Employment Statistics: Differences and Similarities in Job-based and Person-based Employment and Unemployment Estimates

CRS Report for Congress
Employment Statistics: Differences and
Similarities in Job-based and Person-based
Employment and Unemployment Estimates
October 19, 2004
Julie M. Whittaker
Analyst in Applied Microeconomics
Domestic Social Policy

Congressional Research Service ˜ The Library of Congress

Employment Statistics: Differences and Similarities in
Job-based and Person-based Employment and
Unemployment Estimates
Summary
Employment statistics are key indicators of the performance of the economy,
measuring how many jobs exist in the economy as well as how individuals fare in the
labor market. Payroll employment estimates track the number of wage and salary
jobs created and lost in the economy. In comparison to this job-based measure,
(un)employment statistics measure the quantities and ratios of those individuals who
are employed or unable to find employment despite actively seeking jobs.
This paper defines and distinguishes two commonly used groups of employment
estimates: payroll employment (derived from a job-based survey) and employment
and unemployment (derived primarily from a person-based survey) statistics. (A
third unemployment statistic derived from the Unemployment Compensation system,
the Insured Unemployment Rate, is less frequently used as it excludes many groups
of workers.) The report then examines two somewhat counter-intuitive situations:
how there may be aggregate job loss with decreased unemployment rates and how
increased payroll employment does not necessarily imply decreased unemployment.
Payroll employment figures are estimated from a survey of employers in non-
farm industries. These figures directly correspond to each individual job held within
the economy and not to individual workers. A particular worker may hold more than
one position and each of those jobs will be counted as a separate position, increasing
the total number of jobs reported through the payroll employment estimates. In
comparison, individual employment and unemployment statistics measures are based
on individuals and their labor force status. The individual statistics are derived from
the non-institutionalized civilian population. The self-employed, those temporarily
absent from work but who have a job, and certain farm workers are also counted as
workers (although their positions are not counted by payroll employment). A worker
may have one or more jobs, but will be counted only once as employed while a
worker without any job will be considered unemployed.
Although it may seem reasonable to expect that as the number of payroll jobs
increases there would be a similar trend in the number of persons employed, the job-
based measure of payroll employment does not require that the person-based measure
of (un)employment statistics move in the same direction. This paper provides two
examples of these somewhat counter-intuitive situations.
This report will be updated as events warrant.

Contents
Creation of Payroll Employment and (Un)employment Statistics.........1
Data Sources.................................................2
CPS ....................................................2
CES ....................................................3
Calculation of Payroll Employment and Employment and
Unemployment Statistics....................................3
National, State, and Local Payroll Employment Statistics..........3
National Employment and Unemployment Statistics and
Unemployment Rate...................................3
State and Local Employment and Unemployment (Local Area
Unemployment Statistics)...............................4
The Total Unemployment Rate (TUR) and Insured Unemployment
Rate (IUR)...............................................4
What is the Total Unemployment Rate (TUR)?..................4
What Does the Insured Unemployment Rate (IUR) Measure?.......4
What Is the Difference Between the IUR and the TUR?............4
Definitions of Employment......................................5
In the Labor Force and Out of the Labor Force...................5
Employed ................................................5
Payroll Employment.......................................5
Unemployed ..............................................6
Job Loss and Job Gain......................................6
Involuntarily Part-time......................................6
Discouraged Workers.......................................6
Example: Decreased Employment but Increasing Jobs.............7
Total employment as measured by the CES and the CPS...............8
How Falling Employment and a Growing Payroll Employment May
Coexist .................................................10
List of Figures
Figure 1. Current Employment Survey and Current Population Survey
Estimated Employment in Thousands (Seasonally Adjusted)
January 1970-March 2004...................................9
List of Tables
Table 1. CPS and CES Employment and Unemployment Definitions.........7
Table 2. Example: The Employment Situation for September 2004...........8
Table 3. Hispanic Employment and Unemployment Rate, 2000 and 2003.....10
Table 4. The Number of Employed and Payroll Employment Moving in
Opposite Directions, November-December, 2003 (Seasonally Adjusted)..11

