Post-recession trends in nonfarm employment and related economics indicators by LaborStats


									                                       A visual essay:

                      Post-recession trends
                    in nonfarm employment
                 and related economic indicators
                       David Langdon, Rachel Krantz, and Michael Strople

   •    Real GDP

   •    Corporate profits

   •    Unemployment Insurance (UI) claims

   •    The help-wanted index

   •    Manpower Inc.'s employment outlook survey

   •    Consumer confidence (appraisal of current employment conditions)

   •    Federal income tax withholdings

David Langdon is a supervisory economist and Michael Strople is an economist in the Division of Current
Employment Statistics, Bureau of Labor Statistics. Rachel Krantz is an economist in the Division of Labor
Force Statistics, Bureau of Labor Statistics.

                                                                         Monthly Labor Review    September 2004   49
A Visual Essay

       he economy entered a recovery in November 2001 following an 8-month recession, but the labor-
       market recovery began much later. Gross domestic product (GDP) and corporate profits had surged
       before payroll employment reached its August 2003 trough. Employment edged up by about 60,000
a month during the remainder of the year. Gains averaged about 225,000 a month during the first 5
months of 2004 before slowing in the summer. The post-recession disconnect between overall economic
growth and the labor market was unusual. The following analysis reviews this apparent inconsistency,
and also identifies a number of other broad labor-market indicators that paralleled the lagged recovery in
payroll employment. Employment data used in this essay are from the Current Employment Statistics
(CES) survey, a monthly survey of about 160,000 businesses and government agencies, which represent
approximately 400,000 individual worksites. For more information on the CES, see BLS Handbook of
Methods, chapter 2, on the Internet at

                                                                                           Real GDP
During and following the recessions of the      1.     Percent change in real gross domestic product and quarterly change in
early 1980s, the employment trend gener-               total nonfarm employment, seasonally adjusted
ally coincided with the trend in real GDP. In
contrast, sustained employment growth            Quarterly
lagged GDP growth by four quarters after the     employment change                                                             GDP annualized
                                                 (in thousands)                                                                 percent change
1990–91 recession and by eight quarters af-
ter the 2001 recession. Overall, the down-       1,500                                                                                     15
turn in employment associated with the 2001                                   Employment
recession was more prolonged than those as-
sociated with the recessions of the 1980s and
                                                 1,000                                                                        GDP          10
1990–91. (The National Bureau of Eco-
nomic Research, the official arbiter of re-
cessions, has designated these periods as re-
cessions: January-July 1980, July 1981-No-            500                                                                                  5
vember 1982, July 1990-March 1991, and
March-November 2001.)
                                                        0                                                                                  0
The 1990–91 and 2001 recessions were
milder in terms of contracting GDP than those
of the early 1980s. During the 1981–82 re-
                                                     –-500                                                                                –-5
cession, employment contracted by 3 per-
cent, compared with a mild 1 percent during
the other three recessions since 1979.
                                                –-1,000                                                                                   –
                                                             1980   82   84   86    88     90   92      94   96   98   2000     02   04

                                                        SOURCES: Bureau of Economic Analysis and BLS.

50    Monthly Labor Review        September 2004
                                                                                     Corporate profits
Historically, private employment and real        2.     After-tax profits in constant (2000) dollars and quarterly averages of
corporate profits behave similarly at the               total private employment, seasonally adjusted
end of a recession. Both reach a trough as
                                                  Employment                                                                          Corporate profits
the recession ends and then recover               (in thousands)                                                                          (in billions)
quickly. At the end of the 1990–91 reces-         120,000                                                                                      $1,200
sion, both profits and employment contin-
ued to move in parallel, but were lagged in       110,000
their recovery. The parallels broke down                                                                                                        1,000
with the 2001 recession.                          100,000

