bls.gov/ect/
escalator.htm. For information on how to update wage data from any source
to the most recent quarter, see Wayne M. Shelly, “Aging Wage Survey Data
Using the Employment Cost Index” (U.S. Bureau of Labor Statistics, January 29, 2008), http://www.bls.gov/opub/cwc/cm20080122ar01p1.htm.
  32 With the switch to the new occupation (SOC) and industry (NAICS) classification systems, the Bureau of Labor Statistics no longer uses the term
“white collar.” In its place is the designation “professional and related.”
  33 Section 602 of the Fiscal Year 2004 National Defense Authorization Act,
P.L. 108-136, November 24, 2003, and 117 Stat. 1498, amending 37 USC
1009.

• School district property taxes. The Pennsylvania Department of Education uses the ECI as a measure of inflation
to determine the maximum increase allowed for school
district property taxes. Pennsylvania began using the
index in 2005 after the state General Assembly passed,
and the Governor signed into law, Act 72, titled “The
Homeowner Tax Relief Act.” This tax reform initiative was designed to reduce the property tax burden on
Pennsylvania residents.
• Economic price adjustments in long-term purchase
contracts. Long-term purchase contracts may specify that
the ECI is to be used to adjust the labor cost portion of the
contract.
  34 For more information, see “The Employment Cost Index and the
Impact on Medicare Reimbursements” (U.S. Bureau of Labor Statistics),
http://www.bls.gov/ncs/ect/medicare2008_impact.pdf.

13

• Escalator clauses in foreign government contracts. Foreign
governments sometimes use ECI series in contracts with
U.S. firms. For example, the Swiss government uses the
series on durable-goods manufacturing as a wage escalator
in a contract with a U.S. firm that makes computers.
• Economic consulting and forecasting. Various ECI series
are used in models for economic forecasting, including
forecasting ECI values for clients’ use in budgeting and
other activities. ECI series also are used in developing
inflation indexes of personnel costs and other costs for
elementary and secondary schools and for colleges.
ECEC series
The ECEC series measures the average cost to employers
for wages and salaries, and for benefits, per employee hour
worked. As mentioned earlier, the series provides quarterly
data on employer costs per hour worked for total compensation, wages and salaries, total benefits, and the following
benefits: paid leave—vacations, holidays, sick leave, and
personal leave; supplemental pay—premium pay for work
in addition to the regular work schedule (such as overtime,
weekend, and holiday work) and for shift differentials, and
nonproduction bonuses (such as yearend, referral, and attendance bonuses); insurance benefits—life, health, short-term
disability, and long-term disability insurance; retirement and
savings benefits—defined benefit and defined contribution
plans; and legally required benefits—Social Security, Medicare, federal and state unemployment insurance, and workers’
compensation. Cost data are presented both in dollar amounts
and as percentages of total compensation. The ECEC uses
current employment weights to reflect the composition of today’s labor force.
Beginning with the December 2006 data, the NCS
implemented new benchmarking procedures for ECEC
estimates. Current employment weights are used to calculate
cost levels. The weights are derived from two BLS programs:
the QCEW and the CES. Combined, these programs provide
the appropriate industry coverage and currency of data
needed to benchmark all the ECEC series. All other NCS data
products are benchmarked with QCEW data only.35
In most instances, private industry employment weights
used in the ECEC are total employment estimates for twodigit industry groups, such as utilities (NAICS 22) or
wholesale trade (NAICS 42). In a few cases, the employment
weights associated with more detailed industrial categories
are used. Among such categories are the four-digit NAICS
categories elementary and secondary schools (6111), junior
colleges (6122), and colleges and universities (6133), and the
six-digit NAICS category aircraft manufacturing (336411).
For state and local governments, a more aggregated level
reflecting the level of detail published by the CES program
is typically used.
  35 For more information, see “Changes in Calculations for the BLS Employer Costs for Employee Compensation Data, March 2007” (U.S. Bureau
of Labor Statistics), http://www.bls.gov/ncs/ect/sp/ececcalc.pdf.

14

For private and government establishments, employment
data were apportioned on the basis of the sampling weights
assigned to the ECI sample. The ECI, which measures the
change in employer costs for employee compensation, is
calculated with fixed 2002 employment counts in order
to prevent employment shifts among occupations and
industries from influencing the measurement. Therefore,
changes over time in the ECEC series will differ from those
in the ECI.
Historical ECEC data appear in three listings.36 The
first historical listing covers data for the March reference
periods from 1986 to 2001. These data use the SIC and
OCS classification systems. The second listing contains data
for the March, June, September, and December reference
periods from March 2002 to December 2003. These data also
are based on SIC and OCS. The final listing includes data
from March 2004 to the current reference period. These data
are based on NAICS and SOC. Beginning with the quarter
including March 2004, historical data based on NAICS and
on SOC 2000 became available. The new historical tables are
available on request.37
ECEC estimates are shown as costs per hour worked for
total compensation (wages and benefits), expressed both
as dollar amounts and as percentages of compensation.
ECEC estimates are computed for various costs c, including
wages, individual benefits, benefit groups (such as paid
leave), combinations of benefits, total benefits, and total
compensation (total wages plus total benefits).
The ECEC estimates of percentage of total compensation
are calculated from cost aggregates and then rounded to
the first decimal place. This method provides the most
precise estimates of the percentage of total compensation;
estimates calculated from published cost estimates may
differ slightly from those calculated from unpublished cost
aggregates.
The formula for the mean hourly cost c for domain D is