Employment Statistics: Differences and
Similarities in Job-based and Person-based
Employment and Unemployment Estimates
Employment statistics are key indicators of the performance of the economy,
measuring how many jobs/positions exist in the economy as well as how individuals
fare in the labor market. These estimates receive substantial Congressional attention
and are used in policy decisions in many areas including economic stimulus
packages, extending of Unemployment Compensation benefits, and funding of
employment and training programs.
Typically, these statistics are reported in one of two basic forms: payroll
employment statistics and individual (un)employment statistics. Payroll statistics
track the number of filled jobs in the economy and may be used to infer job creation
and loss. (Un)employment statistics measure the quantities and ratios of people in
the labor market who are actively working or seeking jobs. These statistics are used
to estimate the unemployment rate as well. A third type of unemployment statistic,
derived from the Unemployment Compensation system — the Insured
Unemployment Rate — is less frequently used as it excludes many important groups
of workers.
This report explains the data sources used to calculate payroll employment and
individual employment and unemployment estimates. The report also examines
several somewhat counter-intuitive situations, explaining how there may be aggregate
job loss with decreased unemployment rates and how increased payroll numbers do
not necessarily imply decreased numbers of unemployed persons.
Creation of Payroll Employment and
(Un)employment Statistics
Payroll employment statistics and (un)employment statistics receive increased¹
attention during and immediately after recessions occur. These statistics are useful
indicators of how the overall economy and how individuals within the economy are
faring. For example, increasing unemployment rates imply that some of the labor
input (human capital) available to the economy is not used, suggesting production

¹Calculations of the timing of the business cycle (and subsequent designations of recessions
or recoveries) place substantial weight on output measures with only some weight on the
payroll employment statistics. See [http://nber.org/cycles/recessions.html#faq] for more
information.

and/or efficiency losses.² Likewise, measures of payroll employment and the
unemployment numbers are important in quantifying the substantial costs
experienced by workers and their families from the loss of a job as well as associated
losses experienced by society. These worker and family costs include loss of income,
the scarring effect from prolonged periods of joblessness,³ and potentially significant
psychological and emotional costs. Societal costs include increased payments of
unemployment compensation benefits and other pressures on the social welfare
system, and potential increases in crime, drug use, and incarceration.
Data Sources
The Unemployment Compensation (UC) system, measures the number of
unemployed workers covered by the UC system.⁴ However, UC records, which many
people think are the source of total unemployment data, relate only to persons who
have applied for UC benefits. Since it is both impractical to actually count every
unemployed person each month and equally impractical to measure the number of
jobs in each firm, the federal government conducts two monthly sample surveys: the
Current Population Survey (CPS) and the Current Employment Statistics survey
(CES). Both of these data sources are survey based and therefore, the statistics
generated by these surveys are estimates from samples rather than actual counts.
CPS.⁵ The CPS is a monthly survey of households conducted by the Bureau
of the Census for the Bureau of Labor Statistics and designed to represent the civilian
noninstitutional population of the United States. The CPS (approximately 50,000
households) includes certain farm and domestic workers, the self-employed, persons
working 15 hours or more a week as unpaid workers in a family-operated business,
employees of certain nonprofit organizations, and railroad workers. The CPS also
counts employees uncompensated because of temporary absence or strikes, but
excludes workers under 16 years old. It provides a comprehensive body of data on
persons in the labor force, their employment and unemployment, and persons not in
the labor force.

²Okun’s Law is often used to estimate the size of economic loss from unemployment. It
estimates that for every one percentage point increase in unemployment, gross domestic
product drops by 2%. See CRS Report RS21139, Unemployment and Economic Growth,
by Brian Cashell.
³Employers may regard long spells of unemployment as an indicator about workers’
potential productivity, causing a scarring effect. Workers may experience decreased
chances of finding new jobs and ascending their career ladder. Maura Sheehan and Mike
Tomlinson, “Unemployment Duration in an Unemployment Blackspot,” Labour 12 (winter

1998), pp. 643-73.

