View original document

The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.

September 15, 1990

eCONOMIC
COMMeNTORY
Federal Reserve Bank of Cleveland

Can State Employment Declines
Foretell National Business Cycles?
by Randall W. Eberts

Re media accounts have suggested
recent
that the employment performance of states
may offer early signs of a downturn in the
national economy. Some have even hinted
that as few as six states experiencing
employment declines simultaneously is
enough to push the entire U.S. economy
into a recession. For analysts seeking
another leading indicator of national business cycles, the possible existence of
bellwether states is intriguing.
This interest has been heightened by
the perception that, during the 1980s,
the economy was characterized by rolling recessions in which some states or
regions appeared to run counter to the
national trend. If national business
cycles affected states in the same way,
all states would enter and leave recessions at the same time—resembling a
string of boats moored in a harbor, bobbing up and down in unison as a wave
passes below.
Instead, it appears that states enter and
leave national recessions at different
times, and that some states can even
remain untouched by national downturns. For instance, while the U.S.
economy fell into a recession between
1980:IQ and 1981:IIIQ, the energyproducing states of Texas, Louisiana,
and Oklahoma showed no significant
downturn. The same held true during
the subsequent—and much deeper—
recession between 1981:IIIQ and
1982:1 VQ. Yet, as the rest of the nation
began to climb out of the second recesISSN 0428-1276

sion, the oil states slipped into their
own. After enjoying a short recovery
within a year after the national trough,
these three states again descended into
a period of employment decline, which
lasted well into 1987.
Recent episodes of state employment
loss during periods of national business
expansions involve several New
England states, centered around Massachusetts. After years of seemingly
unprecedented growth, that state's economy abruptly changed course in June
1986, subsequently entering a protracted period of employment decline
that still persists. Rhode Island and New
Hampshire immediately joined Massachusetts, and Vermont and Maine were
pulled into the regional recesssion soon
thereafter.
Clearly, the economic performance of
states and regions differs throughout a
national business cycle. It is reasonable
to ask, then, if some pattern develops
among changes in state employment
levels that could offer clues concerning
turning points in the national economy.
This Economic Commentary explores
state employment fluctuations since
1950 to ascertain whether any such
regularities have indeed occurred. The
approach taken here is simple: States
are classified each month as having
either employment growth or employment decline, with growth being
measured as the year-over-year percent-

Is it possible to predict national
recessions by following the employment performance of specific states or
regions? The author investigates this
question from a historical perspective
by comparing the employment patterns of states to the health of the
national economy over the post-WWII
period.

age change in total nonagricultural
employment.
• Tracking State Employment
Declines
A monthly plot of the number of states
showing no year-to-year employment
growth closely tracks national downturns, as measured both by declines in
the U.S. employment level and by the
peaks and troughs of a national business cycle, assigned by the National
Bureau of Economic Research (figure
1 ).2 For each of the seven national
recessions since 1950, the trough coincided (within two months or less) with
the maximum number of states that
would eventually experience employment loss during that specific downturn. At least half of the states (Alaska
and Hawaii are not considered here)
participated in each recession, although
not necessarily the same states. Furthermore, according to past cycles, it takes
at least 20 states posting employment
declines concurrently before a national

FIGURE 1 NUMBER OF STATES EXPERIENCING EMPLOYMENT DECLINES

1956

1961

1966

1971

1976

1981

1986

1990

NOTE: Dashed lines indicate the peaks and troughs of national recessions. Last month plotted is May 1990.
SOURCE: U.S. Department of Labor, Bureau of Labor Statistics.

recession occurs. Since 1950, there
have been three separate incidents in
which a significant number of states
(nine to fourteen) have registered
employment downturns simultaneously
without an ensuing national recession.
One barometer of the severity of a business slide is the number of states that
lack employment growth at the trough.
For instance, during the 1969:IVQ1970:IVQ recession—the least severe
since WWII as measured by real GNP
decline—17 states escaped employment downturns. By contrast, every
state experienced declines during the
1981:IIIQ-1982:IVQ recession—the
worst since WWII.
It is also true that most states experience
employment growth during national
upswings. Over the last 40 years, no
more than 14 states have shown a dip in
employment rates during a national economic expansion; in most instances, that
figure has been closer to zero. An excellent example is the extended period of
growth between February 1961 and
December 1969, when generally only
one or two states sustained sinking
employment levels at any given time.
Other decades have had periods in
which more states ran counter to
national trends than occurred during the
1960s. Although the current expansion
has now surpassed the 1961-69 upswing
to become the longest period of peace-