∑𝑞𝑞∈𝐷𝐷 𝑊𝑊𝑞𝑞′ 𝑌𝑌�𝑐𝑐𝑐𝑐
�
𝑌𝑌𝑐𝑐𝑐𝑐 =
,
∑𝑞𝑞∈𝐷𝐷 𝑊𝑊𝑞𝑞′
where

D is the domain of interest,
Wq’ is the final quote weight for quote q, calculated as the
product of the inverse of the selection probabilities at each
stage of sampling, with one additional factor included to
account for changes in the employment distribution, and

  36 See “Employment Cost Trends” (U.S. Bureau of Labor Statistics), section
titled “ECEC Listings,” http://www.bls.gov/ect/#tables.
  37 See also “Employment Cost Trends” (U.S. Bureau of Labor Statistics),
http://www.bls.gov/ncs/ect/home.htm.) Information on how costs are calculated appears in Albert E. Schwenk, “Measuring Trends in the Structure
and Levels of Employer Costs for Employee Compensation,” Compensation and Working Conditions, summer 1997, pp. 3–14, http://www.bls.gov/
opub/cwc/archive/summer1997art1.pdf.

An example of ECEC data use
• Costs associated with employee compensation. The
International Union of United Automobile, Aerospace
and Agricultural Implement Workers of America, also
known as the United Auto Workers (UAW), has posted
tables using ECEC data on its website. The data are
given in “The Union Advantage—September 2010: The
Union Advantage at a Glance,” which has appeared every
quarter since December 2006. The tables highlight wage
and benefit data from the ECEC for union and nonunion
workers in private industry and for goods-producing and
service-providing industries.40

� 𝑐𝑐𝑐𝑐 is the mean hourly cost c for quote q.
𝑌𝑌

In addition, the formula for the mean hourly cost c as a
percentage of total compensation is

𝑃𝑃𝑐𝑐𝑐𝑐 =
where

𝑌𝑌� 𝑐𝑐𝑐𝑐
× 100,
𝑌𝑌�
𝑇𝑇𝑇𝑇

� 𝑐𝑐𝑐𝑐 is the mean hourly cost c for domain D, as before, and
𝑌𝑌

Incidence and Provisions of Benefits
The NCS collects and publishes data on the incidence of
employer-provided benefits and on the key provisions (terms)
of employee benefit plans, for civilian workers (as defined
by the NCS), workers in private industry, and state and local
government workers. The data on incidence (access to and
participation in employee benefit plans) and key provisions
are published annually. The data on employer-provided
benefits include the following:

� 𝑇𝑇𝑇𝑇 is the mean hourly cost for total compensation for
𝑌𝑌

domain D.

More information on ECEC data
Differences in the estimates for the state and local government
and private industry sectors stem from factors such as
variation in work activities and in occupational structures.
Manufacturing and sales, for example, make up a large part of
private industry work activities but are rare in state and local
government. In contrast, professional and administrative
support occupations (including teachers) account for twothirds of the state and local government workforce but less
than one-half of private industry.38
When respondents do not provide all the data needed,
a procedure for assigning missing values is used. This
imputation procedure is comparable to that used for the
ECI.39

• Health care (medical, dental, vision, and prescription drug
plan coverage, and employee and employer premiums
for individual coverage and family coverage) and the
percentage of establishments offering health benefits;
• Retirement plan coverage (defined benefit and defined
contribution) and the percentage of establishments
offering retirement benefits;
• Life, short-term disability, and long-term disability insurance coverage;
• Paid leave (sick, vacation, jury duty, personal, and family),
paid holidays, unpaid family leave, and nonproduction
bonuses and stock options;