Leading into the 2001 recession, growth of                                                                                                      800
corporate profits stalled and then exhibited          80,000
a relatively sharp and deep decline. Em-
                                                      70,000                    Employment                                                      600
ployment declined as well during the re-
cession, but not as drastically as corporate
profits. Since reaching a trough in the third
quarter of 2001, corporate profits have in-                                                                                                     400
                                                      50,000                                                               Profits
creased by 56 percent. Employment, on the
other hand, remained essentially un-                  40,000
changed until late 2003. Rising productiv-
ity explains, at least in part, the divergence        30,000
in profits and employment.                            20,000                                                                                    0
                                                               1948 52   56   60    64    68   72    76   80   84     88   92    96 2000 04

                                                     NOTE: Corporate profits were deflated using the Implicit price deflator for GDP, 2000 =
                                                  100. Shaded areas denote recessions.
                                                     SOURCES: Bureau of Economic Analysis and BLS.

Corporate profits declined substantially         3.     Total domestic pre-tax profits and total excluding computer and elec-
from the first to fourth quarter of 2001 and            tronic products, information, and transportation and warehousing
then recovered quickly. A large part of this
swing is tied to the financial situation of                                                                                                  Billions
three industry groups: information technol-
ogy (IT)-related manufacturing, information      $750                                                                                           $750
industries, and transportation. (IT-related
manufacturers produce information tech-
nology products. They are classified in           700               Domestic corporate profits,                                                   700
                                                                    excluding selected industries
computers and electronic products manu-
facturing [NAICS 334].) IT-related manufac-
                                                  650                                                                                             650
turing and information industries felt the
impact of the bursting technology bubble,
whereas transportation was affected in part       600                                                                                             600
by the aftermath of the September 11th ter-
rorist attacks. Corporate profits excluding
those industries were much less volatile and      550                                                                                             550
much more in keeping with prior reces-                                                               Domestic corporate profits with in-
sions.                                                                                               ventory valuation adjustment
                                                  500                                                                                             500

                                                  450                                                                                             450
                                                           I      II    III   IV      I     II      III   IV      I      II     III   IV     I
                                                                    2001                   2002                        2003                2004

                                                       SOURCES: Bureau of Economic Analysis and authors' calculations.

                                                                                                Monthly Labor Review            September 2004          51
A Visual Essay

                                                                                      Corporate profits
The difficult financial situation in com-          4.      Quarterly employment changes in total private industries and total
puter and electronic products, information,                private excluding selected industries, seasonally adjusted
and transportation and warehousing re-              Thousands                                                                                            Thousands
sulted in cost-cutting initiatives that in-
                                                    1,000                                                                                                        1,000
cluded numerous layoffs. As a percentage
of private employment, these three indus-               800                                                                                                      800
tries accounted for approximately 9 percent
of employment in the first quarter of 2001;             600                                                                                                      600
however, they accounted for approximately
                                                        400                                                                                                      400
41 percent of declines in private industry
employment through the third quarter of                 200                                                                                                      200
                                                          0                                                                                                      0

                                                    –-200                                                                                                       –-200

                                                    –-400                                                                                                       –-400

                                                    –-600                                                                   Total private
                                                                                                                            Total private                       –-600
                                                                                                                            Total private excluding
                                                                                                                            Total private excluding
                                                    –-800                                                                   selected industries1                –-800

                                                    -1000                                                                                                       – 1,000

                                                    -1200                                                                                                       – 1,200
                                                                 I    II   III   IV   I     II     III        IV        I       II     III    IV     I   II
                                                                       2001               2002                                2003                  2004
                                                         Excludes computers and electronic products, information, and transportation and
                                                       SOURCES: BLS and authors' calculations.

The third quarter of 1997 showed a distinct         5.        After-tax profits and labor costs per unit of real gross value added of
increase in labor costs per unit of real gross                nonfinancial domestic corporate business, seasonally adjusted
value added, which coincided with falling
profits; this trend continued until the third      Labor
quarter of 2001. (These measures are pro-          costs                                                                                                      Profits
duced by the Bureau of Economic Analysis.)         $0.70
                                                     0.7                                                                                                        0.2
Over this timeframe, labor costs increased by
7 cents, while profits declined by 5 cents per
unit of value added. The rising labor costs co-                                        Labor costs
incided with strong private employment               .60
                                                    0.6                                                                                                        .10
growth, an average of 447,000 jobs a quarter.