time growth since WWII, two episodes
in which a relatively large number of
states faced employment drop-offs
have blemished the otherwise excellent
record. The first wave of decline was
caused by the effects of falling oil
prices in the oil-producing states and
by low agricultural prices in the Farm
Belt. The second was associated with
the repercussions of defense spending
rollbacks and financial industry downsizing on the New England economy. Such
instances of employment declines during
economic expansions may represent secular rather than cyclical shifts, since some
states have undergone extended periods of
significant economic restructuring.
• Patterns of Entry and Exit
State employment performance has followed a fairly regular pattern during
each national recession. Sometime
around the peak of each national business cycle, one or two states (usually
not the same ones) will slip into a
downturn. Within a few months, these
states are joined by others, so that by
midway between peak and trough, at
least 50 percent of the states that will
eventually undergo employment loss
have already done so. Once a state's
employment level changes direction
from either growth to decline or vice
versa, it usually continues on that path
for several months.
The pattern of states' entry into and exit
from a national recession typically fol-

lows a "first in-last out" scenario—the
first state to enter a period of employment decline is usually the last to leave
it. Consequently, the timing of turning
points from employment growth to sustained employment loss predicts how
severely a state will be affected by a
downturn.
The 1973-75 recession provides a good
example. Delaware was the first state to
enter that downturn, two months after
the national peak. New York followed
one month later. Both states subsequently experienced comparable durations of employment loss: Delaware's
decline lasted 22 months, while New
York's slump stretched to 26 months.
By contrast, South Dakota, which was
the last state to enter the recession (18
months after Delaware), sustained a
spell of employment decline of only
two months.
While state employment changes followed similar patterns during the other
post-1950 recessions, the average spell
of an employment slump differed from
one recession to the next. Aside from
the twin recessions of 1980, which
many states experienced as one continuous decline, the longest average
period of employment loss was 13.5
months, which occurred during the
1953-54 recession. The spells of
employment loss grew successively
shorter during the next three recessions, bottoming out at slightly more

TABLE 1 STATE MANUFACTURING CHARACTERISTICS AND MONTHS OF EMPLOYMENT DECLINE
Slats
1. West Virginia
2. Rhode Island
3. Pennsylvania
4. Michigan
5. Indiana
6. Illinois
7. Ohio
8. New York
9. Massachusetts
10. Missouri
11. Iowa
12. Louisiana
13. Montana
14. Wisconsin
15. Wyoming
16. Maine
17. Oregon
18. Kentucky
19. Vermont
20. Kansas
21. Washington
22. Minnesota
23. Delaware
24. Oklahoma

Months of
employment
decline

Labor
productivity

Share of
manufacturing

Stats

171
143
140
137
127
124
120
119
115
114
113
107
107
107
106
103
98
95
93
92
92
90
89
87

1.49
3.91
2.90
1.43
2.59
2.94
3.02
4.01
5.61
3.04
4.29
.76
-.34
3.67
-2.92
4.42
2.22
1.88
5.70
3.23
.91
5.22
2.12
4.04

14.30
23.54
20.53
24.78
26.03
18.95
23.31
14.43
18.08
19.02
19.49
11.47
7.56
24.95
4.52
19.55
18.00
19.80
18.43
17.30
17.67
19.08
21.14
14.12

25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.

Months of
employment
decline

Connecticut
New Hampshire
North Dakota
Alabama
Idaho
Nebraska
Tennessee
South Dakota
New Jersey
Arkansas
Mississippi
South Carolina
Texas
Maryland
North Carolina
California
Virginia
Colorado
Georgia
New Mexico
Utah
Arizona
Nevada
Florida

Labor
productivity

Share of
manufacturing

85
85
84
82
82
80
74
70
67
65
65
65
63
59
59
56
52
51
49
48
48
32
25
19

4.51
8.95
2.61
3.77
4.94
3.82
4.30
3.46
3.64
3.91
3.49
5.13
2.75
3.62
3.00
4.18
2.47
3.56
3.99
5.51
5.71
3.05
2.90
4.49

21.44
21.64
6.34
24.17
16.50
13.44
24.32
11.55
17.62
25.80
26.42
26.01
14.24
9.73
28.29
17.24
14.95
13.11
19.29
7.51
14.85
12.90
4.37
10.26

NOTE: Statesarerankedby the number of months of employment decline between January 1950 and May 1990. Labor productivity is measured as the average annual
rate of change of manufacturing output per worker between 1979 and 1986. Manufacturing's share of total employment is the 1990 average.
SOURCE: U.S. Department of Labor, Bureau of Labor Statistics.

than seven months. After that, however, the duration began to increase:
The combined average length of
employment loss for the twin 1980
recessions was 25 months.
Although many states participated in
each recession at the trough, no one
state or group of states ushered in every
national downturn, a fact that dismisses
the notion of a bellwether state or
region. The Midwest has been well represented in each of the seven recessions
since 1950, however. Michigan introduced three of them, and all of the midwestern states experienced employment
slumps during each of the national
recessions. The Midwest's close ties to
national cycles do not stem simply from
the fact that the region employs a large
share of the nation's workers. California, Texas, and Florida also have large
employment bases, but they are typically the last states to enter a recession;
in some cases, they have been able to
avoid recessions entirely.