  A detailed examination of differences in compensation levels and trends
between private industry and state and local government is found in Bradley R. Braden and Stephanie L. Hyland, “Cost of employee compensation in public and private sectors,” Monthly Labor Review, May 1993, pp.
14–21, http://www.bls.gov/opub/mlr/1993/05/art2full.pdf, and in Albert
E. Schwenk, “Compensation Cost Trends in Private Industry and State and
Local Governments,” Compensation and Working Conditions, fall 1999, pp.
13–18, http://www.bls.gov/opub/cwc/archive/fall1999art2.pdf.
For more information on the ECEC calculation procedure, see “Changes
in Variance Estimation Calculations for the BLS Employer Costs for
Employee Compensation Data, March 2007” (U.S. Bureau of Labor
Statistics), http://www.bls.gov/ncs/ect/sp/ececvmet.pdf.
Relative standard errors for all estimates are available to users. (See
“Employment Cost Trends” (U.S. Bureau of Labor Statistics), section titled
“ECT Databases,” http://www.bls.gov/ncs/ect/#tables), published shortly
after the quarterly news release is issued.
For a detailed explanation of how to use standard error data to
analyze differences in changes over time, see Martha A. C. Walker and
Bruce J. Bergman, “Analyzing Year-to-Year Changes in Employer Costs for
Employee Compensation,” Compensation and Working Conditions, spring
1998, pp. 17–27, http://www.bls.gov/opub/cwc/archive/spring1998art3.
pdf. This article supplements Michael K. Lettau, Mark A. Loewenstein, and
Aaron T. Cushner, “Explaining the Differential Growth Rates of the ECI
and the ECEC,” Compensation and Working Conditions, summer 1997, pp.
15–23, http://www.bls.gov/opub/cwc/archive/summer1997art2.pdf, which
examines how differences in the construction of these measures contribute
to differing trends.
  39 For a description, see Song Yi, “Accounting for missing data in the
Employment Cost Index,” Monthly Labor Review, April 2006, pp. 22–27,
http://www.bls.gov/opub/mlr/2006/04/art4full.pdf.
38

• Health promotion benefits;
• Pretax benefits; and
• “Quality of life” benefits, such as long-term care insurance,
a flexible-workplace option, and subsidized commuting.
In addition, the NCS publishes detailed provisions
of coverage in two major benefit areas: health insurance
and retirement plans. A basic set of tables listing detailed
provisions of these plans is published annually, with
additional tables published on a rotating basis. Health
data include medical plan provisions, such as deductibles,
coinsurance, and out-of-pocket maximums, as well as details
of dental, vision, and prescription drug benefits. Provisions
of defined benefit and defined contribution retirement plans,
such as eligibility requirements and benefit formulas, also
are published.
  40 Visit http://www.uaw.org/sites/default/files/unionadvantage0910.pdf.
Note: Links to non-BLS websites are provided for the reader’s convenience
and do not constitute an endorsement.

15

Medical premiums. Estimates of employer and employee
medical premiums are for participants in all medical plans,
with calculations for both individual and family coverage.
The calculations are based on the assumption that all
employees in the occupation have identical coverage, rather
than on actual decisions regarding medical coverage made by
employees within the occupations.
Leave benefits for teachers. Primary, secondary, and special
education teachers typically have a work schedule of 37 or 38
weeks per year. Because of this schedule, they generally are
not offered paid vacation or holidays. In many cases, time off
during winter and spring breaks during the school year is not
considered paid vacation days for the purposes of this survey.
The NCS measures of employer-provided benefits are as
follows:41
• Incidence of benefits. The percentage of all workers who
are provided a particular benefit plan. The incidence can
be either a rate of access to, or a rate of participation in, a
benefit plan.
• Provisions of benefits. The terms of a benefit plan. For
example, a medical plan might charge a $20 copayment
for a doctor’s office visit.
• Access to a benefit plan. Employees are considered to
have access to a benefit plan if the plan is available for
their use.
• Participation in a benefit plan. Employees in contributory
plans are deemed to be participating in an insurance or
retirement plan if they have paid required contributions
and fulfilled any applicable service requirements. Employees in noncontributory plans are counted as participating regardless of whether they have fulfilled the service
requirements.
• Takeup rates. Takeup rates are the percentages of
workers with access to a plan who participate in the
plan. Takeup rates are computed as the number of
workers participating in a plan, divided by the number
of workers with access to the plan, times 100 and
rounded to the nearest whole percentage. Because the
computation of takeup rates is based on the number of
workers, rather than the rounded percentages, the takeup
rates in published tables may differ slightly from the
ratio of participation to access.
• Establishments offering a benefit. The percentage of
establishments (instead of employees, as in the majority
of benefits tables) offering a given benefit.
Formula used to calculate NCS estimates of access to
benefits. The formula for the percentage of employees with
access to a benefit area, such as life insurance, for domain
D is
  41 For a list of NCS terms and definitions pertaining to benefits, see National
Compensation Survey: Glossary of Employee Benefits Terms (U.S. Bureau of
Labor Statistics, July 2012), http://www.bls.gov/ncs/ebs/glossary20112012.
htm and http://www.bls.gov/ncs/ebs/glossary20112012.pdf.