⁴Covered employment includes all positions where the employer pays federal and state
unemployment taxes on the worker’s wages. While almost all jobs are covered by the UC
program, not all workers qualify for UC benfits. Those who file for UC benefits are
different from the general unemployed population. See Stephen Wandner and Andrew
Stettner, “Why are Many Jobless Workers not Applying for Benefits?” Monthly Labor
Review, June 2000, pp. 21-32. Most of those who do not file do so because they think they
are not eligible or because they are optimistic about finding employment.
⁵See [http://www.bls.census.gov/cps/cpsmain.htm].

CES.⁶ The CES is a monthly employer survey collected from payroll records
by the Bureau of Labor Statistics with the cooperation of state agencies. The sample
has approximately 160,000 businesses and government agencies which have about
400,000 separate sites. The CES collects the number of jobs as reported by
establishments. The survey provides information on the employment, hours, and
earnings of jobs that exist on nonfarm payrolls.
In the most basic terms, statistics that count the number of jobs rely upon the
CES, and statistics that count the number of persons employed (or unemployed) rely
upon the CPS. That is, statistics reporting rates of employment, unemployment, or
that report demographic characteristics of workers will generally be from the CPS.
The number of jobs (rather than the number of workers) in the economy and
information about these jobs by industry will typically be calculated from the CES.
The CPS provides measures of all types of work done by the non-
institutionalized civilian population, including part-time work, self-employment, and
some farm employment. In comparison, the CES does not differentiate between full-
or part-time employment, does not measure self-employment, and excludes all
agricultural jobs.
Calculation of Payroll Employment and Employment
and Unemployment Statistics
The rule of thumb that the CES measures the number of jobs and the CPS
measures both the employment of workers and estimates the unemployment rate, is
complicated by detailed analyses and demographic breakdowns. In particular, for
smaller geographic areas or for monthly rather than annual information, these two
surveys are used in conjunction with each other and may be supplemented with
information from other sources including administrative information from state UC
systems and the decennial census.
National, State, and Local Payroll Employment Statistics. These
statistics are from the CES payroll survey which, as noted earlier, includes
employment numbers, employment by industry, hours, and earnings. Information is
also available for smaller geographic regions (states and certain metropolitan areas).
These survey data are used to calculate the net number of jobs created or lost in a
period. The jobs tallied from this information are not linked to individuals but rather
to firms and industries. Generally, these data are released in the BLS publication The
Employment Situation in their preliminary form on the first Friday of the following
month.
National Employment and Unemployment Statistics and
Unemployment Rate. National estimates of the number of employed and
unemployed as well as the unemployment rate are calculated using data collected in
the CPS. The CPS is a nationally representative household survey and includes
employment and unemployment of the non-institutionalized civilian population by
age, sex, race, and ethnicity. The unemployment statistics are defined by individuals

⁶See [http://stats.bls.gov/ces/home.htm].

who have responded to survey questions that indicate they have: (1) searched for
work in the previous four weeks and (2) were unable to find a position. Generally,
these data are released in the BLS publication The Employment Situation on the first
Friday of the following month.
State and Local Employment and Unemployment (Local Area
Unemployment Statistics). While survey information collected for the national
unemployment rate (CPS) includes information on geography, these data may need
to be supplemented with up to three other information sources (the CES, the census,
and state UC information) in order to obtain statistical significance. State and local
employment and unemployment estimates are calculated from a combination of these
different sources depending on the size of the population of the particular area.
!Annual average data for all states, the District of Columbia, New York City
(NYC), and the Los Angeles-Long Beach (LA) metropolitan areas are
estimated directly from the CPS.
!Monthly estimates for these areas are produced using estimating equations
based on regression techniques. These models combine current and historical
data from the CPS, the CES, and state UC data. Generally, these data are
released on the third Friday of the following month.
!Estimates (monthly and annual) for smaller labor market areas (below the
state level, including metropolitan statistical areas other than NYC or LA,
counties, cities of at least 25,000, and cities and towns in New England
regardless of population) are estimated by a procedure that uses techniques
based on inputs from the decennial census, annual population estimates, and
current state UC data. Generally, these data are released on the fourth
Wednesday of the next month.
The Total Unemployment Rate (TUR) and Insured
Unemployment Rate (IUR)
What is the Total Unemployment Rate (TUR)? The Total Unemployment
Rate (TUR) is simply another name for the unemployment rate derived from the CPS
data.
What Does the Insured Unemployment Rate (IUR) Measure?
Statistics on insured unemployment in the United States are collected as a byproduct
of the unemployment compensation program. Workers who lose their jobs and are
covered by the programs typically file claims which serve as notice that they are
beginning a period of unemployment. Those who qualify for benefits are counted in
the insured unemployment statistic, the Insured Unemployment Rate (IUR).
What Is the Difference Between the IUR and the TUR? The IUR is
substantially different than the TUR because it excludes several important groups:
self-employed workers, unpaid family workers, workers in certain not-for-profit
organizations, and several other, primarily seasonal, worker categories. In addition
to those unemployed workers whose last jobs were in the excluded kinds of
employment, the insured unemployed exclude the following: those who have
exhausted their UC benefits; new entrants or reentrants to the labor force;