This apparent tradeoff reversed itself in the                                                                               Profits
fourth quarter of 2001. Unit profits began to        .50
                                                    0.5                                                                                                        00
increase, and unit labor costs decreased.
From the third quarter of 2001 to the third
quarter of 2003, profits per unit increased by
3 cents, while unit labor costs decreased by         .40
                                                    0.4                                                                                                        –.10
2 cents. Private employment averaged de-
clines of 288,000 per quarter over the same
timeframe. This drop in unit labor costs is
unusual, but is a way for firms to return to        .30
                                                    0.3                                                                                                        –.20
profit when their pricing power is weak.                       1980   82   84    86    88     90         92        94        96       98     2000   02   04
Price per unit of value added was flat be-
tween 2001 and 2003, so in order to increase          NOTE: Profits are after taxes, inventory valuation adjustment, and capital consumption adjust-
profits, businesses cut costs, including the re-   ment. The series are deflated such that the price of one unit of gross value added in 2000 equals $1.
duction of labor costs through layoffs.                  SOURCE: Bureau of Economic Analysis.

52     Monthly Labor Review         September 2004
                                                                     Unemployment Insurance (UI) claims
Initial UI claims and the over-the-month        6.     Monthly average of initial UI claims and monthly changes in total non-
change of total nonfarm employment mea-                farm employment, seasonally adjusted
sure the flow of workers. The former mea-
                                                 Over-the-month                                                                                     Number of
sures the flow of laid-off workers claiming      employment changes                                                                             claims, inverse
unemployment insurance compensation for          (in thousands)                                                                                           scale
the first week following a layoff. The latter    750                                                                                                 100,000
measures the net flow of jobs onto payrolls.
Not surprisingly, these flows are negatively                          Total nonfarm over-the-month change
correlated.                                      500                                                                                                  200,000

After the 2001 recession, initial UI claims
hovered around 400,000 monthly, a some-          250                                                                                                  300,000
what lower level than during prior reces-
sions. Claims did not fall and remain be-
low 400,000 until the fourth quarter of                0                                                                                              400,000
2003, coinciding with the return to growth
in total nonfarm employment.
                                                –-250                                                                                                 500,000

                                                –-500                                                                                                 600,000
                                                                    Initial claims

                                                –-750                                                                                                 700,000
                                                     1980      82     84     86        88        90        92     94   96      98   2000   02    04

                                                         NOTE: The August and September 1983 over-the-month changes, –300,000 and +1.1 million,
                                                     reflect a strike of 640,000 workers and their subsequent return.
                                                         SOURCES: DOL Employment and Training Administration and BLS.

The labor market’s post-recovery stagna-        7.     Monthly average of continuing UI claims and total nonfarm employment,
tion is particularly evident in the continu-           seasonally adjusted
ing claims series. Like total nonfarm em-
ployment, continuing claims have tended          Number of                                                                                      Employment (in
                                                 claims                                                                                             thousands)
to reach a turning point at the end of a re-
cession, with employment then rising and        5,000,000                                                                                             140,000
claims declining. The pattern changed                                                                                        Employment
somewhat in the 1990–91 recession, as the
drop in claims stalled between mid-1991
and mid-1992.                                   4,000,000                                                                                             120,000

At the end of the 2001 recession, continu-
ing claims reached 3.67 million, a plateau
that extended through mid-2003. They            3,000,000                                                                                             100,000
did not decline steadily until September
2003, following the August trough in to-
tal nonfarm employment.

                                                2,000,000                                                                   Continuing                80,000

                                                1,000,000                                                                                             60,000
                                                             1967 70       73     76        79        82     85   88   91      94   97 2000 03

                                                     NOTE: Shaded areas denote recessions.
                                                     SOURCES: DOL Employment and Training Administration and BLS.