16

𝐴𝐴𝐷𝐷 =
where

∑𝑞𝑞∈𝐷𝐷 𝑊𝑊𝑞𝑞′ 𝑋𝑋𝑞𝑞
× 100,
∑𝑞𝑞∈𝐷𝐷 𝑊𝑊𝑞𝑞′

D is the domain of interest,
Wq’ is the final weight for quote q, calculated as described in
the section on the calculation of ECEC estimates, and
Xq is 1 if the worker in quote q has access to the benefit being
estimated and 0 otherwise.
Incidence (participation). The formula for the incidence, or
percentage, of employees participating in a benefit area, such
as medical care, for domain D is

∑𝑞𝑞∈𝐷𝐷 ∑𝑗𝑗 ∈𝑞𝑞 𝑊𝑊𝑞𝑞′ 𝑃𝑃𝑞𝑞𝑞𝑞
𝐼𝐼𝐷𝐷 =
× 100,
∑𝑞𝑞∈𝐷𝐷 𝑊𝑊𝑞𝑞′
where

D is the domain of interest,
Wq’ is the final quote weight for quote q, calculated as described
in the section on the calculation of ECEC estimates, and
Pqj is the percentage of workers in quote q who are
participating in plan j.
Other estimates of incidence, such as the percentage
of participants in a benefit area or in a subset of a benefit
area, can be computed in a similar manner, such that the
base includes only those workers who participate in the
benefit. For example, to calculate the percentage of medical
insurance participants in fee-for-service plans in domain D,
a ratio is calculated such that the denominator is the same as
the numerator in the previous formula and the numerator is of
the same form as well, except that the summation is restricted
to those participants in fee-for-service plans.
Average (mean). The formula for the average flat monthly
employee contribution for medical insurance for domain D
is

Y
𝐼𝐼𝐷𝐷D =

where

∑𝑞𝑞∈𝐷𝐷 ∑𝑗𝑗 ∈𝑞𝑞 𝑊𝑊𝑞𝑞′ 𝑌𝑌𝑞𝑞𝑞𝑞 𝑃𝑃𝑞𝑞𝑞𝑞
,
∑𝑞𝑞∈𝐷𝐷 ∑𝑗𝑗 ∈𝑞𝑞 𝑊𝑊𝑞𝑞′ 𝑃𝑃𝑞𝑞𝑞𝑞

D is the domain of interest,
Wq’ is the final quote weight for quote q, calculated as
described in the section on the calculation of ECEC estimates,

Yqj is the average monthly employee contribution to plan j by
workers in quote q, and

• the weighted plan employment (WPEigj) across plans with
a value less than Qigj is less than p percent of the total
weighted plan employment and

Pqj is the percentage of workers in quote q who are
participating in plan j.

• the weighted plan employment (WPEigj) across plans with
a value more than Qigj is less than (100 − p) percent of the
total weighted plan employment.

Other means, such as the average annual deductible for
medical insurance, can be calculated by a similar formula. In
all cases, the averages include only those workers with the
provision in question.
The weighted count of workers participating in plans offered to workers in the sampled occupation and establishment is calculated by multiplying the final benchmarked
quote weight by the participation rate for only those plans
in the quote that meet the specific conditions defined by the
quote condition and the plan conditions. The formula is

It is possible that there are no specific plan records igj for
which both of these properties hold. This occurs when there
exists a plan for which the WPEigj of records whose value
is less than Qigj equals p percent of the total weighted plan
employment. In that situation, the pth percentile is the average
of Qigj and the value on the record with the next-lowest value.
The Qigj values must be sorted in ascending order.
More information on NCS benefits data
Standard errors are available for incidence estimates from
2008 onward. In 2009, the NCS published its first benefits
estimates that include imputed data.42

WPEigj = (OccFWig) × Xig Yigj × Zigj × Pigj,
where
i = establishment,

Examples of Use of NCS Benefits Data
• Planning and improving company benefits. NCS data
commonly are used as a guide when companies choose the
provisions for their benefit plans. In addition, companies
may improve benefit packages to remain competitive in
the labor market. For example, a computer company may
have a difficult time finding qualified computer engineers,
or a car dealership may not be able to attract the best
salesperson. Instead of simply raising the wage, many
companies will enhance or add benefits.

g = occupation within establishment i,
j = plan in occupation g within establishment i,
WPEigj = weighted plan employment of record igj,
OccFWig = final benchmarked quote weight for occupation g
in establishment i,
Xig, Yigj, and Zigj are dummy variables such that

{
{
{

X = 1 if quote ig meets the condition set in the quote (row)
ig
 condition
0 otherwise,

• Lowering turnover rates. To attract and retain workers,
employers may provide additional benefits. These prospective benefits may be traditional or emerging. Employers can search NCS benefits data to evaluate benefits that
employees currently are being offered nationwide.

Yigj= 1 if plan igj meets the condition set in the base
 (denominator) plan condition
0 otherwise,

• Aiding collective bargaining negotiations. Collective bargaining units renegotiate their contracts at various times.
The bargaining unit may want to add a new benefit, such
as subsidized commuting, to an agreement. The bargaining unit and the employer can use NCS benefits data to
assist them in making decisions.