disqualified workers whose unemployment is considered to have resulted from their
own actions rather than from economic conditions, and eligible unemployed persons
who do not file for benefits. Because of these and other limitations, statistics on
insured unemployment, cannot be used as a count of total unemployment in the
United States. In 1988, for example, when there were virtually no extended
unemployment benefits paid to persons who had otherwise exhausted their benefits,
the number receiving UC benefits represented only 31% of the total unemployed. In

1992, when extended UC benefits were in effect, this proportion was 51%.

Definitions of Employment
To better understand the various definitions used in calculating the payroll
employment and (un)employment statistics, it is important to clearly define various
employment concepts as they will have direct implications for the usefulness of the
statistics. Table 1 below lists the manner in which the CPS and CES define various
employment concepts. The first column lists the generally accepted title for the
employment concept for the individual-based estimates, the third column lists the
titles for the job-based estimates. The second column lists the definition according
to the CPS and the fourth column lists the definition according to the CES.
In the Labor Force and Out of the Labor Force. The CPS considers all
non-institutionalized, civilian individuals (16 and older) who either currently have
a job or who have looked for work in the previous four weeks to be in the labor
force. All others are considered to be out of the labor force.
The CES does not measure employment from the perspective of the individual
but rather measures paid employment in certain industries. All positions counted in
the payroll survey may represent persons who according to the CPS are in the labor
force. However, not all persons in the labor force (i.e., the unemployed or the self-
employed) will be represented by the payroll statistics.
Employed. In the CPS, an individual who is paid for working at least 1 hour
or who worked 15 hours without pay in a family-owned business during the previous
week is considered to be employed. An individual is also counted as employed even
if he or she is on vacation, not at work due to temporary illness, on temporary unpaid
leave, or involved in a strike. Unlike the CES, the CPS counts individuals only one
time, even if they hold more than one job.
Payroll Employment. In comparison, the CES tallies each job that exists
within each sampled firm toward estimating aggregate payroll employment statistics
if the position paid a worker for working any portion of the reference pay period (the
pay period that includes the 12th of the month). A position is omitted from this tally
if the person did not receive pay for the entire reference pay period. Employees
working at more than one job — therefore appearing on more than one payroll —
would be counted separately for each appearance. All positions counted by payroll
employment have persons who may qualify as being employed according to the CPS
definition. However, not all employed persons as defined in the CPS (e.g., the self-
employed) will have their positions counted in the payroll employment statistics.