                                                                                                           Monthly Labor Review          September 2004         53
A Visual Essay

                                                                                      The help-wanted index
The help-wanted index measures monthly           8.     The help-wanted index and monthly changes in total nonfarm employ-
help-wanted advertising volume in 51 ma-                ment, seasonally adjusted
jor newspapers across the United States.
                                                  Over-the-month                                                                                     Help-
During the recessions of 1980, 1981–82,           employment changes                                                                                wanted
and 1990–91, a downturn in help-wanted            (in thousands)                                                                                     index
advertising corresponded with job losses;          600                                                                                                150
conversely, an upturn in help-wanted ad-
vertising coincided with job gains. News-                                                    Employment
paper job advertising began its most recent       400                                                                                                 125
decline in 2000 and remained weak through
2003. Nonfarm payroll employment fell
from early 2001 until late 2003.                  200                                                                                                 100

Interestingly, the magnitude of decline in
help-wanted advertising around the 2001                0                                                                                              75
recession was more severe than the declines
of the early 1980s and 1990s. In recent
                                                 –-200                                                   Help-wanted index                            50
years, Internet-based job postings have be-
come a popular method for employers to
advertise job openings; some employers           –-400                                                                                                25
may substitute web-based postings for
newspaper advertising. The shape of the
help-wanted index indicates that newspa-         –-600                                                                                                0
per job advertising notched up in early                    1980    82      84    86     88    90        92        94    96    98   2000   02   04
2004, after showing little change in 2003.
                                                     NOTE: The August and September 1983 over-the-month changes, –300,000 and +1.1 million,
This small uptick coincided with strong          reflect a strike of 640,000 workers and their subsequent return.
gains in total nonfarm employment.
                                                     SOURCES: The Conference Board and BLS.

Manpower Inc.'s net employment outlook                            Manpower Inc. employment outlook survey
measures the difference between the percent
                                                  9.       Net employment outlook and quarterly changes in total nonfarm
of employers expecting to add jobs and
those expecting to shed them. Curiously, in
                                                           employment, seasonally adjusted
only one quarter since 1977 were there more       Quarterly employment                                                             Net employment outlook
                                                  change (in thousands)                                                                         (percent)
firms planning to decrease their staffing lev-
                                                   2,000                                                                                              30
els than firms planning to increase them.
                                                                        Net employment outlook
Trends in the net employment outlook cor-          1,500
relate highly with quarterly changes in to-
tal nonfarm employment. A one-quarter lag
in the employment outlook shows an even            1,000                                                                                              15
higher correlation. Not surprisingly, em-
ployers may develop their outlook for the
upcoming quarter by looking at the one that            500
has just ended.

Like total nonfarm employment, Man-                        0                                                                                          0
power Inc.'s employment outlook behaved
unusually after the 2001 recession. Al-
though the outlook brightened somewhat                –-500
                                                                                   Quarterly change in total nonfarm employment
in 2002, corresponding with relatively
smaller declines in nonfarm employment,
it dimmed again in 2003. At the end of the       –-1,000                                                                                              –-15
                                                               1977 79     81    83    85    87    89        91    93    95   97   99 2001 03
year and into early 2004, the outlook
showed greater improvement, and total
nonfarm employment began to recover.                  SOURCE: Manpower Inc. and BLS.

54    Monthly Labor Review         September 2004
                                                Consumer confidence (appraisal of current employment conditions)
As part of the Conference Board’s monthly       10.       "Jobs plentiful" and monthly changes in total nonfarm employment,
survey of Consumer confidence, respon-                    seasonally adjusted
dents evaluate current employment condi-
tions, judging jobs as either “plentiful,”       Over-the-month employment
“not so plentiful,” or “hard to get.” Focus-     change (in thousands)                                                                       Percent
ing just on those who see “plenty” of jobs,       750                                                                                             60
it is clear that the most recent economic re-
                                                                            Total nonfarm over-the-month change
covery was not accompanied by an imme-
diate improvement in the labor market. The        500                                                                                             50
proportion of respondents finding jobs
plentiful recovered somewhat at the end of
the 2001 recession, only to decline again         250                                                                                             40
during 2002 and through most of 2003.
During the final months of 2003, it began
to rise, coinciding with the first months of          0                                                                                           30
steady, though slow, growth in total non-
farm employment. It is worth remember-
ing that BLS data may condition the public’s    –-250                                                                                             20
perception of labor market conditions, es-
pecially in the case of people who are not      –-500                       "Jobs plentiful"                                                      10
actively looking, or know anyone actively
looking, for work.
                                                –-750                                                                                             0
                                                          1978   80   82   84   86     88      90   92   94     96     98   2000   02        04
                                                    NOTE: The August and September 1983 over-the-month changes, –300,000 and +1.1 million,
                                                reflect a strike of 640,000 workers and their subsequent return.
                                                   SOURCE: Conference Board and BLS.