Zigj= 1 if plan igj meets the condition set in the additional
 (numerator) plan condition
0 otherwise, and
Pigj = percentage of workers in occupation g and establishment
i who are participating in plan j.

• Understanding health benefits data. Health benefits data
are broken out into average contributions for medical
coverage and average plan limits. A new company can
reference these averages when it selects group health
plan coverage, comparing the averages with proposals
that health plan companies have given the new company.
An established company can compare the premiums
it currently pays for health benefits with nationwide
averages. The comparison helps the established company

Calculation of percentiles
Percentiles of benefit provisions are calculated with data only
from those workers in plans that include the provision in
question. Percentile data are used to describe the distribution
of a numeric value, such as a median annual deductible of
$400 and the value $600 at the 90th percentile. The following
percentiles p are calculated: 10, 25, 50 (median), 75, and 90.

  42 For the latest news release on benefits, see “Employee Benefits Survey”
(U.S. Bureau of Labor Statistics), http://www.bls.gov/ebs; for more
information on NCS imputation with benefits data, see Omolola E. Ojo and
Jonathan J. Lisic, “BLS Resumes Estimation of Sample Errors for Benefits
Measures,” Compensation and Working Conditions, May 22, 2008, http://
www.bls.gov/opub/cwc/cm20080520ar01p1.htm.

The pth percentile is the value Qigj, where the plan value of a
quantity is for a specific benefit or a subset of a benefit area,
such that
17

assess its health benefits or negotiate contracts with health
benefit companies.

on earnings, employer costs of compensation, and benefits,
the NCS provides a unique measurement of the labor market,
allowing employer costs to be linked to estimates of specific
plan provisions.
Also, federal, state, and local government agencies use
NCS estimates in administering compensation programs
and in formulating public policy on compensation.
The data are of value to federal and state mediation
and conciliation services and to state employment
compensation agencies in judging the suitability of job
offers. In addition, NCS data are used by government
agencies to

• Assessing and formulating public policy. NCS benefits
data were used to design defined benefit plans and
savings and thrift plans for federal employees. In the
recent debate over a universal health care system, benefits
data on employee premium sharing were considered in
formulating proposals. Data on the amount of retirement
income from employer plans have helped to frame the
debate on the subject of Social Security reform.
• Researching current benefit issues. Students, consultants,
and researchers use benefits data frequently. Students
may be writing a thesis or trying to identify a noteworthy
item on which to focus an assignment. Consultants
may be trying to recommend benefit-related actions
to a company or provide supporting data to clients.
Researchers sometimes want to investigate a particular
issue pertaining to benefits or may focus on a few years
of previous data to develop research on trends or other
benefit issues.

• Evaluate benefits packages,
• Analyze contract settlements,
• Aid in collective bargaining negotiations, and
• Index Medicare payments.
NCS data are used in private industry to
• Adjust wages in long-term contracts,

Use and Limitations of the Data

• Evaluate benefit packages,

NCS data have a variety of uses.43 One use is in economic
analysis, in which a knowledge of levels, structures, and
trends of pay rates and benefit practices is required in
analyzing current economic developments and in studying
wage dispersion and differentials. By using the same survey
methodology and definitions as it goes about collecting data

• Analyze contract settlements,
• Aid in collective bargaining negotiations,
• Guide decisions in locating businesses or plants, and
• Assist in administering wages and salaries.
Although NCS compensation measures have many uses,
their limitations must be kept in mind. The data are subject
to sampling error, which may cause deviations from the
results that would be obtained if the actual records of all
establishments could be used. Nonsampling error is present in
surveys as well. (See the section “Calculation and Reliability
of the Estimates” for more information.)

  43 See, for example, “Employment Cost Trends: How to Use the Employment
Cost Index for Escalation” (U.S. Bureau of Labor Statistics, June 17, 2008),
http://www.bls.gov/ect/escalator.htm. For additional information, see the
subsection “Examples of Data Use” for each data product in the section
“Calculation and Reliability of the Estimates.” For an overview, see Natalie
Kramer, “Earnings and Other Compensation Data at BLS: What Users Seek
and What We Offer,” Compensation and Working Conditions, February 26,
2003, http://www.bls.gov/opub/cwc/cm20030224ar01p1.htm.

18

Technical References
Branch, E. Raphael, and Lowell Mason, “Seasonal
adjustment in the ECI and the conversion to NAICS and
SOC,” Monthly Labor Review, April 2006, pp. 12–21,
http://www.bls.gov/opub/mlr/2006/04/art3full.pdf.

Area sample selection
The criteria for defining Metropolitan, Micropolitan,
and Combined Statistical Areas are published in the
Federal Register (65 FR 82228–82238, December
27, 2000), http://www.whitehouse.gov/omb/fedreg/
metroareas122700.pdf.