Unemployed. The CPS measures unemployment by defining those workers
to be unemployed if they do not have a job but have searched for work in the previous
four weeks and are available to work. The unemployment rate is the ratio of
unemployed persons to the total number of persons in the labor force.
Job Loss and Job Gain. The CES does not measure the number of
unemployed persons. Instead, the CES data from current and previous periods may
be used to calculate the net job loss (gain) from one period to the next. An increase
in the unemployment rate generally would imply an expected decrease in the payroll
employment numbers; however, this is not always true. A more detailed explanation
of how this occurs appears later in this paper.
Involuntarily Part-time.⁷ In the CPS, involuntary part-time workers are
those who report working fewer hours than they wish either due to fewer hours
allotted by the employer or because they were unable to find full-time work. Since
these workers have reported at least one hour of work in the survey week, they are
considered to be in the labor force and to be employed.
In the CES each paid position, regardless of the number of hours, is counted as
payroll employment. In effect, the CES does not measure jobs that have persons who
are working less hours than they wish (involuntarily part-time) any differently than
other jobs within the survey sample.
Discouraged Workers.⁸ In the CPS, these are people currently without a job
but who have expressed a desire to work and have searched for work in the past year;
however, they have made no effort to find a job in the past four weeks due to either
a perceived lack of jobs or a perceived inability to get a job due to some personal
factor such as lack of skills or education. These persons are considered to be out of
the labor force and are not included in calculations of unemployment statistics. The
U.S. Bureau of Labor Statistics does, however, calculate alternatives to the official
unemployment rate which do include discouraged workers among others in their
measures. As these workers do not hold a paid position, the CES data do not include
them.

⁷See CRS Report 98-695, Part-time Job Growth and the Labor Effects of Policy Responses,
by Linda Levine.
⁸ See CRS Report RL322292, Offshoring (a.k.a. Offshore Outsourcing) and Job Insecurity
Among U.S. Workers, by Linda Levine. There is often some misunderstanding surrounding
the difference between a discouraged and a dislocated worker. A dislocated worker has
several years’ job tenure and loses his or her job through plant closings or through
permanent layoff.