• Why the Difference in
Employment Changes among States?
What, then, explains the variation in
state employment cycles? State employ-

ment changes are a combination of
cyclical and secular trends. Secular
trends are associated with structural
shifts across industries and across
regions. For instance, many midwestern states have lost their share of
national manufacturing employment
and population to states in the south
and southwest as a result of shifts in
regional comparative advantage. Consequently, their employment patterns
could result partly from factors that are
not necessarily associated with business cycles.
Business cycles, on the other hand, are
thought to be caused by more temporary
factors. One commonly held view is that
shocks—such as escalating oil prices,
precipitated in the past by OPEC policies
and more recently by the Iraqi conflict—
trigger national recessions. Other types of
shocks, including inventory excesses or
technology and demand shocks, may also
influence the business cycle, but oil
shocks appear to be the most identifiable.
Higher oil prices and even oil shortages
adversely affect industrial output by raising production costs and, in some instances, by depressing production. The
severity of the economic consequences

of these shocks for individual industries, and consequently for states,
depends on the intensity with which
they use oil and oil-based products.
According to the oil-shock view of business cycles, variation in state employment depends on the sensitivity of state
industries to such disturbances, which in
turn is related to the productivity and
composition of a state's industries.
States with less-efficient industries may
be more susceptible to shocks such as
higher oil prices, for example, because
firms located there cannot absorb
increased energy costs without losing
their competitive advantage to regions
with more-efficient businesses. Productivity can differ across states as a result
of the age of equipment and facilities
and the skill of the labor force, among
other factors. Manufacturing productivity, measured as output per worker,
varies greatly among states. As shown in
table 1, between 1979 and 1986, New
Hampshire posted the highest productivity gains, while Wyoming recorded
the lowest. Analysis shows that states
with higher-than-average labor productivity growth had lower-than-average
periods of employment loss.

Another factor affecting a state's susceptibility to shocks is its composition of
industries. Specific industries follow different cyclical patterns and show various
degrees of sensitivity to shocks. For
instance, durable manufacturing historically has exhibited wide swings in performance, stemming in part from the durable nature of the product. The industry
is also heavily dependent on energy in
the production process, making it vulnerable to higher energy costs.
Therefore, it is not surprising that, during the 1980s, those states with a higherthan-average percentage of workers
employed in durable goods manufacturing sustained a greater-than-average
number of months of employment loss.
Conversely, the number of months of
employment decline was negatively
correlated with the share of workers in
the service industry—a sector that
shows few cyclical tendencies. As
shown in table 1, the share of manufacturing employment ranges from a high
of 28.29 percent for North Carolina to
a low of 4.37 percent for Nevada.
Table 1 also ranks states according to
the number of months of total employment decline since 1950, thereby illustrating the relationship between a
state's concentration in manufacturing
and the duration of employment loss.
West Virginia wins the unenviable firstplace position, with 171 months of

Federal Reserve Bank of Cleveland
Research Department
P.O. Box 6387
Cleveland, OH 44101

Address Correction Requested:
Please send corrected mailing label to
the above address.

Material may be reprinted provided that
the source is credited. Please send copies
of reprinted materials to the editor.

employment loss, or a total of 14 years.
Except for Rhode Island, the next
seven states on the list—Pennsylvania,
Michigan, Indiana, Illinois, Ohio, and
New York—are all concentrated in the
nation's industrial heartland.
• Conclusion
In recent months, the number of states
experiencing overall employment loss
has risen from zero in April 1989 to six
in June 1990. Because all states enjoyed
employment gains during most of 1988
and the first half of 1989, the obvious
question is, are we entering a period of
national recession? This Economic
Commentary provides some historical
perspective. Based on employment patterns over the last 40 years, six states
experiencing a slump concurrently does
not necessarily lead to a national recession. Moreover, the six states currently
posting employment losses are concentrated in the Northeast— a region that
has never led any national recession.
The analysis also highlights the regional
dynamics of national business cycles.
Shocks hit different regions at different
times, causing state employment cycles
to vary in timing and in duration of
employment slumps. Although this tendency may have been more pronounced
during the 1980s, rolling recessions in
particular regions also occurred in the
1950s and 1970s. Some shocks permeate enough regions to effect a national

downturn, while others remain contained in a specific area. The severity of
a state's employment downturn depends
on the composition and health of its
industries (particularly manufacturing),
which vary considerably across states.
Therefore, to view the national economy as a monolith rather than as a
mosaic of individual economies would
misrepresent the diverse experiences of
certain areas of our nation over national
business cycles.
•

Footnotes

1. This approach corresponds to the definition used to date national recessions. A different approach would be to examine state
employment changes in relation to state
employment trends.
2. The National Bureau of Economic
Research defines a recession as two successive quarters of decline in the GNP.

Randall W. Eberts is an assistant vice president and economist at the Federal Reserve
Bank of Cleveland. The author wishes to
thank Michael Bryan, Erica Groshen, and
Mark Sniderman for helpful comments and
suggestions. Ralph Day provided expert data
analysis.
The views stated herein are those of the
author and not necessarily those of the
Federal Reserve Bank of Cleveland or of the
Board of Governors of the Federal Reserve
Svstem.

BULK RATE
U.S. Postage Paid
Cleveland, OH
Permit No. 385