Survey concepts
Braden, Bradley R., and Stephanie L. Hyland, “Cost of
employee compensation in public and private sectors,”
Monthly Labor Review, May 1993, pp. 14–21, http://
www.bls.gov/opub/mlr/1993/05/art2full.pdf.

The criteria for defining Core Based Statistical Areas are
published in the Federal Register (65 FR 82228–82238,
December 27, 2000), http://www.whitehouse.gov/
omb/fedreg/metroareas122700.pdf.

Buckley, John E., “Fifty years of BLS surveys on Federal employees’ pay,” Monthly Labor Review, September 2009, pp. 36-46, http://www.bls.gov/opub/
mlr/2009/09/art3full.pdf.

Izsak, Yoel, Lawrence R. Ernst, Erin McNulty, Steven
P. Paben, Chester H. Ponikowski, Glenn Springer,
and Jason Tehonica, “Update on the Redesign of
the National Compensation Survey,” Proceedings
of the American Statistical Association, Section on
Survey Research Methods. Alexandria, VA: American
Statistical Association, 2005, http://www.bls.gov/ore/
pdf/st050140.pdf.

Buckley, John E., “Pay Relatives for U.S. Census Regions
and Divisions, 2006,” Compensation and Working Conditions, August 28, 2008, http://www.bls.gov/opub/
cwc/cm20080826ar01p1.htm.

Izsak, Yoel, Lawrence R. Ernst, Steven P. Paben, Chester H.
Ponikowski, and Jason Tehonica, “Redesign of the
National Compensation Survey,” Proceedings of the
American Statistical Association, Section on Survey
Research Methods. Alexandria, VA: American Statistical Association, 2003, pp. 1978–1985, http://www.
amstat.org/sections/srms/proceedings/y2003/Files/
JSM2003-000382.pdf.

Ford, Jason L., “The New Health Participation and Access
Data from the National Compensation Survey,” Compensation and Working Conditions, October 26, 2009,
http://www.bls.gov/opub/cwc/cm20091022ar01p1.
htm.
Gittleman, Maury B., “Pay relatives for metropolitan areas
in the NCS,” Monthly Labor Review, March 2005, pp. 46–
53, http://www.bls.gov/opub/mlr/2005/03/art4full.pdf.

Tehonica, Jason, “New Area Sample Selected for the
National Compensation Survey,” Compensation and
Working Conditions, April 25, 2005, http://www.bls.
gov/opub/cwc/cm20050318ar01p1.htm.

Kramer, Natalie, and Alan Zilberman, “New Definitions of
Employee Access to Paid Sick Leave and Retirement
Benefits in the National Compensation Survey,” Compensation and Working Conditions, December 23, 2008,
http://www.bls.gov/opub/cwc/cm20081219ar01p1.
htm.

Tehonica, Jason, Lawrence R. Ernst, and Chester H.
Ponikowski, “Phase-in of the Redesigned National
Compensation Survey Area Sample,” Proceedings
of the American Statistical Association, Section on
Survey Research Methods. Alexandria, VA: American
Statistical Association, 2005, pp. 2993–2997, https://
www.amstat.org/sections/srms/Proceedings/y2005/
Files/JSM2005-000156.pdf.

Lettau, Michael K., Jonathan Lisic, Jesus Ranon, Bradley
D. Rhein, Thuy T. Shipp, and Sarah J. Stafira, “Local
Area Employee Benefits Estimates for 15 Metropolitan Areas,” Compensation and Working Conditions,
September 28, 2009, http://www.bls.gov/opub/cwc/
cm20090924ar01p1.htm.

Benchmarking
“Changes in Calculations for the BLS Employer Costs
for Employee Compensation Data, March 2007” (U.S.
Bureau of Labor Statistics), http://www.bls.gov/ncs/
ect/sp/ececcalc.pdf.

Moehrle, Thomas G., “The Cost and Incidence of Referral,
Hiring, and Retention Bonuses,” Compensation and
Working Conditions, winter 2000, pp. 37–42, http://
www.bls.gov/opub/cwc/archive/winter2000art4.pdf.

Seasonal adjustment
Branch, E. Raphael, James A. Buszuwski, Albert E.
Schwenk, and Mark Gough, “Transitional Employment
Cost Indexes for seasonal adjustment,” Monthly Labor
Review, April 2008, pp. 25–39, http://www.bls.gov/
opub/mlr/2008/04/art3full.pdf.

Morton, John, “Variable Pay in the BLS National Compensation Survey,” Compensation and Working Conditions, January 30, 2003, http://www.bls.gov/opub/cwc/
cm20030121yb02p1.htm.