Table 1. CPS and CES Employment
and Unemployment Definitions
^{Current E}^m^ploym^ent
^T^erm^{Current Population Survey}⁽^h^o^u^seho^ld/persoⁿ^lev^e^l)^T^erm^St^a^t^istⁱ^{cs (}^e^st^a^b^lishm^ent
^lev^e^{l j}^o^bs)
^L^a^b^o^r^f^o^r^c^e^A^llⁿ^oⁿ^-ⁱⁿ^s^titu^tioⁿ^a^lized^NA^NA
^civ^ilian^s^{, 1}⁶^y^ear^{s an}^d^o^l^d^e^r^,
^w^h^o^{are w}^o^rkⁱⁿ^g^OR^are
^s^earchⁱⁿ^g^f^o^{r w}^o^rk^.
^O^u^t^of^l^a^bor^Uⁿ^{der 16}^y^e^ars^ol^d^OR^NA^NA
^fo^r^c^eⁱⁿ^stitu^tioⁿ^a^lized^OR^no^t
^s^earchⁱⁿ^g^f^o^{r w}^o^rkⁱⁿ^prev^iou^s
^f^o^u^r^w^eek^s
^Em^pl^oy^ed^A^t^l^eas^t¹⁶^y^e^a^r^s^ol^{d, w}^h^o^Pa^y^r^ol^l^A^ll^j^o^b^sⁱⁿ^estab^lish^m^en^ts
^wo^r^k^e^d^{at l}^e^as^t¹^ho^ur^fo^r^p^a^y^em^pl^oy^m^eⁿ^t^th^{at p}^a^id^{a w}^a^g^eⁱⁿ^th^e
^OR^{15 h}^o^u^r^s^wⁱ^t^h^o^u^t^payⁱⁿ^a^ref^e^ren^{ce peri}^{od ex}^cl^udiⁿ^g^:
^fa^m^ily^-o^wⁿ^e^d^b^u^sine^ss.^T^h^e^y^ag^rⁱ^cu^ltu^r^a^{l, self}^-^e^m^p^lo^y^e^d^,
^m^a^y^b^e^oⁿ^va^c^a^tⁱ^o^{n o}^r^no^t^a^t^uⁿ^p^{aid f}^a^m^ily^w^o^rk^ers^,^an^d
^w^o^r^k^d^u^{e to}^t^e^m^p^o^r^ar^y^illn^ess.^priv^{ate h}^o^u^s^eh^{old w}^o^rk^ers^.
^Uⁿ^em^pl^oy^ed^A^t^l^eas^t^{16 y}^ears^an^{d ol}^{der w}^h^o^j^ob^l^o^s^s^/^g^aⁱⁿ^T^h^e^c^u^r^r^eⁿ^tⁿ^u^m^{ber of}^j^obs
^areⁿ^o^{t cu}^rren^tly^em^ploy^{ed an}^dⁱⁿ^estab^lⁱ^s^h^m^en^{ts less th}^e
^h^a^v^e^s^earch^{ed f}^o^r^w^o^rkⁱⁿ^th^e^s^u^m^of^j^obs^du^riⁿ^g^t^h^e
^prev^iou^s^f^o^u^r^w^eek^s^.^prevⁱ^o^u^s^peri^{od. (N}^eg^atⁱ^v^e
ⁿ^u^m^{ber i}^m^plⁱ^e^s^em^pl^oy^m^eⁿ^t
^lo^{ss, p}^o^s^itiv^eⁿ^u^m^b^e^r^im^p^lies
^em^pl^oy^m^eⁿ^t^g^r^ow^t^h^.)
^I^nvo^l^unt^a^rⁱ^l^y^Work^ers^w^h^{o report}^w^o^rkⁱⁿ^g^NA^A^{ll j}^obs^a^r^e^couⁿ^t^{ed as}
^p^a^r^t^-^tim^e^fe^w^e^r^ho^ur^{s t}^h^aⁿ^t^h^e^y^wⁱ^sh.^pay^r^ol^l^em^pl^oy^m^eⁿ^t^w^h^et^h^e^r
⁽^S^{till co}ⁿ^s^id^er^e^d^t^o^b^eⁱⁿ^th^e^th^{e j}^o^b^{is f}^u^ll-^o^r^p^a^r^t^-^tim^e.
^l^a^{bor f}^o^{rce an}^{d t}^o^be
^em^pl^oy^ed.)
^Disco^u^rag^e^d^W^o^u^l^{d l}ⁱ^k^{e a j}^{ob bu}^t^h^a^sⁿ^o^t^NA^T^h^es^{e w}^o^r^k^ers^{do n}^o^{t h}^o^ld
^wo^r^k^e^r^s^earch^{ed in}^prev^iou^s^f^o^u^r^pai^d^em^pl^oy^m^eⁿ^t^an^{d do n}^o^t
^w^eek^s^du^e^{to perception}^th^{at n}^o^fⁱ^g^u^r^{e in}^to^p^a^y^r^o^l^l
^j^obs^exⁱ^s^t^an^{d w}^o^u^l^d^not^be^em^pl^oy^m^eⁿ^t^.
^hⁱ^red^b^ecau^{se o}^f^lack^o^f^sk^ills.
^Coⁿ^s^id^er^ed^to^b^e^o^u^{t o}^f^th^e
^l^a^{bor f}^o^rce.
^Source:^C^R^{S table created u}^sⁱⁿ^g^B^L^{S data diction}^a^ries^f^o^{r th}^{e C}^E^{S an}^{d C}^P^S.
^NA^{= No}^{t ap}^p^licab^le
Example: Decreased Employment but Increasing Jobs. Table 2
contains data taken from the Bureau of Labor Statistics’ The Employment Situation
for September 2004 (released October 8, 2004). Notice that the number of employed
persons (139,480,000) was greater than the number of jobs reported by the CES
(131,567,000, the (p) indicates that the number is preliminary and is subject to
revision). Since agricultural workers (estimated seasonally unadjusted to be

2,374,000); self-employed workers (9,616,000); and unpaid family members (90,000)