A broader measure of consumer confidence        11.       Net appraisal of current employment conditions and monthly changes
in the labor market would take into account               in total nonfarm employment, seasonally adjusted
all people, not just those who see "plenty"      Over-the-month employment                                                                   Diffusion
of jobs. Such a measure—a pseudo-diffu-          change (in thousands)                                                                           index
sion index—can be calculated by weight-           750                                                                                             80
ing the proportion of people who respond                                    Total nonfarm over-the-month change
“plentiful” as 100, “not so plentiful” as 50,
and “hard to get” as 0, and summing the          500                                                                                              70
results. This index highlights the growing
lag between the end of recessions and the
related labor market recovery. In 1982,          250                                                                                              60
consumers noted a 1-month lag. After the
1990–91 recession came an 11-month lag,
followed by a somewhat fitful recovery;               0                                                                                           50
consumers did not identify sustained labor
market improvement until mid-1992. The
                                                –-250                                                                                             40
lag following the 2001 recession was 22
months, as the diffusion index bottomed out
                                                                                                              "Diffusion index" of respon-
in September 2003. This pattern parallels                                                                     dents' appraisal of current
                                                –-500                                                                                             30
that seen in CES data; employment hit a low                                                                   employment conditions
in August 2003.
                                                –                                                                                                 20
                                                          1978   80   82   84   86     88      90   92   94     96     98 2000 02            04

                                                    NOTE: The August and September 1983 over-the-month changes, –300,000 and +1.1 million,
                                                reflect a strike of 640,000 workers and their subsequent return.
                                                    SOURCE: Conference Board, BLS, and authors' calculations.

                                                                                               Monthly Labor Review         September 2004             55
A Visual Essay

                                                                            Income tax withholdings
Federal income tax withholdings and non-       12.    Federal income tax withholdings and quarterly total nonfarm employ-
farm employment had similar trends during             ment, seasonally adjusted
the last two recessions. Both employment                                                                                               Tax with-
                                               Employment (in                                                                        holdings (in
and withholdings peaked as the recessions      thousands)                                                                               billions)
began and reached low points after the re-      140,000                                                                                     $900
cessions had ended. After the 1990–91 re-
cession, there were no substantive job gains
until the second quarter of 1992, coincident
with a return to growth in income tax          130,000                                                                                       800
withholdings. There are not yet sufficient
data to compare turning points in withhold-                                 Nonfarm
ings and employment for the period after the                                employment
                                               120,000                                                                                       700
2001 recession, although both data series
declined through the third quarter of 2003
and have since regained some of their re-
                                               110,000                                                                                       600
cession-related losses. Income tax cuts en-
acted in 2001 and 2003 are also a factor in                                                           Income tax
the recent decline in withholdings.                                                                   withholdings
                                               100,000                                                                                       500

                                                 90,000                                                                                      400
                                                          1988      90       92        94        96       98      2000       02        04
                                                   NOTE: Quarterly withholdings data are not available for years prior to 1988. Federal personal
                                               income tax withholdings were deflated using BEA's price index for Federal consumption expendi-
                                               tures and gross investment (2000 = 100). Shaded areas denote recessions.
                                                   SOURCES: Bureau of Economic Analysis, BLS, and authors' calculations.

56    Monthly Labor Review      September 2004

To top