19

National Compensation Survey: Guide for Evaluating
Your Firm’s Jobs and Pay (U.S. Bureau of Labor Statistics, October 2003), http://www.bls.gov/ncs/ocs/sp/
ncbr0004.pdf.
Schumann, Richard, “Work schedules in the National
Compensation Survey,” Compensation and Working
Conditions, July 22, 2008, http://www.bls.gov/opub/
cwc/cm20080722ar01p1.htm.
Schwenk, Albert E., “Compensation Cost Trends in Private
Industry and State and Local Governments,” Compensation and Working Conditions, fall 1999, pp. 13–18,
http://www.bls.gov/opub/cwc/archive/fall1999art2.
pdf.
Weinstein, Harriet, and Elizabeth Dietz, “Towards a
Working Definition of Compensation,” Compensation
and Working Conditions, June 1996, pp. 3–9, http://
www.bls.gov/opub/cwc/archive/summer1996art1.
pdf.
Imputation methods
Yi, Song, “Accounting for missing data in the Employment
Cost Index,” Monthly Labor Review, April 2006, pp.
22–27, http://www.bls.gov/opub/mlr/2006/04/art4full.
pdf.

Walker, Martha A.C., and Bruce Bergman, “Analyzing
Year-to-Year Changes in Employer Costs for Employee
Compensation,” Compensation and Working Conditions,
spring 1998, pp. 17–27, http://www.bls.gov/opub/cwc/
archive/spring1998art3.pdf.
Use of fixed weights and estimation of indexes
Costo, Stephanie L., “Introducing 2002 weights for the
Employment Cost Index,” Monthly Labor Review, April
2006, pp. 28–32, http://www.bls.gov/opub/mlr/2006/04/
art5full.pdf.
Ruser, John W., “The Employment Cost Index: what is
it?” Monthly Labor Review, September 2001, pp. 3–16,
http://www.bls.gov/opub/mlr/2001/09/art1full.pdf.
Schwenk, Albert E., “Measuring Trends in the Structure
and Levels of Employer Costs for Employee Compensation,” Compensation and Working Conditions, summer
1997, pp. 3–14, http://www.bls.gov/opub/cwc/archive/
summer1997art1.pdf.
Schwenk, Albert E., “BLS Introduces New Employment
Cost Indexes for 14 Metropolitan Areas,” Compensation
and Working Conditions, September 24, 2008, http://
www.bls.gov/opub/cwc/cm20080922ar01p1.htm.

Stafira, Sarah, “Recent Modification of Imputation Methods
for National Compensation Survey Benefits Data,” Compensation and Working Conditions, August 28, 2009,
http://www.bls.gov/opub/cwc/cm20090825ar01p1.
htm.

NCS conversion to new industry and occupational
classification systems
Branch, E. Raphael, and Lowell Mason, “Seasonal
adjustment in the ECI and the conversion to NAICS and
SOC,” Monthly Labor Review, April 2006, pp. 12–21,
http://www.bls.gov/opub/mlr/2006/04/art3full.pdf.

Measurement of variance
“Changes in Variance Estimation Calculations for the BLS
Employer Costs for Employee Compensation Data,
March 2007” (U.S. Bureau of Labor Statistics), http://
www.bls.gov/ncs/ect/sp/ececvmet.pdf.

Smith, James E., and Robert W. Van Giezen, “Change
Comes to the National Compensation Survey Locality
Wage Bulletins,” Compensation and Working Conditions, January 24, 2007, http://www.bls.gov/opub/cwc/
cm20070122ar01p1.htm.

Lettau, Michael K., Mark A. Loewenstein, and Aaron T.
Cushner, “Explaining the Differential Growth Rates of
the ECI and the ECEC,” Compensation and Working
Conditions, summer 1997, pp. 15–23, http://www.bls.
gov/opub/cwc/archive/summer1997art2.pdf.

Weinstein, Harriet G., and Mark A. Loewenstein, “Comparing Current and Former Industry and Occupation
ECEC Series,” Compensation and Working Conditions, August 25, 2004, http://www.bls.gov/opub/cwc/
cm20040823ar01p1.htm.

Ojo, Omolola E., and Jonathan J. Lisic, “BLS Resumes
Estimation of Sample Errors for Benefits Measures,”
Compensation and Working Conditions, May 22, 2008,
http://www.bls.gov/opub/cwc/cm20080520ar01p1.
htm.

Use of NCS data products
Buckley, John E., “Beyond Basic Benefits: Employee Access
to Other Types of Benefits, 1979–2008,” Compensation
and Working Conditions, May 29, 2009, http://www.bls.
gov/opub/cwc/cm20090527ar01p1.htm.

Schwenk, Albert E., “Measuring Trends in the Structure
and Levels of Employer Costs for Employee Compensation,” Compensation and Working Conditions, summer
1997, pp. 3–14, http://www.bls.gov/opub/cwc/archive/
summer1997art1.pdf.