work in jobs that are not included in the CES estimates but are considered to be

workers by the CPS, it is reasonable to find that the CES number of jobs is less than
the number of employed workers according to the CPS.
From August 2004 to September 2004, the number of persons who were
employed was estimated to have dropped by 19,000; in comparison the net number
of jobs added to the economy, according to the CES, increased by 96,000.
Table 2. Example: The Employment Situation
for September 2004
^{Current Population}^{Current E}^m^ploym^ent
^T^erm^Surv^e^y⁽^h^o^u^seho^ld/persoⁿ^lev^e^l)^T^erm^St^a^t^istⁱ^cs⁽^e^st^a^b^lishm^ent^lev^e^{l j}^o^bs)
⁽^t^ho^usa^nds)⁽^t^ho^usa^nds)
^L^a^{bor f}^o^rce^147,483^N^A^N^A
^O^u^t^of^l^a^{bor f}^o^rce^76,458^N^A^N^A
^Em^pl^oy^ed^139,480^Pay^r^ol^l^131,567(p)
^em^pl^oy^m^eⁿ^t
^Uⁿ^em^pl^oy^ed^8,003^N^A^N^A
^A^ug.^-Se^p^t^.^Cha^nge^-1⁹^A^ug.^-Se^p^t^.^96(p)
^ch^an^g^e
^Uⁿ^em^pl^oy^m^eⁿ^t^rat^e^5.4%^N^A^N^A
^Source:^{CRS tab}^l^{e cr}^eated^f^r^o^m^d^a^ta^p^u^b^lish^e^dⁱⁿ^th^{e B}^u^r^eau^o^f^L^a^b^o^r^{Statistics’}^T^h^{e Em}^pl^oym^e^nt
^S^itu^a^tioⁿ^f^o^{r S}^e^pt^{. 2004.}
^NA^{= No}^{t ap}^p^licab^le
⁽^p⁾⁼^P^r^elimⁱⁿ^ar^y
Total employment as measured by the CES and the CPS
Figure 1 tracks the monthly employment statistics as reported by the CPS and
CES. In general the CPS accounts for more persons working than the CES calculates
for the number of jobs in the economy. This is in part attributable to inclusion in the
CPS of self-employed workers and a few other types of workers whose jobs are not
measured by the CES.
While the overall pattern between the CPS and CES employment estimates has
been fairly consistent, the difference between the two surveys narrowed in the late
1990s but has been increasing in the 2000s. In particular, while the CPS consistently
has estimated job growth during this period, this trend is not reflected with the same⁹

consistency in the CES.
⁹For information on the path of payroll employment following the 2001 recession see CRS
Report RL32047, The “Jobless Recovery” From the 2001 Recession, by Marc Labonte and
Linda Levine.

CRS-9
Figure 1. Current Employment Survey and Current Population Survey
Estimated Employment in Thousands (Seasonally Adjusted)
January 1970-March 2004
^140,⁰⁰⁰
^E^m^pl^oy^m^ent
^130,⁰⁰⁰⁽^C^u^r^r^ent^P^opul^a^tⁱ^oⁿ^Su^r^v^e^y⁾
^120,⁰⁰⁰
^110,⁰⁰⁰^ousands
ⁱ^{ki/CRS-RL32642}^100,⁰⁰⁰ⁱⁿ^Th
^g/w^eⁿ^t^P^a^y^r^o^l^l^E^m^pl^oy^m^ent
^s^.or⁽^C^ur^r^ent^E^m^pl^oy^m^ent
^le^ak^90,⁰⁰⁰^l^oym^St^a^tⁱ^s^tⁱ^c^s⁾
^p
^://wiki^E^m
^h^ttp^80,⁰⁰⁰
^Re^c^e^s^sⁱ^oⁿ
^70,⁰⁰⁰
^60,⁰⁰⁰^Jan^Jan^Jan^Jan^Jan^Jan^Jan^Jan^Jan^Jan^Jan^Jan^Jan^Jan^Jan^Jan^Jan^Jan
^-⁷⁰^-⁷²^-⁷⁴^-⁷⁶^-⁷⁸^-⁸⁰^-⁸²^-⁸⁴^-⁸⁶^-⁸⁸^-⁹⁰^-⁹²^-⁹⁴^-⁹⁶^-⁹⁸^-⁰⁰^-⁰²^-⁰⁴
^Y^ear
^Source:^Fⁱ^g^u^r^{e prepared by}^C^R^S^bas^e^{d on}^Bu^reau^of^L^a^{bor S}^t^atⁱ^s^tⁱ^c^s^dat^a^f^r^om^t^h^{e C}^u^rren^t^Popu^l^a^tⁱ^oⁿ^S^u^rv^ey^(C^PS^{) an}^{d C}^u^rren^t^Em^pl^oy^m^eⁿ^t^{Statistics (}^C^ES)^.