Buckley, John E., “Recent Modifications of Employee
Benefits Data in the National Compensation Survey,”
Compensation and Working Conditions, May 29, 2009,
http://www.bls.gov/opub/cwc/cm20090518ar01p1.
htm.

20

Kramer, Natalie, “Earnings and Other Compensation Data at
BLS: What Users Seek and What We Offer,” Compensation and Working Conditions, February 26, 2003, http://
www.bls.gov/opub/cwc/cm20030224ar01p1.htm.

Shelly, Wayne M., “Aging Wage Survey Data Using the
Employment Cost Index,” Compensation and Working
Conditions, January 29, 2008, http://www.bls.gov/opub/
cwc/cm20080122ar01p1.htm.

Schwenk, Albert E., “BLS Introduces New Employer Costs
for Employee Compensation Data for Private Industry
Workers in 15 Metropolitan Areas,” Compensation and
Working Conditions, September 28, 2009, http://www.
bls.gov/opub/cwc/cm20090921ar01p1.htm.

Wiatrowski, William J., “BLS at 125: using historic principles to track the 21st-century economy,” Monthly Labor
Review, June 2009, pp. 3–25, http://www.bls.gov/opub/
mlr/2009/06/art1full.pdf.

21

Appendix: Major Work Stoppages Program

T

The number of days of idleness is computed by multiplying
the number of workers idled during the reference month by
the number of lost workdays, based on a 5-day workweek
(Monday through Friday), excluding federal holidays. If
applicable, the cumulative number of days of idleness also
is computed for each work stoppage beyond the beginning
reference month.

he Bureau compiles data on work stoppages—strikes
or lockouts—involving 1,000 or more workers for at
least a full day or shift. Such data have been collected
since 1947, with detailed information available since 1993.
Detailed data on work stoppages include monthly and annual
listings of companies and governments involved in a work
stoppage, along with the name of the union involved in the
dispute, the location of the stoppage, the North American
Industry Classification System (NAICS) code, the beginning
and ending dates of the dispute, the number of lost workdays
during the reference month and the cumulative number of
lost workdays from the beginning of the work stoppage, the
number of workers idled by the stoppage, and the number
of days of idleness during the reference month and the
cumulative number of days of idleness from the beginning of
the work stoppage.

Sources of information
Information on work stoppages is obtained from reports
from the Federal Mediation and Conciliation Service, state
labor market information offices, BLS Strike Reports from
the Office of Employment and Unemployment Statistics,
and media sources. One or both parties involved in the work
stoppage (the employer, the union, or some other organization)
is contacted to verify the duration of the stoppage and number
of workers idled thereby.

Definitions and methods
A work stoppage is a strike or a lockout. Because of the
complexity of most labor–management disputes, the Bureau
makes no attempt to distinguish between strikes and lockouts
in its statistics. A strike is a temporary stoppage of work by a
group of employees to express a grievance, enforce a demand,
or protest the terms, conditions, or provisions of a contract. A
lockout is a temporary withholding or denial of employment
by management, typically during a labor dispute. The group
of employees involved in a strike or lockout may or may not
be members of a union.
The number of lost workdays is based on a 5-day
workweek (Monday through Friday), excluding weekends
and federal holidays. If applicable, the cumulative number
of lost workdays also is computed for each work stoppage
beyond the beginning reference month.
The number of workers involved includes all workers
made idle for one shift or longer in establishments directly
involved in a stoppage. Workers involved include those who
initiate the strike, as well as others in the establishment who
honor picket lines or are idled because the facility is closed
down. This number does not account for secondary idleness—
that is, the effects of a stoppage on other establishments or
industries whose employees may be made idle as a result of
shortages of material or services.

22

Availability of data
Data for the major work stoppages series are uninterrupted
and date back to 1947.44 Detailed monthly data include both
monthly and cumulative totals for each work stoppage, as
well as the number of stoppages beginning and in effect
during the month. Annual data present the cumulative
totals for the calendar year. You may obtain a searchable
worksheet containing data on major work stoppages
from 1993 to the present by sending an e-mail request to
[email protected].
From 1947 to 2007, the Bureau of Labor Statistics acted
under the mandate of the Taft–Hartley Act to solicit collective
bargaining agreements and make them available in a publicly
accessible file. In September 2007, responsibility for the
maintenance of collective bargaining agreements and for
the continued collection of those agreements was officially
moved to the U.S. Department of Labor, Office of Labor–
Management Standards.45
  44 For statistics on monthly and annual work stoppages and for detailed
monthly data since 1993, see “Work Stoppages” (U.S. Bureau of Labor
Statistics), http://www.bls.gov/wsp.
  45 For more information, see Drew M. Simmons, “Collective Bargaining Agreements File Moves to New Home,” Compensation and Working Conditions Online, November 30, 2007, http://www.bls.gov/opub/cwc/
cb20071128ar01p1.htm.