How Falling Employment and a Growing Payroll
Employment May Coexist
It may be confusing when economic indicators such as payroll employment
statistics, employment numbers, and unemployment rates seem to suggest opposite
trends. Tables 3 and 4 below demonstrate how this outcome occurs because of the
underlying definitions of the unemployment rate and payroll employment that
measure employment in very different ways.
In any given economy, it is possible for both the number of employed persons
and the unemployment rate to increase at the same time within the same household
survey (the CPS). An interesting case to illuminate this point can be found in
studying the growth of Hispanic employment shown in Table 3 below.
Table 3. Hispanic Employment and Unemployment Rate, 2000
and 2003
YearLabor Force(thousands)Employed(thousands)Unemployed(thousands)UnemploymentRate

2000 16,689 15,735 954 5.7%

2003 18,813 17,372 1,441 7.7%

Difference + 2,124 +1,637 +487 2.0%
Percentage 12.7% 10.4% 51.0% 35.1%
Change
^So^urce^{: CRS tab}^l^{e cr}^eated^u^sⁱⁿ^g^B^L^{S statistics calcu}^lated^f^r^o^m^th^{e CP}^S.
Hispanic employment increased by 1,637,000 persons between 2000 and 2003,
but the Hispanic rate of unemployment increased from 5.7% to 7.7%. The
explanation is that at the same time as 1,637,000 new jobs filled by Hispanics were
added to the economy, approximately 2,124,000 Hispanics entered the workforce.
The growth in employment was more than offset by the number of new job seekers
entering or returning to the labor market.
On average, the labor force grows at approximately 1% per year. As a result,
employment must grow at least at that same rate (1%) to prevent an increase in the
unemployment rate.
It is typical for the unemployment rate to continue to rise after a recession’s end,
such as occurred after November 2001. This may be because some persons reenter
the labor force with the belief that job prospects have improved and thus, are less
willing to accept certain jobs. It may also be due to other unemployed workers who
are unable to find new work in a substantially reconfigured labor market where their
skills are no longer in demand.

Table 4. The Number of Employed and Payroll Employment
Moving in Opposite Directions, November-December, 2003
(Seasonally Adjusted)
^La^bo^r^E^m^p^l^oyed^Payrol^l
^Moⁿ^t^h^Unem^ploym^ent^Ra^te^Force^Pers^ons^E^m^p^l^oym^en^t
⁽^t^ho^usa^nds)⁽^t^ho^usa^nds)⁽^t^ho^usa^nds)
^N^o^v^e^m^b^{er 2003}^5.9%^147,187^138,533^130,123
^Decem^{ber 2003}^5.7%^146,878^138,479^130,124
^Dⁱ^f^f^e^ren^c^e^-^0.2%^-³⁰⁹^-⁵⁴⁺¹
^P^e^rcen^tag^e^C^h^an^g^e^-³^.3%^-^0.2%^-⁰^.0%⁰^.0%
^So^urce^{: C}^R^S^{table created u}^sⁱⁿ^g^B^L^{S s}^t^atis^tics^calcu^{lated f}^r^om^th^{e C}^P^{S an}^{d th}^{e C}^E^S.
Just as Table 3 demonstrated how the number of employed and the
unemployment rate can increase at the same time within the same household survey,
Table 4 delineates how the CES may have nominally increasing payroll employment
while the CPS has a substantial decrease in number of employed persons and at the
same time a decreased unemployment rate.
In December 2003, the national unemployment rate was 5.7% — a statistically
significant decrease from 5.9% in November 2003. At the same time, the estimated
number of employed decreased by approximately 54,000 individuals from
138,533,000 to 138,479,000. The unemployment rate was able to decrease despite
the drop in persons reporting employment because approximately 309,000 persons
reported that they left the labor force.¹⁰ Payroll employment, on the other hand,
remained essentially the same, increasing by an estimated 1,000 positions

¹⁰The number of individuals identifying themselves as discouraged workers decreased from
457,000 in November to 433,000 in December. This decline, however does not take into
account any expected seasonal variations in this statistic and should be viewed with caution.